[PATCH] Allow booting from a Btrfs subvolume.

Done

Details

4 participants

Ludovic Courtès
Pierre Neidhardt
Christopher Baines
Maxim Cournoyer

Owner: unassigned

Submitted by: Maxim Cournoyer

Severity: normal

Maxim Cournoyer wrote on 5 Sep 2019 02:20

Recipients:(name . guix-patches)(address . guix-patches@gnu.org)

Message-ID:87sgpby4p9.fsf@gmail.com

Hello!

I'm sending this patch series to add support for booting off Btrfs

subvolumes. There was some interested shown on #guix, so hopefully

someone can test it on their system :-)

Before this change, it wasn't possible to pass the required options to

the Linux kernel as our init script would ignore them.

I'm not including system tests yet, as this will take a bit more time

and is starting to be a big change in itself.

Attachment: 0001-bootloader-grub-Allow-booting-from-a-Btrfs-subvolume.patch

From 3a628d1e731b2857a4c964484213cce980cb596f Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 16 Jul 2019 18:09:38 +0900
Subject: [PATCH 2/4] build: initrd: Fix "write-cpio-archive" return value.

* gnu/build/linux-initrd.scm (write-cpio-archive): Really return OUTPUT on
success, even when compression is disabled.
---
 gnu/build/linux-initrd.scm | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

Toggle diff (24 lines)diff --git a/gnu/build/linux-initrd.scm b/gnu/build/linux-initrd.scm
index 3aaa06d3a0..ea7de58553 100644
--- a/gnu/build/linux-initrd.scm
+++ b/gnu/build/linux-initrd.scm
@@ -71,8 +71,7 @@ COMPRESS? is true, compress it using GZIP.  On success, return OUTPUT."
       (cpio:write-cpio-archive files port
                                #:file->header cpio:file->cpio-header*)))
 
-  (or (not compress?)
-
+  (if compress?
       ;; Gzip insists on adding a '.gz' suffix and does nothing if the input
       ;; file already has that suffix.  Shuffle files around to placate it.
       (let* ((gz-suffix? (string-suffix? ".gz" output))
@@ -88,7 +87,6 @@ COMPRESS? is true, compress it using GZIP.  On success, return OUTPUT."
                (unless gz-suffix?
                  (rename-file (string-append output ".gz") output))
                output)))
-
       output))
 
 (define (cache-compiled-file-name file)
-- 
2.23.0

From 49ffe9a2645252bb708995169a9f1749f3982385 Mon Sep 17 00:00:00 2001

* gnu/build/linux-boot.scm (mount-root-file-system): Fix typo.

---

gnu/build/linux-boot.scm | 5 +++--

1 file changed, 3 insertions(+), 2 deletions(-)

Toggle diff (18 lines)diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index 7a30ebcef1..b4e6421b27 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -362,8 +362,9 @@ the last argument of `mknod'."
 (define* (mount-root-file-system root type
                                  #:key volatile-root?)
   "Mount the root file system of type TYPE at device ROOT.  If VOLATILE-ROOT?
-is true, mount ROOT read-only and make it a overlay with a writable tmpfs
-using the kernel build-in overlayfs."
+is true, mount ROOT read-only and make it an overlay with a writable tmpfs
+using the kernel built-in overlayfs."
+
   (if volatile-root?
       (begin
         (mkdir-p "/real-root")
-- 
2.23.0

From b56aea9c62b015c8a8b48827f9587b1578c83af3 Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Thu, 18 Jul 2019 04:59:25 +0900
Subject: [PATCH 4/4] linux-boot: Honor "rootflags" kernel argument.

* gnu/build/linux-boot.scm (mount-root-file-system): Add the optional FLAGS
and OPTIONS arguments; and document them.  Pass those to the `mount' calls.
(boot-system): Parse the "rootflags" kernel argument, and use it when calling
`mount-root-file-system'.
* doc/guix.texi (Initial RAM Disk): Document the use of the "rootflags"
argument.
---
 doc/guix.texi            | 19 +++++++++++++++++++
 gnu/build/linux-boot.scm | 22 +++++++++++++---------
 2 files changed, 32 insertions(+), 9 deletions(-)

Toggle diff (100 lines)diff --git a/doc/guix.texi b/doc/guix.texi
index cc7c91ac92..1e093b38a0 100644
--- a/doc/guix.texi
+++ b/doc/guix.texi
@@ -24761,6 +24761,25 @@ Instruct the initial RAM disk as well as the @command{modprobe} command
 must be a comma-separated list of module names---e.g.,
 @code{usbkbd,9pnet}.
 
+@item rootflags=@var{options}@dots{}
+@cindex mount options, passed to initrd
+@cindex rootflags, initrd
+This argument allows passing one or multiple file system specific mount
+options to the @code{mount} procedure used by the init script.  @var{options}
+must be a comma-separated list of option names or option-value pairs.  The
+following example instructs the initial RAM disk to mount the Btrfs subvolume
+named ``rootfs'' as the root file system, and to enable automatic file
+defragmentation:
+
+@example
+rootflags=subvol=rootfs,autodefrag
+@end example
+
+Specifying the subvolume to mount by its name, as shown above, is also used in
+Guix to produce a working Grub configuration for the Grub-based bootloaders
+when using a Btrfs subvolume for the root file system (@xref{Bootloader
+Configuration}).
+
 @item --repl
 Start a read-eval-print loop (REPL) from the initial RAM disk before it
 tries to load kernel modules and to mount the root file system.  Our
diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index b4e6421b27..b2d8f74a71 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -1,6 +1,7 @@
 ;;; GNU Guix --- Functional package management for GNU
 ;;; Copyright © 2013, 2014, 2015, 2016, 2017, 2018, 2019 Ludovic Courtès <ludo@gnu.org>
 ;;; Copyright © 2017 Mathieu Othacehe <m.othacehe@gmail.com>
+;;; Copyright © 2019 Maxim Cournoyer <maxim.cournoyer@gmail.com>
 ;;;
 ;;; This file is part of GNU Guix.
 ;;;
@@ -360,15 +361,16 @@ the last argument of `mknod'."
           (filter-map string->number (scandir "/proc")))))
 
 (define* (mount-root-file-system root type
+                                 #:optional (flags 0) options
                                  #:key volatile-root?)
-  "Mount the root file system of type TYPE at device ROOT.  If VOLATILE-ROOT?
-is true, mount ROOT read-only and make it an overlay with a writable tmpfs
-using the kernel built-in overlayfs."
-
+  "Mount the root file system of type TYPE at device ROOT.  The optional FLAGS
+and OPTIONS arguments behave the same as for the `mount' procedure.  If
+VOLATILE-ROOT?  is true, mount ROOT read-only and make it an overlay with a
+writable tmpfs using the kernel built-in overlayfs."
   (if volatile-root?
       (begin
         (mkdir-p "/real-root")
-        (mount root "/real-root" type MS_RDONLY)
+        (mount root "/real-root" type (logior MS_RDONLY flags) options)
         (mkdir-p "/rw-root")
         (mount "none" "/rw-root" "tmpfs")
 
@@ -385,11 +387,11 @@ using the kernel built-in overlayfs."
                "lowerdir=/real-root,upperdir=/rw-root/upper,workdir=/rw-root/work"))
       (begin
         (check-file-system root type)
-        (mount root "/root" type)))
+        (mount root "/root" type flags options)))
 
   ;; Make sure /root/etc/mtab is a symlink to /proc/self/mounts.
   (false-if-exception
-    (delete-file "/root/etc/mtab"))
+   (delete-file "/root/etc/mtab"))
   (mkdir-p "/root/etc")
   (symlink "/proc/self/mounts" "/root/etc/mtab"))
 
@@ -483,7 +485,8 @@ upon error."
      (mount-essential-file-systems)
      (let* ((args    (linux-command-line))
             (to-load (find-long-option "--load" args))
-            (root    (find-long-option "--root" args)))
+            (root    (find-long-option "--root" args))
+            (rootflags (find-long-option "rootflags" args)))
 
        (when (member "--repl" args)
          (start-repl))
@@ -526,7 +529,8 @@ upon error."
                              ((uuid root) => identity)
                              (else (file-system-label root)))))
              (mount-root-file-system (canonicalize-device-spec root)
-                                     root-fs-type
+                                     root-fs-type 0
+                                     rootflags
                                      #:volatile-root? volatile-root?))
            (mount "none" "/root" "tmpfs"))
 
-- 
2.23.0

Christopher Baines wrote on 8 Sep 2019 18:10

Recipients:(address . maxim.cournoyer@gmail.com)(address . 37305@debbugs.gnu.org)

Message-ID:87mufeojks.fsf@cbaines.net

Maxim Cournoyer <maxim.cournoyer@gmail.com> writes:

Toggle quote (12 lines)> Hello!
>
> I'm sending this patch series to add support for booting off Btrfs
> subvolumes.  There was some interested shown on #guix, so hopefully
> someone can test it on their system :-)
>
> Before this change, it wasn't possible to pass the required options to
> the Linux kernel as our init script would ignore them.
>
> I'm not including system tests yet, as this will take a bit more time
> and is starting to be a big change in itself.

Hey,

I haven't got around to looking at this yet, but I did write a system

test when I reported this bug [1], it's in the

0001-WIP-Btrfs-store-subvolume-test.patch file in the initial

message [2]. Maybe you could see if that passes with your changes here?

1: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=33517

2: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=33517#5

Chris

Ludovic Courtès wrote on 22 Sep 2019 23:43

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)(address . 37305@debbugs.gnu.org)

Message-ID:87y2yg3t3s.fsf@gnu.org

Hi!

Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:

Toggle quote (10 lines)> I'm sending this patch series to add support for booting off Btrfs
> subvolumes.  There was some interested shown on #guix, so hopefully
> someone can test it on their system :-)
>
> Before this change, it wasn't possible to pass the required options to
> the Linux kernel as our init script would ignore them.
>
> I'm not including system tests yet, as this will take a bit more time
> and is starting to be a big change in itself.

Did you have a chance to look at the system test that Chris mentioned?

Toggle quote (13 lines)> From 6858efa540d89c54ce377bfa6a6882e551cd2e56 Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Sun, 14 Jul 2019 20:50:23 +0900
> Subject: [PATCH 1/4] bootloader: grub: Allow booting from a Btrfs subvolume.
>
> * gnu/bootloader/grub.scm (prepend-subvol, arguments->subvol): New procedures.
> (grub-configuration-file): Use ARGUMENTS->SUBVOL to extract the subvolume name
> from the kernel arguments, and prepend it to the kernel and initrd paths using
> the PREPEND-SUBVOL procedure.
> * tests/grub.scm: Add tests for the "arguments->subvol" procedure.
> * doc/guix.texi (File Systems, (Bootloader Configuration): Document the use of
> Btrfs subvolumes.

[...]

Toggle quote (4 lines)

> +@cindex rootflags, Grub

> +@cindex Btrfs root subvolume, Grub

> +To allow using a Btrfs @emph{subvolume} for the root partition, the Grub-based

Nitpick: s/Grub/GRUB/

Toggle quote (16 lines)> --- a/gnu/bootloader/grub.scm
> +++ b/gnu/bootloader/grub.scm
> @@ -3,6 +3,7 @@
>  ;;; Copyright © 2016 Chris Marusich <cmmarusich@gmail.com>
>  ;;; Copyright © 2017 Leo Famulari <leo@famulari.name>
>  ;;; Copyright © 2017 Mathieu Othacehe <m.othacehe@gmail.com>
> +;;; Copyright © 2019 Maxim Cournoyer <maxim.cournoyer@gmail.com>
>  ;;;
>  ;;; This file is part of GNU Guix.
>  ;;;
> @@ -25,6 +26,8 @@
>    #:use-module (guix gexp)
>    #:use-module (gnu artwork)
>    #:use-module (gnu bootloader)
> +  #:use-module ((gnu build linux-modules) #:select (%not-comma))

‘%not-comma’ is not a great API and not quite related to linux-modules

:-), so it’s one of the rare cases where I’d rather keep it private and

duplicate it if needed.

Toggle quote (7 lines)

> +(define (arguments->subvol arguments)

> + "Return any \"subvol\" value given as an option to the \"rootflags\"

> +argument, or #f on failure."

> + (let* ((rootflags (find-long-option "rootflags" arguments))

> + (rootflags-options (and=> rootflags (cut string-tokenize <> %not-comma))))

> + (and=> rootflags-options (cut find-long-option "subvol" <>))))

Maybe rename ‘arguments->subvol’ to ‘linux-arguments-btrfs-subvolume’ or

similar?

Is “rootflags” the commonly-used name here?

Toggle quote (23 lines)> +    (let* ((device (menu-entry-device entry))
> +           (device-mount-point (menu-entry-device-mount-point entry))
> +           (label (menu-entry-label entry))
> +           (kernel (menu-entry-linux entry))
> +           (arguments (menu-entry-linux-arguments entry))
> +           (subvol (arguments->subvol arguments))
> +           (initrd (menu-entry-initrd entry)))
>        ;; Here DEVICE is the store and DEVICE-MOUNT-POINT is its mount point.
>        ;; Use the right file names for KERNEL and INITRD in case
>        ;; DEVICE-MOUNT-POINT is not "/", meaning that the store is on a
>        ;; separate partition.
> -      (let ((kernel  (strip-mount-point device-mount-point kernel))
> -            (initrd  (strip-mount-point device-mount-point initrd)))
> +
> +      ;; Also, in case a subvolume name could be extracted from the "subvol"
> +      ;; option given to the "rootflags" argument of the kernel, it is
> +      ;; prepended to the kernel and initrd paths, to allow booting from
> +      ;; a Btrfs subvolume.
> +      (let ((kernel (prepend-subvol subvol (strip-mount-point
> +                                            device-mount-point kernel)))
> +            (initrd (prepend-subvol subvol (strip-mount-point
> +                                            device-mount-point initrd))))

It might be clearer to have an explicit:

(if subvolume

#~(string-append #$subvolume …)

(strip-mount-point …))

than to hide the ‘if’ in ‘prepend-subvol’.

Regarding the interface, it’s the only time where we extract info from

the user-provided ‘kernel-arguments’ list. Usually it’s the other way

around: we have ‘file-systems’, and from that we build up the list of

kernel arguments.

Do you think it could be made to work similarly? (I’m not familiar with

Btrfs though.) That way, we wouldn’t have to parse the kernel

arguments, the user wouldn’t have to fiddle explicitly with kernel

arguments, and the end result might more easily work with all the

bootloaders.

Toggle quote (14 lines)>  (define (find-long-option option arguments)
>    "Find OPTION among ARGUMENTS, where OPTION is something like \"--load\".
> -Return the value associated with OPTION, or #f on failure."
> +Return the value associated with OPTION, or #f on failure.  Any non-string
> +arguments are ignored."
>    (let ((opt (string-append option "=")))
> -    (and=> (find (cut string-prefix? opt <>)
> +    (and=> (find (lambda (arg)
> +                   (and (string? arg)
> +                        (string-prefix? opt arg)))
>                   arguments)
>             (lambda (arg)
>               (substring arg (+ 1 (string-index arg #\=)))))))

Ignoring non-strings makes for a weird API IMO. :-)

I understand this is because, when using it on the host side, you may

now pass it gexps instead of strings. But perhaps that calls for a new

procedure?

Toggle quote (8 lines)

> From 3a628d1e731b2857a4c964484213cce980cb596f Mon Sep 17 00:00:00 2001

> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>

> Date: Tue, 16 Jul 2019 18:09:38 +0900

> Subject: [PATCH 2/4] build: initrd: Fix "write-cpio-archive" return value.

> * gnu/build/linux-initrd.scm (write-cpio-archive): Really return OUTPUT on

> success, even when compression is disabled.

Good catch, go for it!

Toggle quote (7 lines)

> From 49ffe9a2645252bb708995169a9f1749f3982385 Mon Sep 17 00:00:00 2001

> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>

> Date: Thu, 18 Jul 2019 07:23:48 +0900

> Subject: [PATCH 3/4] linux-boot: Fix typo.

> * gnu/build/linux-boot.scm (mount-root-file-system): Fix typo.

LGTM!

Toggle quote (12 lines)> From b56aea9c62b015c8a8b48827f9587b1578c83af3 Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Thu, 18 Jul 2019 04:59:25 +0900
> Subject: [PATCH 4/4] linux-boot: Honor "rootflags" kernel argument.
>
> * gnu/build/linux-boot.scm (mount-root-file-system): Add the optional FLAGS
> and OPTIONS arguments; and document them.  Pass those to the `mount' calls.
> (boot-system): Parse the "rootflags" kernel argument, and use it when calling
> `mount-root-file-system'.
> * doc/guix.texi (Initial RAM Disk): Document the use of the "rootflags"
> argument.

We’ll have to wait for patch #1, but in the final patch set, it should

probably come before patch #1, no?

Thank you!

Ludo’.

Maxim Cournoyer wrote on 12 Feb 2020 09:47

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)(address . 37305@debbugs.gnu.org)

Message-ID:87k14sfaz7.fsf@gmail.com

Hello Ludovic!

My much delayed answer to your review are the attached, much improved

patches.

The new test passes and I'm already using this on one of my system successfully.

Thanks for your patience,

Maxim

Attachment: 0001-gnu-tests-Reduce-the-time-required-to-run-the-system.patch

From 97d8a635eba34c7cf0708e99bf77ef9bad1344bf Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 12:57:29 -0500
Subject: [PATCH 2/8] gnu: linux-boot: Ensure volatile root is mounted
 read-only.

* gnu/build/linux-boot.scm (mount-root-file-system): Ensure MS_RDONLY is
present among the root file system flags when VOLATILE-ROOT? is #t.
---
 gnu/build/linux-boot.scm | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Toggle diff (21 lines)diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index 3d40a7d05d..4fb711b8f2 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -362,12 +362,12 @@ the last argument of `mknod'."
   "Mount the root file system of type TYPE at device ROOT. If VOLATILE-ROOT? is
 true, mount ROOT read-only and make it an overlay with a writable tmpfs using
 the kernel built-in overlayfs. FLAGS and OPTIONS indicates the options to use
-to mount ROOT."
+to mount ROOT, and behave the same as for the `mount' procedure."
 
   (if volatile-root?
       (begin
         (mkdir-p "/real-root")
-        (mount root "/real-root" type MS_RDONLY options)
+        (mount root "/real-root" type (logior MS_RDONLY flags) options)
         (mkdir-p "/rw-root")
         (mount "none" "/rw-root" "tmpfs")
 
-- 
2.25.0

From 870277ade6c20546566161adbe3e6f4a5c4368a8 Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 23:56:45 -0500
Subject: [PATCH 3/8] file-systems: Add a 'file-system-device->string'
 procedure.

* gnu/system/file-systems.scm (file-system-device->string): New procedure.
* gnu/system.scm (bootable-kernel-arguments): Use it.
* gnu/system/vm.scm (operating-system-uuid): Likewise.
* guix/scripts/system.scm (display-system-generation): Likewise.
---
 gnu/system.scm              | 15 +++++----------
 gnu/system/file-systems.scm | 13 +++++++++++++
 gnu/system/vm.scm           |  8 +-------
 guix/scripts/system.scm     |  7 +------
 4 files changed, 20 insertions(+), 23 deletions(-)

Toggle diff (96 lines)diff --git a/gnu/system.scm b/gnu/system.scm
index 01baa248a2..2e6d03272d 100644
--- a/gnu/system.scm
+++ b/gnu/system.scm
@@ -142,16 +142,11 @@
 (define (bootable-kernel-arguments system root-device)
   "Return a list of kernel arguments (gexps) to boot SYSTEM from ROOT-DEVICE."
   (list (string-append "--root="
-                       (cond ((uuid? root-device)
-
-                              ;; Note: Always use the DCE format because that's
-                              ;; what (gnu build linux-boot) expects for the
-                              ;; '--root' kernel command-line option.
-                              (uuid->string (uuid-bytevector root-device)
-                                            'dce))
-                             ((file-system-label? root-device)
-                              (file-system-label->string root-device))
-                             (else root-device)))
+                       ;; Note: Always use the DCE format because that's what
+                       ;; (gnu build linux-boot) expects for the '--root'
+                       ;; kernel command-line option.
+                       (file-system-device->string root-device
+                                                   #:uuid-type 'dce))
         #~(string-append "--system=" #$system)
         #~(string-append "--load=" #$system "/boot")))
 
diff --git a/gnu/system/file-systems.scm b/gnu/system/file-systems.scm
index d47a514b66..70a6febe3d 100644
--- a/gnu/system/file-systems.scm
+++ b/gnu/system/file-systems.scm
@@ -30,6 +30,7 @@
   #:export (file-system
             file-system?
             file-system-device
+            file-system-device->string
             file-system-title                     ;deprecated
             file-system-mount-point
             file-system-type
@@ -235,6 +236,18 @@ where both FILE1 and FILE2 are absolute file name.  For example:
               (()
                #f)))))))
 
+(define* (file-system-device->string device #:key (uuid-type 'dce))
+  "Return the string representations of the DEVICE field of a <file-system>
+record.  When the device is a UUID, its representation is chosen depening on
+UUID-TYPE, a symbol such as 'dce or 'iso9660."
+  (match device
+    ((? file-system-label?)
+     (file-system-label->string device))
+    ((? uuid?)
+     (uuid->string device))
+    ((? string?)
+     device)))
+
 (define (file-system-needed-for-boot? fs)
   "Return true if FS has the 'needed-for-boot?' flag set, or if it holds the
 store--e.g., if FS is the root file system."
diff --git a/gnu/system/vm.scm b/gnu/system/vm.scm
index 81b2e06ba2..03a511cdde 100644
--- a/gnu/system/vm.scm
+++ b/gnu/system/vm.scm
@@ -609,13 +609,7 @@ TYPE (one of 'iso9660 or 'dce).  Return a UUID object."
     (let ((device (file-system-device fs)))
       (list (file-system-mount-point fs)
             (file-system-type fs)
-            (cond ((file-system-label? device)
-                   (file-system-label->string device))
-                  ((uuid? device)
-                   (uuid->string device))
-                  ((string? device)
-                   device)
-                  (else #f))
+            (file-system-device->string device)
             (file-system-options fs))))
 
   (if (eq? type 'iso9660)
diff --git a/guix/scripts/system.scm b/guix/scripts/system.scm
index e69a3b6c97..b0386a1392 100644
--- a/guix/scripts/system.scm
+++ b/guix/scripts/system.scm
@@ -517,12 +517,7 @@ list of services."
               (cond ((uuid? root-device) 0)
                     ((file-system-label? root-device) 1)
                     (else 2))
-              (cond ((uuid? root-device)
-                     (uuid->string root-device))
-                    ((file-system-label? root-device)
-                     (file-system-label->string root-device))
-                    (else
-                     root-device)))
+              (file-system-device->string root-device))
 
       (format #t (G_ "  kernel: ~a~%") kernel)
 
-- 
2.25.0

From 347cccca292119104383aa7116b5eb7473d8a51b Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 14:00:06 -0500
Subject: [PATCH 4/8] gnu: linux-boot: Refactor boot-system.

The --root option can now be omitted, and inferred from the root file system
declaration instead.

* gnu/build/linux-boot.scm (boot-system): Remove nested definitions for
root-fs-type, root-fs-flags and root-fs-options, and bind those inside the
let* instead.  Make "--root" take precendence over the device field string
representation of the root file system.
* doc/guix.texi (Initial RAM Disk): Document that "--root" can be left
unspecified.
---
 doc/guix.texi            |  7 ++++---
 gnu/build/linux-boot.scm | 42 +++++++++++++++++++---------------------
 2 files changed, 24 insertions(+), 25 deletions(-)

Toggle diff (87 lines)diff --git a/doc/guix.texi b/doc/guix.texi
index 42d7cfa2e8..85cfabc2f3 100644
--- a/doc/guix.texi
+++ b/doc/guix.texi
@@ -25917,9 +25917,10 @@ service activation programs and then spawns the GNU@tie{}Shepherd, the
 initialization system.
 
 @item --root=@var{root}
-Mount @var{root} as the root file system.  @var{root} can be a
-device name like @code{/dev/sda1}, a file system label, or a file system
-UUID.
+Mount @var{root} as the root file system.  @var{root} can be a device
+name like @code{/dev/sda1}, a file system label, or a file system UUID.
+When unspecified, the device name from the root file system of the
+operating system declaration is used.
 
 @item --system=@var{system}
 Have @file{/run/booted-system} and @file{/run/current-system} point to
diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index 4fb711b8f2..28697e7bbf 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -467,26 +467,6 @@ upon error."
   (define (root-mount-point? fs)
     (string=? (file-system-mount-point fs) "/"))
 
-  (define root-fs-type
-    (or (any (lambda (fs)
-               (and (root-mount-point? fs)
-                    (file-system-type fs)))
-             mounts)
-        "ext4"))
-
-  (define root-fs-flags
-    (mount-flags->bit-mask (or (any (lambda (fs)
-                                      (and (root-mount-point? fs)
-                                           (file-system-flags fs)))
-                                    mounts)
-                               '())))
-
-  (define root-fs-options
-    (any (lambda (fs)
-           (and (root-mount-point? fs)
-                (file-system-options fs)))
-         mounts))
-
   (display "Welcome, this is GNU's early boot Guile.\n")
   (display "Use '--repl' for an initrd REPL.\n\n")
 
@@ -495,7 +475,25 @@ upon error."
       (mount-essential-file-systems)
       (let* ((args    (linux-command-line))
              (to-load (find-long-option "--load" args))
-             (root    (find-long-option "--root" args)))
+             (root-fs (find root-mount-point? mounts))
+             (root-fs-type (or (and=> root-fs file-system-type)
+                               "ext4"))
+             (root-device (and=> root-fs file-system-device))
+             (root-device-str (and=> root-device file-system-device->string))
+             ;; --root takes precedence over the 'device' field of the root
+             ;; <file-system> record.
+             (root (or (find-long-option "--root" args)
+                       root-device-str))
+             (root-fs-flags (mount-flags->bit-mask
+                             (or (and=> root-fs file-system-flags)
+                                 '())))
+             (root-fs-options (if root-fs
+                                  (file-system-options root-fs)
+                                  '()))
+             (root-options (if (null? root-fs-options)
+                               #f
+                               (file-system-options->str
+                                root-fs-options))))
 
         (when (member "--repl" args)
           (start-repl))
@@ -541,7 +539,7 @@ upon error."
                                       root-fs-type
                                       #:volatile-root? volatile-root?
                                       #:flags root-fs-flags
-                                      #:options root-fs-options))
+                                      #:options root-options))
             (mount "none" "/root" "tmpfs"))
 
         ;; Mount the specified file systems.
-- 
2.25.0

Attachment: 0005-file-systems-Represent-the-file-system-options-as-an.patch

From de7043998f2a2877f47e7952a435b711a0cbe9b5 Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 14:14:36 -0500
Subject: [PATCH 6/8] gnu: linux-boot: Honor the "--root-options" kernel
 argument.

* gnu/build/linux-boot.scm (boot-system): Parse the "--root-options" kernel
argument, and use it when calling `mount-root-file-system'.  Update doc.
* doc/guix.texi (Initial RAM Disk): Document the use of the "--root-options"
argument.
---
 doc/guix.texi            | 10 ++++++++++
 gnu/build/linux-boot.scm | 15 ++++++++++-----
 2 files changed, 20 insertions(+), 5 deletions(-)

Toggle diff (63 lines)diff --git a/doc/guix.texi b/doc/guix.texi
index 5d526b1aee..d6bfbd7b55 100644
--- a/doc/guix.texi
+++ b/doc/guix.texi
@@ -25935,6 +25935,16 @@ Instruct the initial RAM disk as well as the @command{modprobe} command
 must be a comma-separated list of module names---e.g.,
 @code{usbkbd,9pnet}.
 
+@item --root-options=@var{options}@dots{}
+@cindex mount options for the root file system, passed to initrd
+@cindex rootflags, initrd
+@cindex root-options, initrd
+This argument allows passing one or multiple file system specific mount
+options used by the initrd to mount the root file system.  @var{options}
+must be a comma-separated list of option names or option-value pairs.
+When unspecified, the value of the options field of the root file system
+of the operating system declaration is used.
+
 @item --repl
 Start a read-eval-print loop (REPL) from the initial RAM disk before it
 tries to load kernel modules and to mount the root file system.  Our
diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index 28697e7bbf..f65e942ebc 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -2,6 +2,7 @@
 ;;; Copyright © 2013, 2014, 2015, 2016, 2017, 2018, 2019 Ludovic Courtès <ludo@gnu.org>
 ;;; Copyright © 2017 Mathieu Othacehe <m.othacehe@gmail.com>
 ;;; Copyright © 2019 Guillaume Le Vaillant <glv@posteo.net>
+;;; Copyright © 2020 Maxim Cournoyer <maxim.cournoyer@gmail.com>
 ;;;
 ;;; This file is part of GNU Guix.
 ;;;
@@ -452,7 +453,8 @@ LINUX-MODULE-DIRECTORY, then installing KEYMAP-FILE with 'loadkeys' (if
 KEYMAP-FILE is true), then setting up QEMU guest networking if
 QEMU-GUEST-NETWORKING? is true, calling PRE-MOUNT, mounting the file systems
 specified in MOUNTS, and finally booting into the new root if any.  The initrd
-supports kernel command-line options '--load', '--root', and '--repl'.
+supports kernel command-line options '--load', '--root', '--root-options' and
+'--repl'.
 
 Mount the root file system, specified by the '--root' command-line argument,
 if any.
@@ -490,10 +492,13 @@ upon error."
              (root-fs-options (if root-fs
                                   (file-system-options root-fs)
                                   '()))
-             (root-options (if (null? root-fs-options)
-                               #f
-                               (file-system-options->str
-                                root-fs-options))))
+             ;; --root-options takes precedence over the 'options' field of the
+             ;; root <file-system> record.
+             (root-options (or (find-long-option "--root-options" args)
+                               (if (null? root-fs-options)
+                                   #f
+                                   (file-system-options->string
+                                    root-fs-options)))))
 
         (when (member "--repl" args)
           (start-repl))
-- 
2.25.0

From 549d585266e32cf413ae0511edd315f615f0c3a9 Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 14:27:19 -0500
Subject: [PATCH 7/8] gnu: linux-boot: Filter out file system independent
 options.

This fixes an issue where options such as "defaults", which are understood by
the command line program "mount", are not understood by the system call of the
same name, which is used in the initial RAM disk.

* gnu/system/file-systems.scm (%file-system-independent-mount-options): New variable.
(file-system-independent-mount-option?): New predicate.
* gnu/build/linux-boot.scm (boot-system): Use the above predicate to filter
out system independent mount options.
---
 gnu/build/linux-boot.scm    |  3 ++-
 gnu/system/file-systems.scm | 17 +++++++++++++++++
 2 files changed, 19 insertions(+), 1 deletion(-)

Toggle diff (49 lines)diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index f65e942ebc..8e55797549 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -490,7 +490,8 @@ upon error."
                              (or (and=> root-fs file-system-flags)
                                  '())))
              (root-fs-options (if root-fs
-                                  (file-system-options root-fs)
+                                  (remove file-system-independent-mount-option?
+                                          (file-system-options root-fs))
                                   '()))
              ;; --root-options takes precedence over the 'options' field of the
              ;; root <file-system> record.
diff --git a/gnu/system/file-systems.scm b/gnu/system/file-systems.scm
index eff89f146c..2dcf41ba57 100644
--- a/gnu/system/file-systems.scm
+++ b/gnu/system/file-systems.scm
@@ -46,6 +46,7 @@
             file-system-location
 
             file-system-type-predicate
+            file-system-independent-mount-option?
 
             file-system-label
             file-system-label?
@@ -563,4 +564,20 @@ system has the given TYPE."
   (lambda (fs)
     (string=? (file-system-type fs) type)))
 
+(define %file-system-independent-mount-options
+  ;; Taken from 'man 8 mount'.
+  '("async" "atime" "auto" "noatime" "noauto" "context" "defaults" "dev" "nodev"
+    "diratime" "nodiratime" "dirsync" "exec" "noexec" "group" "iversion"
+    "noiversion" "mand" "nomand" "_netdev" "nofail" "relatime" "norelatime"
+    "strictatime" "nostrictatime" "lazytime" "nolazytime" "suid" "nosuid"
+    "silent" "loud" "owner" "remount" "ro" "rw" "sync" "user" "nouser" "users"))
+
+(define (file-system-independent-mount-option? option)
+  "Predicate to check if a <file-system> option is file system independent."
+  (let ((option-name (if (pair? option)
+                         (car option)
+                         option)))
+    (or (string-prefix-ci? "x-" option-name)
+        (member option-name %file-system-independent-mount-options))))
+
 ;;; file-systems.scm ends here
-- 
2.25.0

Maxim Cournoyer wrote on 13 Feb 2020 21:27

Re: [bug#37305] [PATCH V2] Allow booting from a Btrfs subvolume.

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)(address . 37305@debbugs.gnu.org)

Message-ID:87lfp6b5cs.fsf_-_@gmail.com

I've fixed a few problems I've found with more extensive testing.

Attached is the version 2 of the patch series.

Attachment: 0001-gnu-tests-Reduce-the-time-required-to-run-the-system.patch

From 97d8a635eba34c7cf0708e99bf77ef9bad1344bf Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 12:57:29 -0500
Subject: [PATCH 2/9] gnu: linux-boot: Ensure volatile root is mounted
 read-only.

* gnu/build/linux-boot.scm (mount-root-file-system): Ensure MS_RDONLY is
present among the root file system flags when VOLATILE-ROOT? is #t.
---
 gnu/build/linux-boot.scm | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Toggle diff (21 lines)diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index 3d40a7d05d..4fb711b8f2 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -362,12 +362,12 @@ the last argument of `mknod'."
   "Mount the root file system of type TYPE at device ROOT. If VOLATILE-ROOT? is
 true, mount ROOT read-only and make it an overlay with a writable tmpfs using
 the kernel built-in overlayfs. FLAGS and OPTIONS indicates the options to use
-to mount ROOT."
+to mount ROOT, and behave the same as for the `mount' procedure."
 
   (if volatile-root?
       (begin
         (mkdir-p "/real-root")
-        (mount root "/real-root" type MS_RDONLY options)
+        (mount root "/real-root" type (logior MS_RDONLY flags) options)
         (mkdir-p "/rw-root")
         (mount "none" "/rw-root" "tmpfs")
 
-- 
2.23.0

From e8d6642d3597207657842c9ca4849f8660d06638 Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 23:56:45 -0500
Subject: [PATCH 3/9] file-systems: Add a 'file-system-device->string'
 procedure.

* gnu/system/file-systems.scm (file-system-device->string): New procedure.
* gnu/system.scm (bootable-kernel-arguments): Use it.
* gnu/system/vm.scm (operating-system-uuid): Likewise.
* guix/scripts/system.scm (display-system-generation): Likewise.
---
 gnu/system.scm              | 15 +++++----------
 gnu/system/file-systems.scm | 15 +++++++++++++++
 gnu/system/vm.scm           |  8 +-------
 guix/scripts/system.scm     |  7 +------
 4 files changed, 22 insertions(+), 23 deletions(-)

Toggle diff (98 lines)diff --git a/gnu/system.scm b/gnu/system.scm
index 01baa248a2..2e6d03272d 100644
--- a/gnu/system.scm
+++ b/gnu/system.scm
@@ -142,16 +142,11 @@
 (define (bootable-kernel-arguments system root-device)
   "Return a list of kernel arguments (gexps) to boot SYSTEM from ROOT-DEVICE."
   (list (string-append "--root="
-                       (cond ((uuid? root-device)
-
-                              ;; Note: Always use the DCE format because that's
-                              ;; what (gnu build linux-boot) expects for the
-                              ;; '--root' kernel command-line option.
-                              (uuid->string (uuid-bytevector root-device)
-                                            'dce))
-                             ((file-system-label? root-device)
-                              (file-system-label->string root-device))
-                             (else root-device)))
+                       ;; Note: Always use the DCE format because that's what
+                       ;; (gnu build linux-boot) expects for the '--root'
+                       ;; kernel command-line option.
+                       (file-system-device->string root-device
+                                                   #:uuid-type 'dce))
         #~(string-append "--system=" #$system)
         #~(string-append "--load=" #$system "/boot")))
 
diff --git a/gnu/system/file-systems.scm b/gnu/system/file-systems.scm
index d47a514b66..fc383d8a5a 100644
--- a/gnu/system/file-systems.scm
+++ b/gnu/system/file-systems.scm
@@ -30,6 +30,7 @@
   #:export (file-system
             file-system?
             file-system-device
+            file-system-device->string
             file-system-title                     ;deprecated
             file-system-mount-point
             file-system-type
@@ -235,6 +236,20 @@ where both FILE1 and FILE2 are absolute file name.  For example:
               (()
                #f)))))))
 
+(define* (file-system-device->string device #:key uuid-type)
+  "Return the string representations of the DEVICE field of a <file-system>
+record.  When the device is a UUID, its representation is chosen depening on
+UUID-TYPE, a symbol such as 'dce or 'iso9660."
+  (match device
+    ((? file-system-label?)
+     (file-system-label->string device))
+    ((? uuid?)
+     (if uuid-type
+         (uuid->string (uuid-bytevector device) uuid-type)
+         (uuid->string device)))
+    ((? string?)
+     device)))
+
 (define (file-system-needed-for-boot? fs)
   "Return true if FS has the 'needed-for-boot?' flag set, or if it holds the
 store--e.g., if FS is the root file system."
diff --git a/gnu/system/vm.scm b/gnu/system/vm.scm
index 81b2e06ba2..03a511cdde 100644
--- a/gnu/system/vm.scm
+++ b/gnu/system/vm.scm
@@ -609,13 +609,7 @@ TYPE (one of 'iso9660 or 'dce).  Return a UUID object."
     (let ((device (file-system-device fs)))
       (list (file-system-mount-point fs)
             (file-system-type fs)
-            (cond ((file-system-label? device)
-                   (file-system-label->string device))
-                  ((uuid? device)
-                   (uuid->string device))
-                  ((string? device)
-                   device)
-                  (else #f))
+            (file-system-device->string device)
             (file-system-options fs))))
 
   (if (eq? type 'iso9660)
diff --git a/guix/scripts/system.scm b/guix/scripts/system.scm
index e69a3b6c97..b0386a1392 100644
--- a/guix/scripts/system.scm
+++ b/guix/scripts/system.scm
@@ -517,12 +517,7 @@ list of services."
               (cond ((uuid? root-device) 0)
                     ((file-system-label? root-device) 1)
                     (else 2))
-              (cond ((uuid? root-device)
-                     (uuid->string root-device))
-                    ((file-system-label? root-device)
-                     (file-system-label->string root-device))
-                    (else
-                     root-device)))
+              (file-system-device->string root-device))
 
       (format #t (G_ "  kernel: ~a~%") kernel)
 
-- 
2.23.0

From 4f6e3955957beb5287e9d5a5d33b74725836e1ac Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 14:00:06 -0500
Subject: [PATCH 4/9] gnu: linux-boot: Refactor boot-system.

The --root option can now be omitted, and inferred from the root file system
declaration instead.

* gnu/build/linux-boot.scm (boot-system): Remove nested definitions for
root-fs-type, root-fs-flags and root-fs-options, and bind those inside the
let* instead.  Make "--root" take precendence over the device field string
representation of the root file system.
* doc/guix.texi (Initial RAM Disk): Document that "--root" can be left
unspecified.
---
 doc/guix.texi            |  7 ++++---
 gnu/build/linux-boot.scm | 42 +++++++++++++++++++---------------------
 2 files changed, 24 insertions(+), 25 deletions(-)

Toggle diff (87 lines)diff --git a/doc/guix.texi b/doc/guix.texi
index 42d7cfa2e8..85cfabc2f3 100644
--- a/doc/guix.texi
+++ b/doc/guix.texi
@@ -25917,9 +25917,10 @@ service activation programs and then spawns the GNU@tie{}Shepherd, the
 initialization system.
 
 @item --root=@var{root}
-Mount @var{root} as the root file system.  @var{root} can be a
-device name like @code{/dev/sda1}, a file system label, or a file system
-UUID.
+Mount @var{root} as the root file system.  @var{root} can be a device
+name like @code{/dev/sda1}, a file system label, or a file system UUID.
+When unspecified, the device name from the root file system of the
+operating system declaration is used.
 
 @item --system=@var{system}
 Have @file{/run/booted-system} and @file{/run/current-system} point to
diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index 4fb711b8f2..28697e7bbf 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -467,26 +467,6 @@ upon error."
   (define (root-mount-point? fs)
     (string=? (file-system-mount-point fs) "/"))
 
-  (define root-fs-type
-    (or (any (lambda (fs)
-               (and (root-mount-point? fs)
-                    (file-system-type fs)))
-             mounts)
-        "ext4"))
-
-  (define root-fs-flags
-    (mount-flags->bit-mask (or (any (lambda (fs)
-                                      (and (root-mount-point? fs)
-                                           (file-system-flags fs)))
-                                    mounts)
-                               '())))
-
-  (define root-fs-options
-    (any (lambda (fs)
-           (and (root-mount-point? fs)
-                (file-system-options fs)))
-         mounts))
-
   (display "Welcome, this is GNU's early boot Guile.\n")
   (display "Use '--repl' for an initrd REPL.\n\n")
 
@@ -495,7 +475,25 @@ upon error."
       (mount-essential-file-systems)
       (let* ((args    (linux-command-line))
              (to-load (find-long-option "--load" args))
-             (root    (find-long-option "--root" args)))
+             (root-fs (find root-mount-point? mounts))
+             (root-fs-type (or (and=> root-fs file-system-type)
+                               "ext4"))
+             (root-device (and=> root-fs file-system-device))
+             (root-device-str (and=> root-device file-system-device->string))
+             ;; --root takes precedence over the 'device' field of the root
+             ;; <file-system> record.
+             (root (or (find-long-option "--root" args)
+                       root-device-str))
+             (root-fs-flags (mount-flags->bit-mask
+                             (or (and=> root-fs file-system-flags)
+                                 '())))
+             (root-fs-options (if root-fs
+                                  (file-system-options root-fs)
+                                  '()))
+             (root-options (if (null? root-fs-options)
+                               #f
+                               (file-system-options->str
+                                root-fs-options))))
 
         (when (member "--repl" args)
           (start-repl))
@@ -541,7 +539,7 @@ upon error."
                                       root-fs-type
                                       #:volatile-root? volatile-root?
                                       #:flags root-fs-flags
-                                      #:options root-fs-options))
+                                      #:options root-options))
             (mount "none" "/root" "tmpfs"))
 
         ;; Mount the specified file systems.
-- 
2.23.0

Attachment: 0005-file-systems-Represent-the-file-system-options-as-an.patch

From 67135c925b07f2e077b4cd852e07178691a10164 Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 14:14:36 -0500
Subject: [PATCH 6/9] gnu: linux-boot: Honor the "--root-options" kernel
 argument.

* gnu/build/linux-boot.scm (boot-system): Parse the "--root-options" kernel
argument, and use it when calling `mount-root-file-system'.  Update doc.
* doc/guix.texi (Initial RAM Disk): Document the use of the "--root-options"
argument.
---
 doc/guix.texi            | 10 ++++++++++
 gnu/build/linux-boot.scm | 15 ++++++++++-----
 2 files changed, 20 insertions(+), 5 deletions(-)

Toggle diff (63 lines)diff --git a/doc/guix.texi b/doc/guix.texi
index 5d526b1aee..d6bfbd7b55 100644
--- a/doc/guix.texi
+++ b/doc/guix.texi
@@ -25935,6 +25935,16 @@ Instruct the initial RAM disk as well as the @command{modprobe} command
 must be a comma-separated list of module names---e.g.,
 @code{usbkbd,9pnet}.
 
+@item --root-options=@var{options}@dots{}
+@cindex mount options for the root file system, passed to initrd
+@cindex rootflags, initrd
+@cindex root-options, initrd
+This argument allows passing one or multiple file system specific mount
+options used by the initrd to mount the root file system.  @var{options}
+must be a comma-separated list of option names or option-value pairs.
+When unspecified, the value of the options field of the root file system
+of the operating system declaration is used.
+
 @item --repl
 Start a read-eval-print loop (REPL) from the initial RAM disk before it
 tries to load kernel modules and to mount the root file system.  Our
diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index 28697e7bbf..f65e942ebc 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -2,6 +2,7 @@
 ;;; Copyright © 2013, 2014, 2015, 2016, 2017, 2018, 2019 Ludovic Courtès <ludo@gnu.org>
 ;;; Copyright © 2017 Mathieu Othacehe <m.othacehe@gmail.com>
 ;;; Copyright © 2019 Guillaume Le Vaillant <glv@posteo.net>
+;;; Copyright © 2020 Maxim Cournoyer <maxim.cournoyer@gmail.com>
 ;;;
 ;;; This file is part of GNU Guix.
 ;;;
@@ -452,7 +453,8 @@ LINUX-MODULE-DIRECTORY, then installing KEYMAP-FILE with 'loadkeys' (if
 KEYMAP-FILE is true), then setting up QEMU guest networking if
 QEMU-GUEST-NETWORKING? is true, calling PRE-MOUNT, mounting the file systems
 specified in MOUNTS, and finally booting into the new root if any.  The initrd
-supports kernel command-line options '--load', '--root', and '--repl'.
+supports kernel command-line options '--load', '--root', '--root-options' and
+'--repl'.
 
 Mount the root file system, specified by the '--root' command-line argument,
 if any.
@@ -490,10 +492,13 @@ upon error."
              (root-fs-options (if root-fs
                                   (file-system-options root-fs)
                                   '()))
-             (root-options (if (null? root-fs-options)
-                               #f
-                               (file-system-options->str
-                                root-fs-options))))
+             ;; --root-options takes precedence over the 'options' field of the
+             ;; root <file-system> record.
+             (root-options (or (find-long-option "--root-options" args)
+                               (if (null? root-fs-options)
+                                   #f
+                                   (file-system-options->string
+                                    root-fs-options)))))
 
         (when (member "--repl" args)
           (start-repl))
-- 
2.23.0

From cb060af5ea56427e1fd63ced5f9c9edd3ae61f76 Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 14:27:19 -0500
Subject: [PATCH 7/9] gnu: linux-boot: Filter out file system independent
 options.

This fixes an issue where options such as "defaults", which are understood by
the command line program "mount", are not understood by the system call of the
same name, which is used in the initial RAM disk.

* gnu/system/file-systems.scm (%file-system-independent-mount-options): New variable.
(file-system-independent-mount-option?): New predicate.
* gnu/build/linux-boot.scm (boot-system): Use the above predicate to filter
out system independent mount options.
---
 gnu/build/linux-boot.scm    |  3 ++-
 gnu/system/file-systems.scm | 17 +++++++++++++++++
 2 files changed, 19 insertions(+), 1 deletion(-)

Toggle diff (49 lines)diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index f65e942ebc..8e55797549 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -490,7 +490,8 @@ upon error."
                              (or (and=> root-fs file-system-flags)
                                  '())))
              (root-fs-options (if root-fs
-                                  (file-system-options root-fs)
+                                  (remove file-system-independent-mount-option?
+                                          (file-system-options root-fs))
                                   '()))
              ;; --root-options takes precedence over the 'options' field of the
              ;; root <file-system> record.
diff --git a/gnu/system/file-systems.scm b/gnu/system/file-systems.scm
index 6dc0e6814e..2c3c159d04 100644
--- a/gnu/system/file-systems.scm
+++ b/gnu/system/file-systems.scm
@@ -46,6 +46,7 @@
             file-system-location
 
             file-system-type-predicate
+            file-system-independent-mount-option?
 
             file-system-label
             file-system-label?
@@ -565,4 +566,20 @@ system has the given TYPE."
   (lambda (fs)
     (string=? (file-system-type fs) type)))
 
+(define %file-system-independent-mount-options
+  ;; Taken from 'man 8 mount'.
+  '("async" "atime" "auto" "noatime" "noauto" "context" "defaults" "dev" "nodev"
+    "diratime" "nodiratime" "dirsync" "exec" "noexec" "group" "iversion"
+    "noiversion" "mand" "nomand" "_netdev" "nofail" "relatime" "norelatime"
+    "strictatime" "nostrictatime" "lazytime" "nolazytime" "suid" "nosuid"
+    "silent" "loud" "owner" "remount" "ro" "rw" "sync" "user" "nouser" "users"))
+
+(define (file-system-independent-mount-option? option)
+  "Predicate to check if a <file-system> option is file system independent."
+  (let ((option-name (if (pair? option)
+                         (car option)
+                         option)))
+    (or (string-prefix-ci? "x-" option-name)
+        (member option-name %file-system-independent-mount-options))))
+
 ;;; file-systems.scm ends here
-- 
2.23.0

Attachment: 0008-bootloader-grub-Allow-booting-from-a-Btrfs-subvolume.patch

Ludovic Courtès wrote on 14 Feb 2020 18:22

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)(address . 37305@debbugs.gnu.org)

Message-ID:8736bdf5il.fsf@gnu.org

Hi Maxim!

Great to see Btrfs support improving; many people will love that!

Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:

Toggle quote (16 lines)> From 3640bea548826e1c1ec9b766da1fdfe4791d7452 Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Sun, 17 Nov 2019 06:01:00 +0900
> Subject: [PATCH 1/9] gnu: tests: Reduce the time required to run the system
>  tests.
>
> When setting the GUIX_DEV_HACKS environment variable, the Guix package used
> inside the instrumented VMs recycles the binaries already found in the Guix
> checkout of the developer instead of rebuilding Guix from scratch.  This
> brings the time required for this component from 20+ minutes down to 2-3
> minutes on an X200 machine.
>
> * gnu/packages/package-management.scm (current-guix/pre-built): New procedure.
> * build-aux/run-system-tests.scm (tests-for-channel-instance): Use it, when
> GUIX_DEV_HACKS is defined.

I understand the need, but I’d really like to avoid that; it’s too

fragile IMO.

But I have good news! First, commit

887fd835a7c90f720d36a211478012547feaead0 really improved things by

avoiding the full ‘guix’ package rebuild (and we’re only talking about

the installation tests; other tests are just fine.) Second, there are

improvements to Guile that will appear in 3.0.1/2.2.7 that make

compilation of big files roughly twice as fast.

There’s still room for improvement, but I’d rather work in those

directions. WDYT?

Toggle quote (9 lines)

> From 97d8a635eba34c7cf0708e99bf77ef9bad1344bf Mon Sep 17 00:00:00 2001

> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>

> Date: Tue, 11 Feb 2020 12:57:29 -0500

> Subject: [PATCH 2/9] gnu: linux-boot: Ensure volatile root is mounted

> read-only.

> * gnu/build/linux-boot.scm (mount-root-file-system): Ensure MS_RDONLY is

> present among the root file system flags when VOLATILE-ROOT? is #t.

(You can drop the “gnu:” prefix.)

OK!

Toggle quote (11 lines)> From e8d6642d3597207657842c9ca4849f8660d06638 Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Tue, 11 Feb 2020 23:56:45 -0500
> Subject: [PATCH 3/9] file-systems: Add a 'file-system-device->string'
>  procedure.
>
> * gnu/system/file-systems.scm (file-system-device->string): New procedure.
> * gnu/system.scm (bootable-kernel-arguments): Use it.
> * gnu/system/vm.scm (operating-system-uuid): Likewise.
> * guix/scripts/system.scm (display-system-generation): Likewise.

OK!

Toggle quote (15 lines)> From 4f6e3955957beb5287e9d5a5d33b74725836e1ac Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Tue, 11 Feb 2020 14:00:06 -0500
> Subject: [PATCH 4/9] gnu: linux-boot: Refactor boot-system.
>
> The --root option can now be omitted, and inferred from the root file system
> declaration instead.
>
> * gnu/build/linux-boot.scm (boot-system): Remove nested definitions for
> root-fs-type, root-fs-flags and root-fs-options, and bind those inside the
> let* instead.  Make "--root" take precendence over the device field string
> representation of the root file system.
> * doc/guix.texi (Initial RAM Disk): Document that "--root" can be left
> unspecified.

[...]

Toggle quote (9 lines)

> @item --root=@var{root}

> -Mount @var{root} as the root file system. @var{root} can be a

> -device name like @code{/dev/sda1}, a file system label, or a file system

> -UUID.

> +Mount @var{root} as the root file system. @var{root} can be a device

> +name like @code{/dev/sda1}, a file system label, or a file system UUID.

> +When unspecified, the device name from the root file system of the

> +operating system declaration is used.

Oh! Does it always work? That makes me wonder why we’ve been carrying

‘--root’ and I’m not sure if I’m forgetting a good reason to do it that

way.

Toggle quote (24 lines)>        (mount-essential-file-systems)
>        (let* ((args    (linux-command-line))
>               (to-load (find-long-option "--load" args))
> -             (root    (find-long-option "--root" args)))
> +             (root-fs (find root-mount-point? mounts))
> +             (root-fs-type (or (and=> root-fs file-system-type)
> +                               "ext4"))
> +             (root-device (and=> root-fs file-system-device))
> +             (root-device-str (and=> root-device file-system-device->string))
> +             ;; --root takes precedence over the 'device' field of the root
> +             ;; <file-system> record.
> +             (root (or (find-long-option "--root" args)
> +                       root-device-str))
> +             (root-fs-flags (mount-flags->bit-mask
> +                             (or (and=> root-fs file-system-flags)
> +                                 '())))
> +             (root-fs-options (if root-fs
> +                                  (file-system-options root-fs)
> +                                  '()))
> +             (root-options (if (null? root-fs-options)
> +                               #f
> +                               (file-system-options->str
> +                                root-fs-options))))

Since ‘file-system-device->string’ is lossy, I think we should do it the

other way around: convert the ‘--root’ string to a <file-system>.

Perhaps that bit can be moved to a separate procedure, like

‘root-string->file-system’.

Also, the (or … "ext4") bit doesn’t sound great, but perhaps it’ll be

unnecessary if we do with something as outlined above?

Toggle quote (19 lines)> From af61745d8b686755a5d9deb9e21c9eac624fb43e Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Wed, 25 Sep 2019 22:43:41 +0900
> Subject: [PATCH 5/9] file-systems: Represent the file system options as an
>  alist.
>
> This allows accessing the parameter values easily, without having to parse a
> string.
>
> * gnu/system/file-systems.scm (<file-system>): Update the default value of the
> OPTIONS field, doc.
> (%file-system-options): Field accessor renamed from `file-system-options'.
> (file-system-options, file-system-options->string): New procedures.
> * gnu/build/file-systems.scm (mount-file-system): Adapt.
> * gnu/services/base.scm (file-system->fstab-entry): Likewise.
> * tests/file-systems.scm: New tests.
> * doc/guix.texi (File Systems): Document the modified default value of the
> 'file-system-options' field.

The main issue I see with this change is that mount(2) takes raw strings

for the options. There’s a convention to have those strings look like

“KEY1=VALUE1,KEY2=VALUE2”, but it’s just a convention.

As a rule of thumb, I’d rather have our interface be as close as

possible to the actual mount(2) interface, which means taking strings.

Now, we can surely add helper procedures to parse options that follow

the above conventions.

WDYT?

Toggle quote (9 lines)

> + ;; Support the deprecated options format (a string).

> + (define (options-string->options-list str)

> + (let ((option-list (string-split str #\,)))

> + (map (lambda (param)

> + (if (string-contains param "=")

> + (apply cons (string-split param #\=))

> + param))

> + option-list)))

I think we’d want to split only on the first ‘=’ sign, meaning we should

use ‘string-index’ etc. instead of ‘string-split’.

Toggle quote (11 lines)> From 67135c925b07f2e077b4cd852e07178691a10164 Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Tue, 11 Feb 2020 14:14:36 -0500
> Subject: [PATCH 6/9] gnu: linux-boot: Honor the "--root-options" kernel
>  argument.
>
> * gnu/build/linux-boot.scm (boot-system): Parse the "--root-options" kernel
> argument, and use it when calling `mount-root-file-system'.  Update doc.
> * doc/guix.texi (Initial RAM Disk): Document the use of the "--root-options"
> argument.

Hmm do we really need this extra option? :-)

(Also, in hindsight, I think it was a mistake to call them

‘--something’. Following the common naming convention, we should rather

call these options ‘gnu.something’.)

Toggle quote (15 lines)> From cb060af5ea56427e1fd63ced5f9c9edd3ae61f76 Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Tue, 11 Feb 2020 14:27:19 -0500
> Subject: [PATCH 7/9] gnu: linux-boot: Filter out file system independent
>  options.
>
> This fixes an issue where options such as "defaults", which are understood by
> the command line program "mount", are not understood by the system call of the
> same name, which is used in the initial RAM disk.
>
> * gnu/system/file-systems.scm (%file-system-independent-mount-options): New variable.
> (file-system-independent-mount-option?): New predicate.
> * gnu/build/linux-boot.scm (boot-system): Use the above predicate to filter
> out system independent mount options.

[...]

Toggle quote (8 lines)

> +(define %file-system-independent-mount-options

> + ;; Taken from 'man 8 mount'.

> + '("async" "atime" "auto" "noatime" "noauto" "context" "defaults" "dev" "nodev"

> + "diratime" "nodiratime" "dirsync" "exec" "noexec" "group" "iversion"

> + "noiversion" "mand" "nomand" "_netdev" "nofail" "relatime" "norelatime"

> + "strictatime" "nostrictatime" "lazytime" "nolazytime" "suid" "nosuid"

> + "silent" "loud" "owner" "remount" "ro" "rw" "sync" "user" "nouser" "users"))

I’d rather avoid it. In general, as much as possible, I think we should

pass options to the kernel without trying to be “smart”. It’s often the

safe strategy.

WDYT?

Toggle quote (20 lines)> From 6cf2ece21683e98544f8f46675aef58d5a7231fd Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Sun, 14 Jul 2019 20:50:23 +0900
> Subject: [PATCH 8/9] bootloader: grub: Allow booting from a Btrfs subvolume.
>
> * gnu/bootloader/grub.scm (grub-configuration-file) [btrfs-subvolume-path]:
> New parameter.  When it is defined, prepend its value to the kernel and
> initrd file paths.
> * gnu/bootloader/depthcharge.scm (depthcharge-configuration-file): Adapt.
> * gnu/bootloader/extlinux.scm (extlinux-configuration-file): Likewise.
> * gnu/system/file-systems.scm (btrfs-subvolume?)
> (btrfs-store-subvolume-path): New procedures.
> * gnu/system.scm (operating-system-bootcfg): Specify the Btrfs subvolume path
> of the GNU store to the `operating-system-bootcfg' procedure, using the new
> BTRFS-SUBVOLUME-PATH argument.
> * doc/guix.texi (File Systems): Add a Btrfs subsection to document the use of
> subvolumes.  Document the new `properties' field of the `<file-system>'
> record.
> * gnu/tests/install.scm: Add test "btrfs-root-on-subvolume-os".

Neat!

Toggle quote (13 lines)>  (define* (grub-configuration-file config entries
>                                    #:key
>                                    (system (%current-system))
> -                                  (old-entries '()))
> +                                  (old-entries '())
> +                                  btrfs-subvolume-path)
>    "Return the GRUB configuration file corresponding to CONFIG, a
>  <bootloader-configuration> object, and where the store is available at
>  STORE-FS, a <file-system> object.  OLD-ENTRIES is taken to be a list of menu
> -entries corresponding to old generations of the system."
> +entries corresponding to old generations of the system.  BTRFS-SUBVOLUME-PATH
> +may be used to specify on which subvolume a Btrfs root file system resides."

(Nitpick: s/path/file name/ :-))

It’s a bit problematic that (1) GRUB needs explicit Btrfs support, and

(2) other bootloaders will silently ignore the option, due to

#:allow-other-keys.

I don’t have a better idea, but it’d be great if Btrfs support could be

made bootloader-independent, and if it could be somewhat

not-too-btrfs-specific, if that is possible at all.

Thoughts?

Toggle quote (18 lines)> +  (properties       file-system-properties        ; list of name-value pairs
> +                    (default '()))
>    (location         file-system-location
>                      (default (current-source-location))
>                      (innate)))
> @@ -582,4 +589,48 @@ system has the given TYPE."
>      (or (string-prefix-ci? "x-" option-name)
>          (member option-name %file-system-independent-mount-options))))
>  
> +(define (btrfs-subvolume? fs)
> +  "Predicate to check if FS, a file-system object, is a Btrfs subvolume."
> +  (and-let* ((btrfs-file-system? (string= "btrfs" (file-system-type fs)))
> +             (option-keys (map (match-lambda
> +                                 ((key . value) key)
> +                                 (key key))
> +                               (file-system-options fs))))
> +    (find (cut string-prefix? "subvol" <>) option-keys)))

I wonder if we can avoid special support in the <file-system> API for

Btrfs.

Toggle quote (5 lines)

> + (error "The store is on a Btrfs subvolume, but the \

> +subvolume name is unknown.

> +Hint: Define the \"btrfs-subvolume-path\" file system property or

> +use the \"subvol\" Btrfs file system option."))))

Rather use ‘raise’ with ‘&message’ and ‘&fix-hint’ conditions.

Pheeew, that was a long patch series!

Perhaps we can split the continuation of this thread in sizable chunks!

Thanks for working on this,

Ludo’.

Maxim Cournoyer wrote on 16 Feb 2020 06:36

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)(address . 37305@debbugs.gnu.org)

Message-ID:87blpzozz7.fsf@gmail.com

Hello!

Thanks for the prompt feedback.

Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (35 lines)> Hi Maxim!
>
> Great to see Btrfs support improving; many people will love that!
>
> Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:
>
>> From 3640bea548826e1c1ec9b766da1fdfe4791d7452 Mon Sep 17 00:00:00 2001
>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>> Date: Sun, 17 Nov 2019 06:01:00 +0900
>> Subject: [PATCH 1/9] gnu: tests: Reduce the time required to run the system
>>  tests.
>>
>> When setting the GUIX_DEV_HACKS environment variable, the Guix package used
>> inside the instrumented VMs recycles the binaries already found in the Guix
>> checkout of the developer instead of rebuilding Guix from scratch.  This
>> brings the time required for this component from 20+ minutes down to 2-3
>> minutes on an X200 machine.
>>
>> * gnu/packages/package-management.scm (current-guix/pre-built): New procedure.
>> * build-aux/run-system-tests.scm (tests-for-channel-instance): Use it, when
>> GUIX_DEV_HACKS is defined.
>
> I understand the need, but I’d really like to avoid that; it’s too
> fragile IMO.
>
> But I have good news!  First, commit
> 887fd835a7c90f720d36a211478012547feaead0 really improved things by
> avoiding the full ‘guix’ package rebuild (and we’re only talking about
> the installation tests; other tests are just fine.)  Second, there are
> improvements to Guile that will appear in 3.0.1/2.2.7 that make
> compilation of big files roughly twice as fast.
>
> There’s still room for improvement, but I’d rather work in those
> directions.  WDYT?

With a little bit more love (inheriting from the Guix package from the

inferior captured at build time), I don't see the hack being any less

fragile than the Guix checkout compiled and ran with ./pre-inst-env.

There's no arguing that it *is* a hack, but:

1) Being labeled as such (GUIX_DEV_HACKS)

2) Being undocumented

3) Only being enabled explicitly (through that GUIX_DEV_HACKS

environment variable)

4) Can be reverted easily in the future when the default, clean implementation is

fast enough to obsolete it.

To me means its targeted to developers who understand the nature of the

hack and is provided without any warranty.

And while it's exciting that Guile's compilation time is going to be

improved, no compiler's going to be as efficient as "no compilation"

;-).

Toggle quote (11 lines)>> From 97d8a635eba34c7cf0708e99bf77ef9bad1344bf Mon Sep 17 00:00:00 2001
>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>> Date: Tue, 11 Feb 2020 12:57:29 -0500
>> Subject: [PATCH 2/9] gnu: linux-boot: Ensure volatile root is mounted
>>  read-only.
>>
>> * gnu/build/linux-boot.scm (mount-root-file-system): Ensure MS_RDONLY is
>> present among the root file system flags when VOLATILE-ROOT? is #t.
>
> (You can drop the “gnu:” prefix.)

Done.

I never know before looking at past logs (and then sometimes it's a

mixed bag). Is there any mechanical process for selecting the right

commit prefix? :-)

Toggle quote (45 lines)> OK!
>
>> From e8d6642d3597207657842c9ca4849f8660d06638 Mon Sep 17 00:00:00 2001
>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>> Date: Tue, 11 Feb 2020 23:56:45 -0500
>> Subject: [PATCH 3/9] file-systems: Add a 'file-system-device->string'
>>  procedure.
>>
>> * gnu/system/file-systems.scm (file-system-device->string): New procedure.
>> * gnu/system.scm (bootable-kernel-arguments): Use it.
>> * gnu/system/vm.scm (operating-system-uuid): Likewise.
>> * guix/scripts/system.scm (display-system-generation): Likewise.
>
> OK!
>
>> From 4f6e3955957beb5287e9d5a5d33b74725836e1ac Mon Sep 17 00:00:00 2001
>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>> Date: Tue, 11 Feb 2020 14:00:06 -0500
>> Subject: [PATCH 4/9] gnu: linux-boot: Refactor boot-system.
>>
>> The --root option can now be omitted, and inferred from the root file system
>> declaration instead.
>>
>> * gnu/build/linux-boot.scm (boot-system): Remove nested definitions for
>> root-fs-type, root-fs-flags and root-fs-options, and bind those inside the
>> let* instead.  Make "--root" take precendence over the device field string
>> representation of the root file system.
>> * doc/guix.texi (Initial RAM Disk): Document that "--root" can be left
>> unspecified.
>
> [...]
>
>>  @item --root=@var{root}
>> -Mount @var{root} as the root file system.  @var{root} can be a
>> -device name like @code{/dev/sda1}, a file system label, or a file system
>> -UUID.
>> +Mount @var{root} as the root file system.  @var{root} can be a device
>> +name like @code{/dev/sda1}, a file system label, or a file system UUID.
>> +When unspecified, the device name from the root file system of the
>> +operating system declaration is used.
>
> Oh!  Does it always work?  That makes me wonder why we’ve been carrying
> ‘--root’ and I’m not sure if I’m forgetting a good reason to do it that
> way.

If the documentation is accurate, it should :-), given that --root gets
written as a string to the GRUB configuration file, and that the doc
says it's possible to give it as a device name, label or UUID.

About why providing options such as --root or --root-options in the
first place; I pondered about this as well, especially after making the
file systems from operating system able to be mounted with all their
(file system independent -- more on that later) options.  A reason I
came up with was that it allows to experiment at the GRUB command line
and change the root device, or perhaps the root options.  One use case
would be debugging the right options to pass to a file system driver in
case of a mistake in the operating system declaration.

Toggle quote (30 lines)>
>>        (mount-essential-file-systems)
>>        (let* ((args    (linux-command-line))
>>               (to-load (find-long-option "--load" args))
>> -             (root    (find-long-option "--root" args)))
>> +             (root-fs (find root-mount-point? mounts))
>> +             (root-fs-type (or (and=> root-fs file-system-type)
>> +                               "ext4"))
>> +             (root-device (and=> root-fs file-system-device))
>> +             (root-device-str (and=> root-device file-system-device->string))
>> +             ;; --root takes precedence over the 'device' field of the root
>> +             ;; <file-system> record.
>> +             (root (or (find-long-option "--root" args)
>> +                       root-device-str))
>> +             (root-fs-flags (mount-flags->bit-mask
>> +                             (or (and=> root-fs file-system-flags)
>> +                                 '())))
>> +             (root-fs-options (if root-fs
>> +                                  (file-system-options root-fs)
>> +                                  '()))
>> +             (root-options (if (null? root-fs-options)
>> +                               #f
>> +                               (file-system-options->str
>> +                                root-fs-options))))
>
> Since ‘file-system-device->string’ is lossy, I think we should do it the
> other way around: convert the ‘--root’ string to a <file-system>.
> Perhaps that bit can be moved to a separate procedure, like
> ‘root-string->file-system’.

Done!

Toggle quote (3 lines)

> Also, the (or … "ext4") bit doesn’t sound great, but perhaps it’ll be

> unnecessary if we do with something as outlined above?

That bit was already there; it exists because we allow an operating
system definition to be made without any root file system (for use with
the 'guix system container' or 'guix system vm' commands IIRC).  What
could be simpler would be to always require such information to be
embedded in the operating system declaration: when an interactive user
command would need to modify the root file system, it could create a new
operating-system object, inheriting from the original one.

Toggle quote (31 lines)>> From af61745d8b686755a5d9deb9e21c9eac624fb43e Mon Sep 17 00:00:00 2001
>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>> Date: Wed, 25 Sep 2019 22:43:41 +0900
>> Subject: [PATCH 5/9] file-systems: Represent the file system options as an
>>  alist.
>>
>> This allows accessing the parameter values easily, without having to parse a
>> string.
>>
>> * gnu/system/file-systems.scm (<file-system>): Update the default value of the
>> OPTIONS field, doc.
>> (%file-system-options): Field accessor renamed from `file-system-options'.
>> (file-system-options, file-system-options->string): New procedures.
>> * gnu/build/file-systems.scm (mount-file-system): Adapt.
>> * gnu/services/base.scm (file-system->fstab-entry): Likewise.
>> * tests/file-systems.scm: New tests.
>> * doc/guix.texi (File Systems): Document the modified default value of the
>> 'file-system-options' field.
>
> The main issue I see with this change is that mount(2) takes raw strings
> for the options.  There’s a convention to have those strings look like
> “KEY1=VALUE1,KEY2=VALUE2”, but it’s just a convention.
>
> As a rule of thumb, I’d rather have our interface be as close as
> possible to the actual mount(2) interface, which means taking strings.
>
> Now, we can surely add helper procedures to parse options that follow
> the above conventions.
>
> WDYT?

To me, it's an implementation detail that I'd rather abstract away (or
make optional, like in this patch).  Just like we provide a higher level
configuration for services instead of requiring the user to input the
configuration in the native format of the tool (or allowing for both).
The idea for this format was taken from a discussion here:
http://issues.guix.info/issue/33517#3.

Are we really targeting mount(2)?  The commit
9d3053819dfd834a1c29a03427c41d8524b8a7d5 (which you co-authored :-))
mentions 'man 8 mount' for the file system options. This should be
stressed in the manual, as someone attempting to pass file system
independent options (perhaps taking their options line from a previous
/etc/fstab config), could prevent the system from booting.

I mistakenly was under the impression that mount(8) was targeted, since
the main interaction I observed between a file system object and my
system was the /etc/fstab file getting produced.

Doesn't targeting mount(2) mean that it becomes impossible to give all
the options someone would possibly want to produce their /etc/fstab
file?  Suppose you want the 'ro' option for one of your partition in
/etc/fstab; that wouldn't fly currently, unless I'm missing something.
It seems to me we should target mount(8), and internally translate the
file system independent mount(8) options into mount(2) flags or
vice-versa but targeting the higher level and going down makes more
sense to me.

Toggle quote (12 lines)>> +  ;; Support the deprecated options format (a string).
>> +  (define (options-string->options-list str)
>> +    (let ((option-list (string-split str #\,)))
>> +      (map (lambda (param)
>> +             (if (string-contains param "=")
>> +                 (apply cons (string-split param #\=))
>> +                 param))
>> +           option-list)))
>
> I think we’d want to split only on the first ‘=’ sign, meaning we should
> use ‘string-index’ etc. instead of ‘string-split’.

Good catch!

Toggle quote (13 lines)>> From 67135c925b07f2e077b4cd852e07178691a10164 Mon Sep 17 00:00:00 2001
>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>> Date: Tue, 11 Feb 2020 14:14:36 -0500
>> Subject: [PATCH 6/9] gnu: linux-boot: Honor the "--root-options" kernel
>>  argument.
>>
>> * gnu/build/linux-boot.scm (boot-system): Parse the "--root-options" kernel
>> argument, and use it when calling `mount-root-file-system'.  Update doc.
>> * doc/guix.texi (Initial RAM Disk): Document the use of the "--root-options"
>> argument.
>
> Hmm do we really need this extra option?  :-)

It is not strictly needed but allows the user to experiment/troubleshoot

with the init RAM disk from GRUB as discussed earlier for --root. Do

you think it has enough value to be kept?

Toggle quote (4 lines)

> (Also, in hindsight, I think it was a mistake to call them

> ‘--something’. Following the common naming convention, we should rather

> call these options ‘gnu.something’.)

Is this convention detailed somewhere? I haven't found it in 'Standards'.

Toggle quote (29 lines)>> From cb060af5ea56427e1fd63ced5f9c9edd3ae61f76 Mon Sep 17 00:00:00 2001
>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>> Date: Tue, 11 Feb 2020 14:27:19 -0500
>> Subject: [PATCH 7/9] gnu: linux-boot: Filter out file system independent
>>  options.
>>
>> This fixes an issue where options such as "defaults", which are understood by
>> the command line program "mount", are not understood by the system call of the
>> same name, which is used in the initial RAM disk.
>>
>> * gnu/system/file-systems.scm (%file-system-independent-mount-options): New variable.
>> (file-system-independent-mount-option?): New predicate.
>> * gnu/build/linux-boot.scm (boot-system): Use the above predicate to filter
>> out system independent mount options.
>
> [...]
>
>> +(define %file-system-independent-mount-options
>> +  ;; Taken from 'man 8 mount'.
>> +  '("async" "atime" "auto" "noatime" "noauto" "context" "defaults" "dev" "nodev"
>> +    "diratime" "nodiratime" "dirsync" "exec" "noexec" "group" "iversion"
>> +    "noiversion" "mand" "nomand" "_netdev" "nofail" "relatime" "norelatime"
>> +    "strictatime" "nostrictatime" "lazytime" "nolazytime" "suid" "nosuid"
>> +    "silent" "loud" "owner" "remount" "ro" "rw" "sync" "user" "nouser" "users"))
>
> I’d rather avoid it.  In general, as much as possible, I think we should
> pass options to the kernel without trying to be “smart”.  It’s often the
> safe strategy.

Let's revisit this topic after we've sorted the mount(8) vs mount(2)

problem above.

I'll address the rest in a part 2.

Thank you for your time and patience!

Maxim

Ludovic Courtès wrote on 16 Feb 2020 12:11

Making system installation tests faster

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)(address . 37305@debbugs.gnu.org)

Message-ID:8736ba3hxo.fsf_-_@gnu.org

Hi Maxim,

Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:

Toggle quote (41 lines)> Ludovic Courtès <ludo@gnu.org> writes:
>
>> Hi Maxim!
>>
>> Great to see Btrfs support improving; many people will love that!
>>
>> Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:
>>
>>> From 3640bea548826e1c1ec9b766da1fdfe4791d7452 Mon Sep 17 00:00:00 2001
>>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>>> Date: Sun, 17 Nov 2019 06:01:00 +0900
>>> Subject: [PATCH 1/9] gnu: tests: Reduce the time required to run the system
>>>  tests.
>>>
>>> When setting the GUIX_DEV_HACKS environment variable, the Guix package used
>>> inside the instrumented VMs recycles the binaries already found in the Guix
>>> checkout of the developer instead of rebuilding Guix from scratch.  This
>>> brings the time required for this component from 20+ minutes down to 2-3
>>> minutes on an X200 machine.
>>>
>>> * gnu/packages/package-management.scm (current-guix/pre-built): New procedure.
>>> * build-aux/run-system-tests.scm (tests-for-channel-instance): Use it, when
>>> GUIX_DEV_HACKS is defined.
>>
>> I understand the need, but I’d really like to avoid that; it’s too
>> fragile IMO.
>>
>> But I have good news!  First, commit
>> 887fd835a7c90f720d36a211478012547feaead0 really improved things by
>> avoiding the full ‘guix’ package rebuild (and we’re only talking about
>> the installation tests; other tests are just fine.)  Second, there are
>> improvements to Guile that will appear in 3.0.1/2.2.7 that make
>> compilation of big files roughly twice as fast.
>>
>> There’s still room for improvement, but I’d rather work in those
>> directions.  WDYT?
>
> With a little bit more love (inheriting from the Guix package from the
> inferior captured at build time), I don't see the hack being any less
> fragile than the Guix checkout compiled and ran with ./pre-inst-env.

It’s grabbing files from the working tree, with heuristics to minimize

the changes in incorporating files that shouldn’t be there; I think it’s

fragile. :-)

It also breaks that idea that things compiled by Guix are well under

control.

Toggle quote (16 lines)> There's no arguing that it *is* a hack, but:
>
> 1) Being labeled as such (GUIX_DEV_HACKS)
> 2) Being undocumented
> 3) Only being enabled explicitly (through that GUIX_DEV_HACKS
> environment variable)
> 4) Can be reverted easily in the future when the default, clean implementation is
> fast enough to obsolete it.
>
> To me means its targeted to developers who understand the nature of the
> hack and is provided without any warranty.
>
> And while it's exciting that Guile's compilation time is going to be
> improved, no compiler's going to be as efficient as "no compilation"
> ;-).

True! But still, we’d have to maintain that in the meantime and deal

with any “surprising” effects it has for those using it.

FWIW, when testing (gnu installer tests), I only have to rebuild

“guix-system-tests.drv”, which is fast. And in fact, I can also not

rebuild anything by doing:

(define operating-system-with-current-guix identity)

because in this particular case, I know it’s not necessary to have the

current ‘guix’ package.

I suppose this would also be an option in your case, when testing the

Btrfs changes, no?

Perhaps we should make this case more easily accessible through an

environment variable?

Thanks,

Ludo’.

Maxim Cournoyer wrote on 18 Feb 2020 14:37

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)(address . 37305@debbugs.gnu.org)

Message-ID:87sgj8j9sf.fsf@gmail.com

Hi Ludovic,

Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (49 lines)> Hi Maxim,
>
> Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:
>
>> Ludovic Courtès <ludo@gnu.org> writes:
>>
>>> Hi Maxim!
>>>
>>> Great to see Btrfs support improving; many people will love that!
>>>
>>> Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:
>>>
>>>> From 3640bea548826e1c1ec9b766da1fdfe4791d7452 Mon Sep 17 00:00:00 2001
>>>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>>>> Date: Sun, 17 Nov 2019 06:01:00 +0900
>>>> Subject: [PATCH 1/9] gnu: tests: Reduce the time required to run the system
>>>>  tests.
>>>>
>>>> When setting the GUIX_DEV_HACKS environment variable, the Guix package used
>>>> inside the instrumented VMs recycles the binaries already found in the Guix
>>>> checkout of the developer instead of rebuilding Guix from scratch.  This
>>>> brings the time required for this component from 20+ minutes down to 2-3
>>>> minutes on an X200 machine.
>>>>
>>>> * gnu/packages/package-management.scm (current-guix/pre-built): New procedure.
>>>> * build-aux/run-system-tests.scm (tests-for-channel-instance): Use it, when
>>>> GUIX_DEV_HACKS is defined.
>>>
>>> I understand the need, but I’d really like to avoid that; it’s too
>>> fragile IMO.
>>>
>>> But I have good news!  First, commit
>>> 887fd835a7c90f720d36a211478012547feaead0 really improved things by
>>> avoiding the full ‘guix’ package rebuild (and we’re only talking about
>>> the installation tests; other tests are just fine.)  Second, there are
>>> improvements to Guile that will appear in 3.0.1/2.2.7 that make
>>> compilation of big files roughly twice as fast.
>>>
>>> There’s still room for improvement, but I’d rather work in those
>>> directions.  WDYT?
>>
>> With a little bit more love (inheriting from the Guix package from the
>> inferior captured at build time), I don't see the hack being any less
>> fragile than the Guix checkout compiled and ran with ./pre-inst-env.
>
> It’s grabbing files from the working tree, with heuristics to minimize
> the changes in incorporating files that shouldn’t be there; I think it’s
> fragile.  :-)

Fragile or not, it's helped me getting this change tested much faster

than I could have had otherwise, which is the point of that hack :-).

Benchmarks below.

Toggle quote (22 lines)> It also breaks that idea that things compiled by Guix are well under
> control.
>
>> There's no arguing that it *is* a hack, but:
>>
>> 1) Being labeled as such (GUIX_DEV_HACKS)
>> 2) Being undocumented
>> 3) Only being enabled explicitly (through that GUIX_DEV_HACKS
>> environment variable)
>> 4) Can be reverted easily in the future when the default, clean implementation is
>> fast enough to obsolete it.
>>
>> To me means its targeted to developers who understand the nature of the
>> hack and is provided without any warranty.
>>
>> And while it's exciting that Guile's compilation time is going to be
>> improved, no compiler's going to be as efficient as "no compilation"
>> ;-).
>
> True!  But still, we’d have to maintain that in the meantime and deal
> with any “surprising” effects it has for those using it.

Yeah.

Toggle quote (12 lines)> FWIW, when testing (gnu installer tests), I only have to rebuild
> “guix-system-tests.drv”, which is fast.  And in fact, I can also not
> rebuild anything by doing:
>
>   (define operating-system-with-current-guix identity)
>
> because in this particular case, I know it’s not necessary to have the
> current ‘guix’ package.
>
> I suppose this would also be an option in your case, when testing the
> Btrfs changes, no?

I think it is important that the 'installer' image corresponds to the

'current-guix' in my case, because it has to generate the correct GRUB

configuration file at 'guix system init' time.

Below are some analysis of the derivations triggered in different scenarios.

Without the hack:

Toggle snippet (112 lines)./pre-inst-env make check-system TESTS=btrfs-root-on-subvolume-os
Compiling Scheme modules...
random seed for tests: 1581852580
Computing Guix derivation for 'x86_64-linux'... \
Running 1 system tests...
substitute: updating substitutes from 'http://10.42.0.199:80'... 100.0%
substitute: updating substitutes from 'https://ci.guix.gnu.org'... 100.0%
The following derivations will be built:
   /gnu/store/3g9bawa6c876lml0318g2knqdcv2ahv9-btrfs-root-on-subvolume-os.drv
   /gnu/store/vswbrzp8ab5wz0fyzrj9j72mqcbd6c4a-installation.drv
   /gnu/store/nf1wd5bgminylzvwq4mjs5293hpr1mmi-disk-image.drv
   /gnu/store/45334hsl3hs5yx13x7qz1b1sc83vkbzj-module-import-compiled.drv
   /gnu/store/dmi03i3jv9wwkw10fm57g7fq97kwlscf-module-import.drv
   /gnu/store/4cp5pqyn554rdmv9l4vfhnw7lgasv18q-raw-initrd.drv
   /gnu/store/vx147spsmrxkiy0qcmvxn341fpz4fmrv-init.drv
   /gnu/store/d6ri0lh953vaksiz5h8jx681h3i35vvm-module-import-compiled.drv
   /gnu/store/b0gv94llsbvrlzkxs3naz7crcv1glxy1-grub.cfg.drv
   /gnu/store/gcslrfymsqsp6kqckla41zkyaa7fy9in-raw-initrd.drv
   /gnu/store/m8bbfcdn1irn7a41syy9arl1yz34hbln-init.drv
   /gnu/store/hbfgqnralir92slqr9hhfw8cxqrlgv06-system.drv
   /gnu/store/0v4aq6yprk8az183rpdjx90jnxw3w838-parameters.drv
   /gnu/store/8daaamb92bbkpdwz96a19r81f7vnk40k-gc-roots.drv
   /gnu/store/f48dwgib52ciaq1nh06nhks6g7gvy7bx-system.drv
   /gnu/store/4ll7vvq5wxq8c2xl8qw5dbim7nrz0yda-raw-initrd.drv
   /gnu/store/mxdfkkfyy3ld6656a01z0lw7gxgjns25-init.drv
   /gnu/store/72fjlv3kp75629as4bc49yp9hgf80pv7-boot.drv
   /gnu/store/xra5iv1ayyiml1kvpcjg5rkrgjbzy6lb-activate.scm.drv
   /gnu/store/xww4xc7rpc1bw1xk9sn6vy3hazjk1gcq-module-import-compiled.drv
   /gnu/store/y8hd9wzr9ypn2pfcl5al0dckijzw36x6-shepherd.conf.drv
   /gnu/store/14mdcbdyqrihnl7x8lqygncxdbsvmhv0-shepherd-file-system--home.go.drv
   /gnu/store/clv6h5sxwgchw1kkf9cnafgvd5ms4p4r-shepherd-file-system--home.scm.drv
   /gnu/store/zc84ngrjdvbc4kibdvzipf9sxjj50rgz-module-import-compiled.drv
   /gnu/store/6cca3yng6k4zl812ysdgmj5lyw8lmwln-shepherd-file-system--gnu-store.go.drv
   /gnu/store/rplp3l38j9jm0fism0ka1iqbfhx7h3ir-shepherd-file-system--gnu-store.scm.drv
   /gnu/store/729f2ranwr6klr90rvrx5pfka30iwazr-shepherd-user-homes.go.drv
   /gnu/store/2cadf4viazfvzlazj996vmbpq3cw1hbr-shepherd-user-homes.scm.drv
   /gnu/store/7j784x4jpxqlcisikn8cin4s0n1vv7z7-shepherd-term-auto.go.drv
   /gnu/store/zgz6p1mss3jwra7d4vzh1dxy9b4llnlg-shepherd-term-auto.scm.drv
   /gnu/store/7n6y2ljl0gp4s568wqbj5iwdkw5hrhk6-shepherd-file-system--dev-shm.go.drv
   /gnu/store/apsplfgbzzq38b5q7p06jxjn37kx74b2-shepherd-file-system--dev-shm.scm.drv
   /gnu/store/qj459c7mlrls9imm1nw98lw9fgqh4yyy-shepherd-udev.go.drv
   /gnu/store/78klph6xipr232a938qjhm8v4qviacvd-shepherd-udev.scm.drv
   /gnu/store/z103dcbbvz3inx6ax7s966zxmy1s38hn-shepherd-file-system--dev-pts.go.drv
   /gnu/store/5ss1d8nwn4dv9vb69h1iyw884b61y487-shepherd-file-system--dev-pts.scm.drv
   /gnu/store/z1f6a7gfxxszaa0lm68xfl9hbglz48cw-parameters.drv
   /gnu/store/vvg05l3pgkyz1rky6f2cn0d1r2i1b6yr-system.drv
   /gnu/store/1k5k3kd11l3wabcc8sci8m04nvc6wllw-raw-initrd.drv
   /gnu/store/51ay1gi0ajsbrgm7wy6j1njyy6pcjlzb-init.drv
   /gnu/store/jjhnm39rqckl6r0jf7l43000nlydb2vj-parameters.drv
   /gnu/store/wdm0acks0pzl3hiam7knffl7n6xmyvrm-boot.drv
   /gnu/store/59mhndnjhsq6yspag4miv4b71rb8v683-activate.scm.drv
   /gnu/store/z0ik75q9nshaik97jsigd69vriyjijgp-shepherd.conf.drv
   /gnu/store/6pxsd37k5nl5viqadd57n0hah53nzcp7-shepherd-user-homes.go.drv
   /gnu/store/y9rj23r1xmyb5b4fwrh75pb5yr61clvp-shepherd-user-homes.scm.drv
   /gnu/store/h0xbdk4hw1i4byy8xnrnaqmlwnfgrswz-profile.drv
   /gnu/store/sxdgmvfjakapmbyqmxjgvarsaaxc2mw7-profile.drv
   /gnu/store/hxcadzr771dzl99kylz64d9qz2qai8wd-guix-000000000.drv
   /gnu/store/95lcx1xx41p4vcwqlwli7phny6hpp8b0-guix-manual.drv
   /gnu/store/9yz0spp1h158zdh8wmg2r7sspnkm0qfc-guix-translated-texinfo.drv
   /gnu/store/g6ggidnpaahws37qgx15mvkhfll0ns06-guix-daemon.drv
   /gnu/store/jpb2jfk00q5nh5fg3k3cf6hcyw49zz84-guix-command.drv
   /gnu/store/14zcma6hkqszaab37hda736i24r41rv4-guix-module-union.drv
   /gnu/store/sfzd86425cl744lr58yh877i7z5qbd68-guix-000000000-modules.drv
   /gnu/store/71a40hzi2d2wdn1vjl5amfq00caqlh3m-guix-system-tests-modules.drv
   /gnu/store/kc3zg69yabaglvllj4l7hzgs72v9mk26-guix-system-tests.drv
   /gnu/store/5zhgzvnyswm8g50aha95rvdx6dpmw329-guix-packages-base.drv
   /gnu/store/d7df7qw5allqcm2qn75dn60wz1dg50cq-guix-system.drv
   /gnu/store/ily3pqqvfkjz5622fxz6mxg31jbiqb21-guix-packages.drv
   /gnu/store/xgpy1v810xi76w8ivpfi7gg5h7vg0ska-guix-cli.drv
   /gnu/store/7ih7cmbhz1p7a4m7yrgy0s2jawqc666n-guix-packages-base-modules.drv
   /gnu/store/j15kzs5m07pq3nmgqr5rr47cryq8arfs-guix-system-modules.drv
   /gnu/store/rhk9cknm5wbk30slzzrr52app8mfn4a2-guix-packages-modules.drv
   /gnu/store/x05mz632h3ph0k9n3s5y14m51qy3k929-guix-cli-modules.drv
   /gnu/store/fs2g9iin23iq4505jsxjnrjaxnxp1wzl-inferior-script.scm.drv
   /gnu/store/y4hv2s267zjz2gzja3cj428a7z1dln98-profile.drv
   /gnu/store/kx5h4kxkz28gahlgkxni02i0rf8558vd-boot.drv
   /gnu/store/nhi2s5zcyd5frjqag4jnp7fvhzbx6kd6-activate.scm.drv
   /gnu/store/qnchikcfqx8hn0a7cn45f0pnf664s0jk-shepherd.conf.drv
   /gnu/store/7n4cqlsbn7hywlwssxvax6j6b57gm1q4-shepherd-user-homes.go.drv
   /gnu/store/dl333r3dshzi5nb4qqrvcsp2hmk3hd9p-shepherd-user-homes.scm.drv
   /gnu/store/i96lsb7shixcczc62bfadbg1g73qp38l-shepherd-udev.go.drv
   /gnu/store/yv3maqmnqqpiqj2jk5ss574pzvgh9jin-shepherd-udev.scm.drv
   /gnu/store/s5w901l6zl0ql800bc9cndnskqdhnsxm-shepherd-file-system--tmp.go.drv
   /gnu/store/a5nb6vhs86g76ww5bmpphi398pggnga7-shepherd-file-system--tmp.scm.drv
   /gnu/store/sjn7zb56x72lh6i7arkh6gpd639g7d17-shepherd-guix-daemon.go.drv
   /gnu/store/rww6z6m3nz6s7h54xklwj3plylg1f0yc-shepherd-guix-daemon.scm.drv
   /gnu/store/lmbja82mymwsn3kmxwfy8g7s46xac443-linux-vm-loader.drv
   /gnu/store/z8y0v374himjlig4qwjisv1xa7rk7r0l-builder-in-linux-vm.drv
   /gnu/store/14bw0x4dq0g262vk4pjh3dlha6x3gz0s-module-import-compiled.drv
   /gnu/store/nbkfy5wjpia5y7hj1p9j5hmbv42j4zp5-module-import.drv
The following grafts will be made:
   /gnu/store/qn8j1cqp2lfvghiv4sy5rps4kv2hz02y-guile3.0-git-0.3.0.drv
   /gnu/store/dnl877iincgcw7lw493h5nd4llxqjw7a-libgit2-0.28.4.drv
The following profile hooks will be built:
   /gnu/store/96274clz8449s7s18a91sa31d1ilmi0p-fonts-dir.drv
   /gnu/store/30lvdbci7v9zqgn3nqjy1vgbpv5k4b98-ca-certificate-bundle.drv
   /gnu/store/b5dspr8576n0j3vl2cjsr7zrfdp380cw-guix-package-cache.drv
   /gnu/store/chjdvqjbrnb7cqla43n5rvqia5d4f4mg-fonts-dir.drv
   /gnu/store/kh7wixk3i5db2ym48ddyrdbggw8ccxp8-info-dir.drv
   /gnu/store/x3bp2acfy04w9r9l11w30l9z0hpsii5a-manual-database.drv
   /gnu/store/bhla9fihy35pgc3bwszfxvsfjir71v88-glib-schemas.drv
   /gnu/store/isb4444058amfr4sipqmffcrgisx75gs-ca-certificate-bundle.drv
   /gnu/store/l8qgqaxq1xcb8jk3lqmzzrfiv8hzpna1-xdg-desktop-database.drv
   /gnu/store/x0vhd98kg66pzfwffbi45f76126by816-xdg-mime-database.drv
   /gnu/store/xs4k7nmrpbjsy0a5k6lsbdxc2n4i23g4-info-dir.drv
   /gnu/store/zffqfc9w73sy815av7hwincbqpl7hbkr-manual-database.drv
building /gnu/store/dmi03i3jv9wwkw10fm57g7fq97kwlscf-module-import.drv...
successfully built /gnu/store/dmi03i3jv9wwkw10fm57g7fq97kwlscf-module-import.drv
building /gnu/store/9yz0spp1h158zdh8wmg2r7sspnkm0qfc-guix-translated-texinfo.drv...
^Cmake: *** [Makefile:5711: check-system] Interrupt

With the hack:

Toggle snippet (82 lines)./pre-inst-env make check-system TESTS=btrfs-root-on-subvolume-os GUIX_DEV_HACKS=1
Compiling Scheme modules...
random seed for tests: 1581851372
Running 1 system tests...
The following derivations will be built:
   /gnu/store/l4z1363klr8a58qmm3pjzy2y5bxsr6bd-btrfs-root-on-subvolume-os.drv
   /gnu/store/zlg6gxml4awbkbs2s6n1b0z3ysrqvy0h-installation.drv
   /gnu/store/89vdpn33f47y0510im604snwfmbpswmm-disk-image.drv
   /gnu/store/45334hsl3hs5yx13x7qz1b1sc83vkbzj-module-import-compiled.drv
   /gnu/store/4cp5pqyn554rdmv9l4vfhnw7lgasv18q-raw-initrd.drv
   /gnu/store/vx147spsmrxkiy0qcmvxn341fpz4fmrv-init.drv
   /gnu/store/d6ri0lh953vaksiz5h8jx681h3i35vvm-module-import-compiled.drv
   /gnu/store/58912rdjjk4jdv767svb2axp0llmn8ch-linux-vm-loader.drv
   /gnu/store/hvcm9abi0br38plwpngqlvw1l0060fni-builder-in-linux-vm.drv
   /gnu/store/14bw0x4dq0g262vk4pjh3dlha6x3gz0s-module-import-compiled.drv
   /gnu/store/nbkfy5wjpia5y7hj1p9j5hmbv42j4zp5-module-import.drv
   /gnu/store/shc8hzzqi40pdpip0773i2g9x2apnn0h-grub.cfg.drv
   /gnu/store/gcslrfymsqsp6kqckla41zkyaa7fy9in-raw-initrd.drv
   /gnu/store/m8bbfcdn1irn7a41syy9arl1yz34hbln-init.drv
   /gnu/store/zifmvsbiw6x4gs15k29l1nvyzxkszafw-system.drv
   /gnu/store/0v4aq6yprk8az183rpdjx90jnxw3w838-parameters.drv
   /gnu/store/8daaamb92bbkpdwz96a19r81f7vnk40k-gc-roots.drv
   /gnu/store/f48dwgib52ciaq1nh06nhks6g7gvy7bx-system.drv
   /gnu/store/4ll7vvq5wxq8c2xl8qw5dbim7nrz0yda-raw-initrd.drv
   /gnu/store/mxdfkkfyy3ld6656a01z0lw7gxgjns25-init.drv
   /gnu/store/72fjlv3kp75629as4bc49yp9hgf80pv7-boot.drv
   /gnu/store/xra5iv1ayyiml1kvpcjg5rkrgjbzy6lb-activate.scm.drv
   /gnu/store/xww4xc7rpc1bw1xk9sn6vy3hazjk1gcq-module-import-compiled.drv
   /gnu/store/y8hd9wzr9ypn2pfcl5al0dckijzw36x6-shepherd.conf.drv
   /gnu/store/14mdcbdyqrihnl7x8lqygncxdbsvmhv0-shepherd-file-system--home.go.drv
   /gnu/store/clv6h5sxwgchw1kkf9cnafgvd5ms4p4r-shepherd-file-system--home.scm.drv
   /gnu/store/zc84ngrjdvbc4kibdvzipf9sxjj50rgz-module-import-compiled.drv
   /gnu/store/6cca3yng6k4zl812ysdgmj5lyw8lmwln-shepherd-file-system--gnu-store.go.drv
   /gnu/store/rplp3l38j9jm0fism0ka1iqbfhx7h3ir-shepherd-file-system--gnu-store.scm.drv
   /gnu/store/729f2ranwr6klr90rvrx5pfka30iwazr-shepherd-user-homes.go.drv
   /gnu/store/2cadf4viazfvzlazj996vmbpq3cw1hbr-shepherd-user-homes.scm.drv
   /gnu/store/7j784x4jpxqlcisikn8cin4s0n1vv7z7-shepherd-term-auto.go.drv
   /gnu/store/zgz6p1mss3jwra7d4vzh1dxy9b4llnlg-shepherd-term-auto.scm.drv
   /gnu/store/7n6y2ljl0gp4s568wqbj5iwdkw5hrhk6-shepherd-file-system--dev-shm.go.drv
   /gnu/store/apsplfgbzzq38b5q7p06jxjn37kx74b2-shepherd-file-system--dev-shm.scm.drv
   /gnu/store/qj459c7mlrls9imm1nw98lw9fgqh4yyy-shepherd-udev.go.drv
   /gnu/store/78klph6xipr232a938qjhm8v4qviacvd-shepherd-udev.scm.drv
   /gnu/store/z103dcbbvz3inx6ax7s966zxmy1s38hn-shepherd-file-system--dev-pts.go.drv
   /gnu/store/5ss1d8nwn4dv9vb69h1iyw884b61y487-shepherd-file-system--dev-pts.scm.drv
   /gnu/store/z1f6a7gfxxszaa0lm68xfl9hbglz48cw-parameters.drv
   /gnu/store/vvg05l3pgkyz1rky6f2cn0d1r2i1b6yr-system.drv
   /gnu/store/1k5k3kd11l3wabcc8sci8m04nvc6wllw-raw-initrd.drv
   /gnu/store/51ay1gi0ajsbrgm7wy6j1njyy6pcjlzb-init.drv
   /gnu/store/jjhnm39rqckl6r0jf7l43000nlydb2vj-parameters.drv
   /gnu/store/wdm0acks0pzl3hiam7knffl7n6xmyvrm-boot.drv
   /gnu/store/59mhndnjhsq6yspag4miv4b71rb8v683-activate.scm.drv
   /gnu/store/z0ik75q9nshaik97jsigd69vriyjijgp-shepherd.conf.drv
   /gnu/store/6pxsd37k5nl5viqadd57n0hah53nzcp7-shepherd-user-homes.go.drv
   /gnu/store/y9rj23r1xmyb5b4fwrh75pb5yr61clvp-shepherd-user-homes.scm.drv
   /gnu/store/kammkm5z0ihjqalg0ijn69ffhdyzki47-profile.drv
   /gnu/store/07kxlp8a99mvx2zidnlz6rm4ak2az8d0-guix-1.0.1-13.50299ad+.drv
   /gnu/store/m9783fc7m360dk582b3mv2wgrsdz6amk-boot.drv
   /gnu/store/3v1whxi5qkp9cyfc3vjpvziwc9ms7y7k-shepherd.conf.drv
   /gnu/store/50halfax0zbiwpy47z6374ybh1mq1d5r-shepherd-guix-daemon.go.drv
   /gnu/store/g4m4mqiz9z4yd2iwzmb6f81vidw0k4bg-shepherd-guix-daemon.scm.drv
   /gnu/store/7n4cqlsbn7hywlwssxvax6j6b57gm1q4-shepherd-user-homes.go.drv
   /gnu/store/dl333r3dshzi5nb4qqrvcsp2hmk3hd9p-shepherd-user-homes.scm.drv
   /gnu/store/i96lsb7shixcczc62bfadbg1g73qp38l-shepherd-udev.go.drv
   /gnu/store/yv3maqmnqqpiqj2jk5ss574pzvgh9jin-shepherd-udev.scm.drv
   /gnu/store/s5w901l6zl0ql800bc9cndnskqdhnsxm-shepherd-file-system--tmp.go.drv
   /gnu/store/a5nb6vhs86g76ww5bmpphi398pggnga7-shepherd-file-system--tmp.scm.drv
   /gnu/store/nhi2s5zcyd5frjqag4jnp7fvhzbx6kd6-activate.scm.drv
The following profile hooks will be built:
   /gnu/store/4j2hrilvx5rfycqs1sccd39jy5yjhxy0-xdg-mime-database.drv
   /gnu/store/4rdg1bi6xxn85zq5cjqc3bdfksx14913-info-dir.drv
   /gnu/store/c59p1krnc51qlbr5vnxjx6bg3rrixca4-glib-schemas.drv
   /gnu/store/drf1pi6m91b6qra3p6hrknfzh7gcd51z-fonts-dir.drv
   /gnu/store/fwal8vsrjzpbwljnihvv66k9ikkfx2ma-xdg-desktop-database.drv
   /gnu/store/jzp4ps09ipy0kcc6ybx5967vi3xlq7aa-manual-database.drv
   /gnu/store/m7yl2xvk07l37l19wp2hfbvhzl3583ji-ca-certificate-bundle.drv
building /gnu/store/45334hsl3hs5yx13x7qz1b1sc83vkbzj-module-import-compiled.drv...
process 21756 acquired build slot '/var/guix/offload/10.42.0.199:22/0'
load on machine '10.42.0.199' is 0.0 (normalized: 0.0)
building /gnu/store/07kxlp8a99mvx2zidnlz6rm4ak2az8d0-guix-1.0.1-13.50299ad+.drv...
^Cmake: *** [Makefile:5711: check-system] Interrupt

The diff between those two looks like:

Toggle snippet (50 lines)--- /tmp/ediffD0RQCf	2020-02-16 12:10:53.643411325 -0500
+++ /tmp/ediffKmWrdZ	2020-02-16 12:10:53.647411377 -0500
@@ -1,15 +1,18 @@
-   /gnu/store/3g9bawa6c876lml0318g2knqdcv2ahv9-btrfs-root-on-subvolume-os.drv
-   /gnu/store/vswbrzp8ab5wz0fyzrj9j72mqcbd6c4a-installation.drv
-   /gnu/store/nf1wd5bgminylzvwq4mjs5293hpr1mmi-disk-image.drv
+   /gnu/store/l4z1363klr8a58qmm3pjzy2y5bxsr6bd-btrfs-root-on-subvolume-os.drv
+   /gnu/store/zlg6gxml4awbkbs2s6n1b0z3ysrqvy0h-installation.drv
+   /gnu/store/89vdpn33f47y0510im604snwfmbpswmm-disk-image.drv
    /gnu/store/45334hsl3hs5yx13x7qz1b1sc83vkbzj-module-import-compiled.drv
-   /gnu/store/dmi03i3jv9wwkw10fm57g7fq97kwlscf-module-import.drv
    /gnu/store/4cp5pqyn554rdmv9l4vfhnw7lgasv18q-raw-initrd.drv
    /gnu/store/vx147spsmrxkiy0qcmvxn341fpz4fmrv-init.drv
    /gnu/store/d6ri0lh953vaksiz5h8jx681h3i35vvm-module-import-compiled.drv
-   /gnu/store/b0gv94llsbvrlzkxs3naz7crcv1glxy1-grub.cfg.drv
+   /gnu/store/58912rdjjk4jdv767svb2axp0llmn8ch-linux-vm-loader.drv
+   /gnu/store/hvcm9abi0br38plwpngqlvw1l0060fni-builder-in-linux-vm.drv
+   /gnu/store/14bw0x4dq0g262vk4pjh3dlha6x3gz0s-module-import-compiled.drv
+   /gnu/store/nbkfy5wjpia5y7hj1p9j5hmbv42j4zp5-module-import.drv
+   /gnu/store/shc8hzzqi40pdpip0773i2g9x2apnn0h-grub.cfg.drv
    /gnu/store/gcslrfymsqsp6kqckla41zkyaa7fy9in-raw-initrd.drv
    /gnu/store/m8bbfcdn1irn7a41syy9arl1yz34hbln-init.drv
-   /gnu/store/hbfgqnralir92slqr9hhfw8cxqrlgv06-system.drv
+   /gnu/store/zifmvsbiw6x4gs15k29l1nvyzxkszafw-system.drv
    /gnu/store/0v4aq6yprk8az183rpdjx90jnxw3w838-parameters.drv
    /gnu/store/8daaamb92bbkpdwz96a19r81f7vnk40k-gc-roots.drv
    /gnu/store/f48dwgib52ciaq1nh06nhks6g7gvy7bx-system.drv
@@ -44,56 +47,24 @@
    /gnu/store/z0ik75q9nshaik97jsigd69vriyjijgp-shepherd.conf.drv
    /gnu/store/6pxsd37k5nl5viqadd57n0hah53nzcp7-shepherd-user-homes.go.drv
    /gnu/store/y9rj23r1xmyb5b4fwrh75pb5yr61clvp-shepherd-user-homes.scm.drv
-   /gnu/store/h0xbdk4hw1i4byy8xnrnaqmlwnfgrswz-profile.drv
-   /gnu/store/sxdgmvfjakapmbyqmxjgvarsaaxc2mw7-profile.drv
-   /gnu/store/hxcadzr771dzl99kylz64d9qz2qai8wd-guix-000000000.drv
-   /gnu/store/95lcx1xx41p4vcwqlwli7phny6hpp8b0-guix-manual.drv
-   /gnu/store/9yz0spp1h158zdh8wmg2r7sspnkm0qfc-guix-translated-texinfo.drv
-   /gnu/store/g6ggidnpaahws37qgx15mvkhfll0ns06-guix-daemon.drv
-   /gnu/store/jpb2jfk00q5nh5fg3k3cf6hcyw49zz84-guix-command.drv
-   /gnu/store/14zcma6hkqszaab37hda736i24r41rv4-guix-module-union.drv
-   /gnu/store/sfzd86425cl744lr58yh877i7z5qbd68-guix-000000000-modules.drv
-   /gnu/store/71a40hzi2d2wdn1vjl5amfq00caqlh3m-guix-system-tests-modules.drv
-   /gnu/store/kc3zg69yabaglvllj4l7hzgs72v9mk26-guix-system-tests.drv
-   /gnu/store/5zhgzvnyswm8g50aha95rvdx6dpmw329-guix-packages-base.drv
-   /gnu/store/d7df7qw5allqcm2qn75dn60wz1dg50cq-guix-system.drv
-   /gnu/store/ily3pqqvfkjz5622fxz6mxg31jbiqb21-guix-packages.drv
-   /gnu/store/xgpy1v810xi76w8ivpfi7gg5h7vg0ska-guix-cli.drv
-   /gnu/store/7ih7cmbhz1p7a4m7yrgy0s2jawqc666n-guix-packages-base-modules.drv
-   /gnu/store/j15kzs5m07pq3nmgqr5rr47cryq8arfs-guix-system-modules.drv
-   /gnu/store/rhk9cknm5wbk30slzzrr52app8mfn4a2-gu

This message was truncated. Download the full message here.

Maxim Cournoyer wrote on 18 Feb 2020 22:27

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)(address . 37305@debbugs.gnu.org)

Message-ID:87sgj7wpqa.fsf@gmail.com

Hello Ludovic!

I have more benchmark results. Not to be compared with the previous

results, as these were executed on a much faster machine (24 cores vs 4

cores), with an SSD instead of a rotative disk.

Commands used:

current-guix/pre-built:

Toggle snippet (15 lines)time ./pre-inst-env guix system build -e \
'(begin
  (use-modules
   (gnu packages package-management))
  (parameterize
      ((current-guix-package
	(current-guix/pre-built)))
    ((@@
      (gnu tests install)
      operating-system-with-current-guix)
     (@@
      (gnu tests install)
      %btrfs-root-os))))'

current-guix (old fashionned):

Toggle snippet (15 lines)time ./pre-inst-env guix system build -e \
'(begin
  (use-modules
   (gnu packages package-management))
  (parameterize
      ((current-guix-package
	(current-guix)))
    ((@@
      (gnu tests install)
      operating-system-with-current-guix)
     (@@
      (gnu tests install)
      %btrfs-root-os))))'

new current-guix (guix self):

Toggle snippet (27 lines)time ./pre-inst-env guix system build -e \
'(begin
   (use-modules ((gnu ci) #:select (channel-instance->package))
		(guix monads)
		(guix channels)
		(guix store)
		((guix status) #:select (with-status-verbosity))
		((guix git-download) #:select (git-predicate))
		(guix utils)
		(gnu packages package-management))
   (with-store store
     (with-status-verbosity 2
       (run-with-store store
	 (mlet* %store-monad
	     ((source-dir -> "/home/mcournoyer/src/guix")
	      (source (interned-file source-dir
				     "guix-source"
				     #:recursive? #t
				     #:select? (or (git-predicate source-dir)
						   (const #t))))
	      (instance -> (checkout->channel-instance source))
	      (new-guix-current -> (channel-instance->package instance)))
	   (return  (parameterize ((current-guix-package new-guix-current))
		      ((@@ (gnu tests install) operating-system-with-current-guix)
		       (@@ (gnu tests install) %btrfs-root-os)))))))))'

Methodology:
1. Break at each commit of the current change set, run make, and run the three
   above commands in the order they appear.
2. Record the "real" component of the time results of each
   commit/command in the table below.

| Commit                                                       | current-guix/pre-built | old current-guix | new current-guix |
|--------------------------------------------------------------+------------------------+------------------+------------------|
| linux-boot: Ensure volatile root is mounted read-only.       | 2m2.245s               | 13m39.669s       | 3m47.066s        |
| file-systems: Add a 'file-system-device->string' procedure.  | 2m12.576s              | 13m42.906s       | 3m43.212s        |
| gnu: linux-boot: Refactor boot-system.                       | 1m53.928s              | 13m48.825s       | 9m53.133s        |
| file-systems: Represent the file system options as an alist. | 2m12.530s              | ""               | 10m3.523s        |
| gnu: linux-boot: Honor the "--root-options" kernel argument. | ""                     | ""               | 11m10.448s       |
| gnu: linux-boot: Filter out file system independent options. | ""                     | ""               | 5m4.997s         |
| bootloader: grub: Allow booting from a Btrfs subvolume.      | 2m13.375s              | 13m44.671s       | 3m38.815s        |
|--------------------------------------------------------------+------------------------+------------------+------------------|

This machine appears to be 5 times faster (at this task) than the
previous one used, to give an idea of the time someone would have to
wait just to build the operating system to test.  An x200 would be even
slightly slower that than.

One of the things that seems costly is rebuilding the documentation.

Maxim

Maxim Cournoyer wrote on 19 Feb 2020 03:52

Re: [bug#37305] Allow booting from a Btrfs subvolume [review part 2]

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)(address . 37305@debbugs.gnu.org)

Message-ID:875zg2xtsb.fsf@gmail.com

Hello Ludovic,

Here comes part 2!

Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (37 lines)>> From 6cf2ece21683e98544f8f46675aef58d5a7231fd Mon Sep 17 00:00:00 2001
>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>> Date: Sun, 14 Jul 2019 20:50:23 +0900
>> Subject: [PATCH 8/9] bootloader: grub: Allow booting from a Btrfs subvolume.
>>
>> * gnu/bootloader/grub.scm (grub-configuration-file) [btrfs-subvolume-path]:
>> New parameter.  When it is defined, prepend its value to the kernel and
>> initrd file paths.
>> * gnu/bootloader/depthcharge.scm (depthcharge-configuration-file): Adapt.
>> * gnu/bootloader/extlinux.scm (extlinux-configuration-file): Likewise.
>> * gnu/system/file-systems.scm (btrfs-subvolume?)
>> (btrfs-store-subvolume-path): New procedures.
>> * gnu/system.scm (operating-system-bootcfg): Specify the Btrfs subvolume path
>> of the GNU store to the `operating-system-bootcfg' procedure, using the new
>> BTRFS-SUBVOLUME-PATH argument.
>> * doc/guix.texi (File Systems): Add a Btrfs subsection to document the use of
>> subvolumes.  Document the new `properties' field of the `<file-system>'
>> record.
>> * gnu/tests/install.scm: Add test "btrfs-root-on-subvolume-os".
>
> Neat!
>
>>  (define* (grub-configuration-file config entries
>>                                    #:key
>>                                    (system (%current-system))
>> -                                  (old-entries '()))
>> +                                  (old-entries '())
>> +                                  btrfs-subvolume-path)
>>    "Return the GRUB configuration file corresponding to CONFIG, a
>>  <bootloader-configuration> object, and where the store is available at
>>  STORE-FS, a <file-system> object.  OLD-ENTRIES is taken to be a list of menu
>> -entries corresponding to old generations of the system."
>> +entries corresponding to old generations of the system.  BTRFS-SUBVOLUME-PATH
>> +may be used to specify on which subvolume a Btrfs root file system resides."
>
> (Nitpick: s/path/file name/ :-))

I stand corrected! From '(Standards)GNU Manuals':

Please do not use the term "pathname" that is used in Unix

documentation; use "file name" (two words) instead. We use the term

"path" only for search paths, which are lists of directory names.

Toggle quote (10 lines)> It’s a bit problematic that (1) GRUB needs explicit Btrfs support, and
> (2) other bootloaders will silently ignore the option, due to
> #:allow-other-keys.
>
> I don’t have a better idea, but it’d be great if Btrfs support could be
> made bootloader-independent, and if it could be somewhat
> not-too-btrfs-specific, if that is possible at all.
>
> Thoughts?

I have no idea how Btrfs subvolumes are handled (if at all) on U-Boot or

other bootloaders than GRUB. All I know is that for GRUB they need to

handle subvolumes in a special manner in their own grub-mkconfig tool

(which we bypass).

Also, I'm afraid subvolumes are very Btrfs specific :-). It doesn't

exist in traditional file systems like EXT4. I think ZFS must have

something similar, though.

Toggle quote (28 lines)>> +  (properties       file-system-properties        ; list of name-value pairs
>> +                    (default '()))
>>    (location         file-system-location
>>                      (default (current-source-location))
>>                      (innate)))
>> @@ -582,4 +589,48 @@ system has the given TYPE."
>>      (or (string-prefix-ci? "x-" option-name)
>>          (member option-name %file-system-independent-mount-options))))
>>
>> +(define (btrfs-subvolume? fs)
>> +  "Predicate to check if FS, a file-system object, is a Btrfs subvolume."
>> +  (and-let* ((btrfs-file-system? (string= "btrfs" (file-system-type fs)))
>> +             (option-keys (map (match-lambda
>> +                                 ((key . value) key)
>> +                                 (key key))
>> +                               (file-system-options fs))))
>> +    (find (cut string-prefix? "subvol" <>) option-keys)))
>
> I wonder if we can avoid special support in the <file-system> API for
> Btrfs.
>
>> +              (error "The store is on a Btrfs subvolume, but the \
>> +subvolume name is unknown.
>> +Hint: Define the \"btrfs-subvolume-path\" file system property or
>> +use the \"subvol\" Btrfs file system option."))))

> Rather use ‘raise’ with ‘&message’ and ‘&fix-hint’ conditions.

I tried this, but importing (guix utils) to acces &fix-hint caused the init

RAM disk to fail mysteriously:

Toggle snippet (63 lines)[    0.614503] Run /init as init process
GC Warning: pthread_getattr_np or pthread_attr_getstack failed for main thread
GC Warning: Couldn't read /proc/stat
Backtrace:
In ice-9/boot-9.scm:
   222:29 19 (map1 (((ice-9 match)) ((rnrs bytevectors)) ((srfi 
   222:29 18 (map1 (((rnrs bytevectors)) ((srfi srfi-1)) ((srfi 
   222:29 17 (map1 (((srfi srfi-1)) ((srfi srfi-2)) ((srfi #)) (#) 
   222:29 16 (map1 (((srfi srfi-2)) ((srfi srfi-9)) ((srfi #)) (#) 
   222:29 15 (map1 (((srfi srfi-9)) ((srfi srfi-26)) ((srfi #)) (#) 
   222:29 14 (map1 (((srfi srfi-26)) ((srfi srfi-35)) ((srfi # #)) 
   222:29 13 (map1 (((srfi srfi-35)) ((srfi srfi-9 gnu)) ((guix 
))[    0.657578] random: fast init done
   222:29 12 (map1 (((srfi srfi-9 gnu)) ((guix records)) ((guix 
   222:29 11 (map1 (((guix records)) ((guix utils)) ((gnu system 
))))
   222:17 10 (map1 (((guix utils)) ((gnu system uuid))))
  2800:17  9 (resolve-interface (guix utils) #:select _ #:hide _ # _ 
In ice-9/threads.scm:
    390:8  8 (_ _)
In ice-9/boot-9.scm:
  2726:13  7 (_)
In ice-9/threads.scm:
    390:8  6 (_ _)
In ice-9/boot-9.scm:
  2994:20  5 (_)
   2312:4  4 (save-module-excursion #<procedure 7f455b6d48d0 at ice-
  3014:26  3 (_)
In unknown file:
           2 (primitive-load-path "guix/utils" #<procedure 7f455b8fd
In guix/utils.scm:
   508:24  1 (_)
In unknown file:
           0 (dynamic-func "strverscmp" #<dynamic-object #f>)
ERROR: In procedure dynamic-func:
In procedure dynamic-pointer: Symbol not found: strverscmp
[    0.697002] Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000100
[    0.697894] CPU: 0 PID: 1 Comm: init Not tainted 5.4.18-gnu #1
[    0.698592] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba5276e321-prebuilt.qemu.org 04/01/2014
[    0.699938] Call Trace:
[    0.700240]  dump_stack+0x6d/0x8d
[    0.700640]  panic+0x10b/0x2f4
[    0.701010]  do_exit+0x7e2/0xb80
[    0.701398]  ? wake_up_state+0x1f/0x30
[    0.701847]  ? signal_wake_up_state+0x24/0x40
[    0.702374]  do_group_exit+0x44/0xa0
[    0.702805]  __x64_sys_exit_group+0x1c/0x20
[    0.703306]  do_syscall_64+0x5a/0x190
[    0.703748]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[    0.704347] RIP: 0033:0x5df9c6
[    0.704716] Code: 00 00 00 be 3c 00 00 00 eb 19 66 2e 0f 1f 84 00 00 00 00 00 89 d7 89 f0 0f 05 48 3d 00 f0 ff ff 77 22 f4 89 d7 44 89 c0 0f 05 <48> 3d 00 f0 ff ff 76 e2 f7 d8 64 41 89 01 eb da 66 2e 0f 1f 84 00
[    0.706919] RSP: 002b:00007ffd9fa65228 EFLAGS: 00000246 ORIG_RAX: 00000000000000e7
[    0.707808] RAX: ffffffffffffffda RBX: 00000000007c51f0 RCX: 00000000005df9c6
[    0.708643] RDX: 0000000000000001 RSI: 000000000000003c RDI: 0000000000000001
[    0.709482] RBP: 0000000000000001 R08: 00000000000000e7 R09: ffffffffffffffb0
[    0.710328] R10: 0000000000436340 R11: 0000000000000246 R12: 00000000007c51f0
[    0.711166] R13: 0000000000000001 R14: 0000000000000000 R15: 0000000000000003
[    0.712026] Kernel Offset: 0x13000000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff)
[    0.713288] Rebooting in 1 seconds..
file-size: /gnu/store/zfi66vny0h10d180xajgm4pq2vnvmc2z-nss-certs-3.49.1/etc/ssl/certs/NetLock_Arany_=Class_Gold=_F??tan??s??tv??ny:2.6.73.65.44.228.0.16.pem: No such file or directory
file-size: /gnu/store/04xpkgf9zlhcngyr6gnhl4rb8g6v6i1i-profile/etc/ssl/certs/NetLock_Arany_=Class_Gold=_F??tan??s??tv??ny:2.6.73.65.44.228.0.16.pem: No such file or directory

The exception I had refactored to use with &fix-hint looked like:

Toggle snippet (10 lines)

(raise (condition

(&message

(message "The store is on a Btrfs subvolume, \

but the subvolume name is unknown."))

(&fix-hint

(hint "Define the \"btrfs-subvolume-file-name\" \

file system property or use the \"subvol\" Btrfs file system

option."))))

So, I ended up using just a &message condition with the hint embedded in it.

The attached patch (v3) incorporate the agreed changes so far.

The btrfs-root-on-subvolume-os test still passes.

Thank you for your patience,

Maxim

Attachment: 0001-gnu-tests-Reduce-the-time-required-to-run-the-system.patch

From c6e00c0eb8b2ed0c758e9c1be28c6e93f4795a64 Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 12:57:29 -0500
Subject: [PATCH 2/8] linux-boot: Ensure volatile root is mounted read-only.

* gnu/build/linux-boot.scm (mount-root-file-system): Ensure MS_RDONLY is
present among the root file system flags when VOLATILE-ROOT? is #t.
---
 gnu/build/linux-boot.scm | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Toggle diff (21 lines)diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index 3d40a7d05d..4fb711b8f2 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -362,12 +362,12 @@ the last argument of `mknod'."
   "Mount the root file system of type TYPE at device ROOT. If VOLATILE-ROOT? is
 true, mount ROOT read-only and make it an overlay with a writable tmpfs using
 the kernel built-in overlayfs. FLAGS and OPTIONS indicates the options to use
-to mount ROOT."
+to mount ROOT, and behave the same as for the `mount' procedure."
 
   (if volatile-root?
       (begin
         (mkdir-p "/real-root")
-        (mount root "/real-root" type MS_RDONLY options)
+        (mount root "/real-root" type (logior MS_RDONLY flags) options)
         (mkdir-p "/rw-root")
         (mount "none" "/rw-root" "tmpfs")
 
-- 
2.25.0

From c367fc7efeb8ff15c22a98f32098bbcdbf1457b6 Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 23:56:45 -0500
Subject: [PATCH 3/8] file-systems: Add a 'file-system-device->string'
 procedure.

* gnu/system/file-systems.scm (file-system-device->string): New procedure.
* gnu/system.scm (bootable-kernel-arguments): Use it.
* gnu/system/vm.scm (operating-system-uuid): Likewise.
* guix/scripts/system.scm (display-system-generation): Likewise.
---
 gnu/system.scm              | 15 +++++----------
 gnu/system/file-systems.scm | 15 +++++++++++++++
 gnu/system/vm.scm           |  8 +-------
 guix/scripts/system.scm     |  7 +------
 4 files changed, 22 insertions(+), 23 deletions(-)

Toggle diff (98 lines)diff --git a/gnu/system.scm b/gnu/system.scm
index 01baa248a2..2e6d03272d 100644
--- a/gnu/system.scm
+++ b/gnu/system.scm
@@ -142,16 +142,11 @@
 (define (bootable-kernel-arguments system root-device)
   "Return a list of kernel arguments (gexps) to boot SYSTEM from ROOT-DEVICE."
   (list (string-append "--root="
-                       (cond ((uuid? root-device)
-
-                              ;; Note: Always use the DCE format because that's
-                              ;; what (gnu build linux-boot) expects for the
-                              ;; '--root' kernel command-line option.
-                              (uuid->string (uuid-bytevector root-device)
-                                            'dce))
-                             ((file-system-label? root-device)
-                              (file-system-label->string root-device))
-                             (else root-device)))
+                       ;; Note: Always use the DCE format because that's what
+                       ;; (gnu build linux-boot) expects for the '--root'
+                       ;; kernel command-line option.
+                       (file-system-device->string root-device
+                                                   #:uuid-type 'dce))
         #~(string-append "--system=" #$system)
         #~(string-append "--load=" #$system "/boot")))
 
diff --git a/gnu/system/file-systems.scm b/gnu/system/file-systems.scm
index d47a514b66..3b599efa8e 100644
--- a/gnu/system/file-systems.scm
+++ b/gnu/system/file-systems.scm
@@ -30,6 +30,7 @@
   #:export (file-system
             file-system?
             file-system-device
+            file-system-device->string
             file-system-title                     ;deprecated
             file-system-mount-point
             file-system-type
@@ -235,6 +236,20 @@ where both FILE1 and FILE2 are absolute file name.  For example:
               (()
                #f)))))))
 
+(define* (file-system-device->string device #:key uuid-type)
+  "Return the string representations of the DEVICE field of a <file-system>
+record.  When the device is a UUID, its representation is chosen depending on
+UUID-TYPE, a symbol such as 'dce or 'iso9660."
+  (match device
+    ((? file-system-label?)
+     (file-system-label->string device))
+    ((? uuid?)
+     (if uuid-type
+         (uuid->string (uuid-bytevector device) uuid-type)
+         (uuid->string device)))
+    ((? string?)
+     device)))
+
 (define (file-system-needed-for-boot? fs)
   "Return true if FS has the 'needed-for-boot?' flag set, or if it holds the
 store--e.g., if FS is the root file system."
diff --git a/gnu/system/vm.scm b/gnu/system/vm.scm
index 81b2e06ba2..03a511cdde 100644
--- a/gnu/system/vm.scm
+++ b/gnu/system/vm.scm
@@ -609,13 +609,7 @@ TYPE (one of 'iso9660 or 'dce).  Return a UUID object."
     (let ((device (file-system-device fs)))
       (list (file-system-mount-point fs)
             (file-system-type fs)
-            (cond ((file-system-label? device)
-                   (file-system-label->string device))
-                  ((uuid? device)
-                   (uuid->string device))
-                  ((string? device)
-                   device)
-                  (else #f))
+            (file-system-device->string device)
             (file-system-options fs))))
 
   (if (eq? type 'iso9660)
diff --git a/guix/scripts/system.scm b/guix/scripts/system.scm
index e69a3b6c97..b0386a1392 100644
--- a/guix/scripts/system.scm
+++ b/guix/scripts/system.scm
@@ -517,12 +517,7 @@ list of services."
               (cond ((uuid? root-device) 0)
                     ((file-system-label? root-device) 1)
                     (else 2))
-              (cond ((uuid? root-device)
-                     (uuid->string root-device))
-                    ((file-system-label? root-device)
-                     (file-system-label->string root-device))
-                    (else
-                     root-device)))
+              (file-system-device->string root-device))
 
       (format #t (G_ "  kernel: ~a~%") kernel)
 
-- 
2.25.0

Attachment: 0004-linux-boot-Refactor-boot-system.patch

Attachment: 0005-file-systems-Represent-the-file-system-options-as-an.patch

From 9e988968566f004f887080b02712043520d9327e Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 14:14:36 -0500
Subject: [PATCH 6/8] linux-boot: Honor the "--root-options" kernel argument.

* gnu/build/linux-boot.scm (boot-system): Parse the "--root-options" kernel
argument, and use it when calling `mount-root-file-system'.  Update doc.
* doc/guix.texi (Initial RAM Disk): Document the use of the "--root-options"
argument.
---
 doc/guix.texi            | 10 ++++++++++
 gnu/build/linux-boot.scm | 16 +++++++++++-----
 2 files changed, 21 insertions(+), 5 deletions(-)

Toggle diff (71 lines)diff --git a/doc/guix.texi b/doc/guix.texi
index 5d526b1aee..d6bfbd7b55 100644
--- a/doc/guix.texi
+++ b/doc/guix.texi
@@ -25935,6 +25935,16 @@ Instruct the initial RAM disk as well as the @command{modprobe} command
 must be a comma-separated list of module names---e.g.,
 @code{usbkbd,9pnet}.
 
+@item --root-options=@var{options}@dots{}
+@cindex mount options for the root file system, passed to initrd
+@cindex rootflags, initrd
+@cindex root-options, initrd
+This argument allows passing one or multiple file system specific mount
+options used by the initrd to mount the root file system.  @var{options}
+must be a comma-separated list of option names or option-value pairs.
+When unspecified, the value of the options field of the root file system
+of the operating system declaration is used.
+
 @item --repl
 Start a read-eval-print loop (REPL) from the initial RAM disk before it
 tries to load kernel modules and to mount the root file system.  Our
diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index c3229fa292..063daa4ca4 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -2,6 +2,7 @@
 ;;; Copyright © 2013, 2014, 2015, 2016, 2017, 2018, 2019 Ludovic Courtès <ludo@gnu.org>
 ;;; Copyright © 2017 Mathieu Othacehe <m.othacehe@gmail.com>
 ;;; Copyright © 2019 Guillaume Le Vaillant <glv@posteo.net>
+;;; Copyright © 2020 Maxim Cournoyer <maxim.cournoyer@gmail.com>
 ;;;
 ;;; This file is part of GNU Guix.
 ;;;
@@ -452,7 +453,8 @@ LINUX-MODULE-DIRECTORY, then installing KEYMAP-FILE with 'loadkeys' (if
 KEYMAP-FILE is true), then setting up QEMU guest networking if
 QEMU-GUEST-NETWORKING? is true, calling PRE-MOUNT, mounting the file systems
 specified in MOUNTS, and finally booting into the new root if any.  The initrd
-supports kernel command-line options '--load', '--root', and '--repl'.
+supports kernel command-line options '--load', '--root', '--root-options' and
+'--repl'.
 
 Mount the root file system, specified by the '--root' command-line argument,
 if any.
@@ -493,6 +495,7 @@ upon error."
              (root-fs-options (if root-fs
                                   (file-system-options root-fs)
                                   '()))
+
              ;; --root takes precedence over the 'device' field of the root
              ;; <file-system> record.
              ;; TODO: Add options for the root file system type and the root
@@ -500,10 +503,13 @@ upon error."
              (root-device (or (and=> (find-long-option "--root" args)
                                      device-string->file-system-device)
                               root-fs-device))
-             (root-options (if (null? root-fs-options)
-                               #f
-                               (file-system-options->string
-                                root-fs-options))))
+             ;; --root-options takes precedence over the 'options' field of the
+             ;; root <file-system> record.
+             (root-options (or (find-long-option "--root-options" args)
+                               (if (null? root-fs-options)
+                                   #f
+                                   (file-system-options->string
+                                    root-fs-options)))))
 
         (when (member "--repl" args)
           (start-repl))
-- 
2.25.0

From 2cc5b27defbd2b9fd75c70482ac2ecf045b880e3 Mon Sep 17 00:00:00 2001
From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Date: Tue, 11 Feb 2020 14:27:19 -0500
Subject: [PATCH 7/8] linux-boot: Filter out file system independent options.

This fixes an issue where options such as "defaults", which are understood by
the command line program "mount", are not understood by the system call of the
same name, which is used in the initial RAM disk.

* gnu/system/file-systems.scm (%file-system-independent-mount-options): New variable.
(file-system-independent-mount-option?): New predicate.
* gnu/build/linux-boot.scm (boot-system): Use the above predicate to filter
out system independent mount options.
---
 gnu/build/linux-boot.scm    |  3 ++-
 gnu/system/file-systems.scm | 17 +++++++++++++++++
 2 files changed, 19 insertions(+), 1 deletion(-)

Toggle diff (49 lines)diff --git a/gnu/build/linux-boot.scm b/gnu/build/linux-boot.scm
index 063daa4ca4..f77eeecbe3 100644
--- a/gnu/build/linux-boot.scm
+++ b/gnu/build/linux-boot.scm
@@ -493,7 +493,8 @@ upon error."
                              (or (and=> root-fs file-system-flags)
                                  '())))
              (root-fs-options (if root-fs
-                                  (file-system-options root-fs)
+                                  (remove file-system-independent-mount-option?
+                                          (file-system-options root-fs))
                                   '()))
 
              ;; --root takes precedence over the 'device' field of the root
diff --git a/gnu/system/file-systems.scm b/gnu/system/file-systems.scm
index c205feae70..4f0c5ad99e 100644
--- a/gnu/system/file-systems.scm
+++ b/gnu/system/file-systems.scm
@@ -46,6 +46,7 @@
             file-system-location
 
             file-system-type-predicate
+            file-system-independent-mount-option?
 
             file-system-label
             file-system-label?
@@ -567,4 +568,20 @@ system has the given TYPE."
   (lambda (fs)
     (string=? (file-system-type fs) type)))
 
+(define %file-system-independent-mount-options
+  ;; Taken from 'man 8 mount'.
+  '("async" "atime" "auto" "noatime" "noauto" "context" "defaults" "dev" "nodev"
+    "diratime" "nodiratime" "dirsync" "exec" "noexec" "group" "iversion"
+    "noiversion" "mand" "nomand" "_netdev" "nofail" "relatime" "norelatime"
+    "strictatime" "nostrictatime" "lazytime" "nolazytime" "suid" "nosuid"
+    "silent" "loud" "owner" "remount" "ro" "rw" "sync" "user" "nouser" "users"))
+
+(define (file-system-independent-mount-option? option)
+  "Predicate to check if a <file-system> option is file system independent."
+  (let ((option-name (if (pair? option)
+                         (car option)
+                         option)))
+    (or (string-prefix-ci? "x-" option-name)
+        (member option-name %file-system-independent-mount-options))))
+
 ;;; file-systems.scm ends here
-- 
2.25.0

Attachment: 0008-bootloader-grub-Allow-booting-from-a-Btrfs-subvolume.patch

Ludovic Courtès wrote on 20 Feb 2020 10:55

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)(address . 37305@debbugs.gnu.org)

Message-ID:87lfoxbn2e.fsf@gnu.org

Hi Maxim,

Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:

Toggle quote (10 lines)>>> +              (error "The store is on a Btrfs subvolume, but the \
>>> +subvolume name is unknown.
>>> +Hint: Define the \"btrfs-subvolume-path\" file system property or
>>> +use the \"subvol\" Btrfs file system option."))))
>
>> Rather use ‘raise’ with ‘&message’ and ‘&fix-hint’ conditions.
>
> I tried this, but importing (guix utils) to acces &fix-hint caused the init
> RAM disk to fail mysteriously:

Oh, my bad. We should move ‘&fix-hint’ to (guix diagnostics)

eventually.

In the meantime, I’d say just raise a ‘&message’ and leave the hint as a

comment (it’s not supposed to be a user-facing interface). Or maybe you

could define a specific error condition type for this?

Thanks,

Ludo’.

PS: I’ll comment on the other bits ASAP!

Maxim Cournoyer wrote on 24 Feb 2020 15:23

Re: [bug#37305] [PATCH V3] Allow booting from a Btrfs subvolume.

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)(address . 37305@debbugs.gnu.org)

Message-ID:871rqkf4ix.fsf_-_@gmail.com

I've added a `normalize-file' procedure to the attached patch, which

simplify producing correct store references in the GRUB configuration

file generated.

This fixed the GRUB video mode issues reported in a another issue.

Thanks!

Maxim

Attachment: 0001-bootloader-grub-Allow-booting-from-a-Btrfs-subvolume.patch

Ludovic Courtès wrote on 24 Feb 2020 17:02

Re: [bug#37305] [PATCH V2] Allow booting from a Btrfs subvolume.

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)(address . 37305@debbugs.gnu.org)

Message-ID:87imjwkm6u.fsf@gnu.org

Hi Maxim,

Resuming review of this series… Sorry for the delay!

Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:

Toggle quote (17 lines)>>> From 97d8a635eba34c7cf0708e99bf77ef9bad1344bf Mon Sep 17 00:00:00 2001
>>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>>> Date: Tue, 11 Feb 2020 12:57:29 -0500
>>> Subject: [PATCH 2/9] gnu: linux-boot: Ensure volatile root is mounted
>>>  read-only.
>>>
>>> * gnu/build/linux-boot.scm (mount-root-file-system): Ensure MS_RDONLY is
>>> present among the root file system flags when VOLATILE-ROOT? is #t.
>>
>> (You can drop the “gnu:” prefix.)
>
> Done.
>
> I never know before looking at past logs (and then sometimes it's a
> mixed bag).  Is there any mechanical process for selecting the right
> commit prefix? :-)

“gnu:” is for changes to (gnu packages). The idea is that the prefix

should reflect what subsystem the commit is modifying. But yeah,

looking at ‘git log’ can be inspiring. :-)

Toggle quote (17 lines)>>>  @item --root=@var{root}
>>> -Mount @var{root} as the root file system.  @var{root} can be a
>>> -device name like @code{/dev/sda1}, a file system label, or a file system
>>> -UUID.
>>> +Mount @var{root} as the root file system.  @var{root} can be a device
>>> +name like @code{/dev/sda1}, a file system label, or a file system UUID.
>>> +When unspecified, the device name from the root file system of the
>>> +operating system declaration is used.
>>
>> Oh!  Does it always work?  That makes me wonder why we’ve been carrying
>> ‘--root’ and I’m not sure if I’m forgetting a good reason to do it that
>> way.
>
> If the documentation is accurate, it should :-), given that --root gets
> written as a string to the GRUB configuration file, and that the doc
> says it's possible to give it as a device name, label or UUID.

Yes, ‘--root’ can resolve labels and UUIDs; my question was more about

why we have it in the first place.

Toggle quote (9 lines)

> About why providing options such as --root or --root-options in the

> first place; I pondered about this as well, especially after making the

> file systems from operating system able to be mounted with all their

> (file system independent -- more on that later) options. A reason I

> came up with was that it allows to experiment at the GRUB command line

> and change the root device, or perhaps the root options. One use case

> would be debugging the right options to pass to a file system driver in

> case of a mistake in the operating system declaration.

Yes, that makes sense. It’s certainly useful to have ‘--root’ at least

as an option.

Toggle quote (23 lines)>> The main issue I see with this change is that mount(2) takes raw strings
>> for the options.  There’s a convention to have those strings look like
>> “KEY1=VALUE1,KEY2=VALUE2”, but it’s just a convention.
>>
>> As a rule of thumb, I’d rather have our interface be as close as
>> possible to the actual mount(2) interface, which means taking strings.
>>
>> Now, we can surely add helper procedures to parse options that follow
>> the above conventions.
>>
>> WDYT?
>
> To me, it's an implementation detail that I'd rather abstract away (or
> make optional, like in this patch).  Just like we provide a higher level
> configuration for services instead of requiring the user to input the
> configuration in the native format of the tool (or allowing for both).
> The idea for this format was taken from a discussion here:
> http://issues.guix.info/issue/33517#3.
>
> Are we really targeting mount(2)?  The commit
> 9d3053819dfd834a1c29a03427c41d8524b8a7d5 (which you co-authored :-))
> mentions 'man 8 mount' for the file system options.

Right, mount(8) documents file system options that can be passed to

mount(2).

What does it mean to target mount(8) vs. mount(2)? To me, mount(8) is a

CLI to mount(2) that provides additional features to make the CLI more

convenient: the “defaults” option, a way to pass mount(2) flags as

options (like “ro”, “remount”, “bind”), /etc/fstab handling, etc.

Guix System handles /etc/fstab differently and “defaults” makes little

sense in our API (one can just use leave the default value of the

‘options’ field.)

I think mount(8) is actually a good illustration of what not to do. It

ends up mixing things that are separate in the mount(2) API, and that

doesn’t improve clarity and future-proof-ness (what if a file system has

a “bind” option, etc.).

But again, I think the helper procedures that you propose to move back

and forth between the string and the alist representations are very

welcome. I just wouldn’t hard-code that directly in our API.

WDYT?

Toggle quote (17 lines)>>> From 67135c925b07f2e077b4cd852e07178691a10164 Mon Sep 17 00:00:00 2001
>>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>>> Date: Tue, 11 Feb 2020 14:14:36 -0500
>>> Subject: [PATCH 6/9] gnu: linux-boot: Honor the "--root-options" kernel
>>>  argument.
>>>
>>> * gnu/build/linux-boot.scm (boot-system): Parse the "--root-options" kernel
>>> argument, and use it when calling `mount-root-file-system'.  Update doc.
>>> * doc/guix.texi (Initial RAM Disk): Document the use of the "--root-options"
>>> argument.
>>
>> Hmm do we really need this extra option?  :-)
>
> It is not strictly needed but allows the user to experiment/troubleshoot
> with the init RAM disk from GRUB as discussed earlier for --root.  Do
> you think it has enough value to be kept?

I’d rather avoid it for now. Less code is better. :-)

Toggle quote (6 lines)

>> (Also, in hindsight, I think it was a mistake to call them

>> ‘--something’. Following the common naming convention, we should rather

>> call these options ‘gnu.something’.)

> Is this convention detailed somewhere? I haven't found it in 'Standards'.

It’s a convention of the Linux kernel, I don’t know if it’s documented.

That’s it!

Ludo’.

Maxim Cournoyer wrote on 3 Mar 2020 06:00

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)(address . 37305@debbugs.gnu.org)

Message-ID:87y2sijane.fsf@gmail.com

Hello Ludovic!

Resuming review; I hope I'll be able to bring this series to completion

this time.

[...]

Toggle quote (4 lines)

> “gnu:” is for changes to (gnu packages). The idea is that the prefix

> should reflect what subsystem the commit is modifying. But yeah,

> looking at ‘git log’ can be inspiring. :-)

OK. Makes sense. Thanks for taking the time to explain!

[...]

Toggle quote (24 lines)>> About why providing options such as --root or --root-options in the
>> first place; I pondered about this as well, especially after making the
>> file systems from operating system able to be mounted with all their
>> (file system independent -- more on that later) options.  A reason I
>> came up with was that it allows to experiment at the GRUB command line
>> and change the root device, or perhaps the root options.  One use case
>> would be debugging the right options to pass to a file system driver in
>> case of a mistake in the operating system declaration.
>
> Yes, that makes sense.  It’s certainly useful to have ‘--root’ at least
> as an option.
>
>>> The main issue I see with this change is that mount(2) takes raw strings
>>> for the options.  There’s a convention to have those strings look like
>>> “KEY1=VALUE1,KEY2=VALUE2”, but it’s just a convention.
>>>
>>> As a rule of thumb, I’d rather have our interface be as close as
>>> possible to the actual mount(2) interface, which means taking strings.
>>>
>>> Now, we can surely add helper procedures to parse options that follow
>>> the above conventions.
>>>
>>> WDYT?

I don't feel too strongly about it, so I will adapt to your preference.

[...]

Toggle quote (21 lines)>> Are we really targeting mount(2)?  The commit
>> 9d3053819dfd834a1c29a03427c41d8524b8a7d5 (which you co-authored :-))
>> mentions 'man 8 mount' for the file system options.
>
> Right, mount(8) documents file system options that can be passed to
> mount(2).
>
> What does it mean to target mount(8) vs. mount(2)?  To me, mount(8) is a
> CLI to mount(2) that provides additional features to make the CLI more
> convenient: the “defaults” option, a way to pass mount(2) flags as
> options (like “ro”, “remount”, “bind”), /etc/fstab handling, etc.
>
> Guix System handles /etc/fstab differently and “defaults” makes little
> sense in our API (one can just use leave the default value of the
> ‘options’ field.)
>
> I think mount(8) is actually a good illustration of what not to do.  It
> ends up mixing things that are separate in the mount(2) API, and that
> doesn’t improve clarity and future-proof-ness (what if a file system has
> a “bind” option, etc.).

I think I agree. It may be cleaner to use the API of mount(2), while

mapping the low level stuff (binary flags?) to saner symbols, like it's

already done.

Toggle quote (4 lines)

> But again, I think the helper procedures that you propose to move back

> and forth between the string and the alist representations are very

> welcome. I just wouldn’t hard-code that directly in our API.

Okay. I'll keep them around in (gnu system file-systems); they're handy

when we need to check the presence of an option in the option string.

Toggle quote (21 lines)> WDYT?
>
>>>> From 67135c925b07f2e077b4cd852e07178691a10164 Mon Sep 17 00:00:00 2001
>>>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>>>> Date: Tue, 11 Feb 2020 14:14:36 -0500
>>>> Subject: [PATCH 6/9] gnu: linux-boot: Honor the "--root-options" kernel
>>>>  argument.
>>>>
>>>> * gnu/build/linux-boot.scm (boot-system): Parse the "--root-options" kernel
>>>> argument, and use it when calling `mount-root-file-system'.  Update doc.
>>>> * doc/guix.texi (Initial RAM Disk): Document the use of the "--root-options"
>>>> argument.
>>>
>>> Hmm do we really need this extra option?  :-)
>>
>> It is not strictly needed but allows the user to experiment/troubleshoot
>> with the init RAM disk from GRUB as discussed earlier for --root.  Do
>> you think it has enough value to be kept?
>
> I’d rather avoid it for now.  Less code is better.  :-)

Done!

Toggle quote (8 lines)

>>> (Also, in hindsight, I think it was a mistake to call them

>>> ‘--something’. Following the common naming convention, we should rather

>>> call these options ‘gnu.something’.)

>> Is this convention detailed somewhere? I haven't found it in 'Standards'.

> It’s a convention of the Linux kernel, I don’t know if it’s documented.

Well, with Hurd support coming, perhaps the standards used for a

particular kernel are not too relevant? :-)

One last thing that I'll try to understand before considering this

finished: I think adding a special properties field to the file system

record might not be necessary, as it seems that the subvol option can

take an absolute path as well as a relative path (I thought it only

accepted subvolume *names*.)

Will send the 3rd revision soon.

Thank you!

Maxim

Maxim Cournoyer wrote on 7 Mar 2020 05:01

Re: [bug#37305] Making system installation tests faster

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)(address . 37305@debbugs.gnu.org)

Message-ID:87eeu4ke45.fsf@gmail.com

Hello, this is a small update, to reflect recent changes to the code

base.

Maxim Cournoyer <maxim.cournoyer@gmail.com> writes:

Toggle quote (71 lines)> Hello Ludovic!
>
> I have more benchmark results.  Not to be compared with the previous
> results, as these were executed on a much faster machine (24 cores vs 4
> cores), with an SSD instead of a rotative disk.
>
>
> Commands used:
>
> current-guix/pre-built:
>
> time ./pre-inst-env guix system build -e \
> '(begin
>   (use-modules
>    (gnu packages package-management))
>   (parameterize
>       ((current-guix-package
> 	(current-guix/pre-built)))
>     ((@@
>       (gnu tests install)
>       operating-system-with-current-guix)
>      (@@
>       (gnu tests install)
>       %btrfs-root-os))))'
>
>
> current-guix (old fashionned):
>
> time ./pre-inst-env guix system build -e \
> '(begin
>   (use-modules
>    (gnu packages package-management))
>   (parameterize
>       ((current-guix-package
> 	(current-guix)))
>     ((@@
>       (gnu tests install)
>       operating-system-with-current-guix)
>      (@@
>       (gnu tests install)
>       %btrfs-root-os))))'
>
>
> new current-guix (guix self):
>
> time ./pre-inst-env guix system build -e \
> '(begin
>    (use-modules ((gnu ci) #:select (channel-instance->package))
> 		(guix monads)
> 		(guix channels)
> 		(guix store)
> 		((guix status) #:select (with-status-verbosity))
> 		((guix git-download) #:select (git-predicate))
> 		(guix utils)
> 		(gnu packages package-management))
>    (with-store store
>      (with-status-verbosity 2
>        (run-with-store store
> 	 (mlet* %store-monad
> 	     ((source-dir -> "/home/mcournoyer/src/guix")
> 	      (source (interned-file source-dir
> 				     "guix-source"
> 				     #:recursive? #t
> 				     #:select? (or (git-predicate source-dir)
> 						   (const #t))))
> 	      (instance -> (checkout->channel-instance source))
> 	      (new-guix-current -> (channel-instance->package instance)))
> 	   (return  (parameterize ((current-guix-package new-guix-current))
> 		      ((@@ (gnu tests install) operating-system-with-current-guix)
> 		       (@@ (gnu tests install) %btrfs-root-os)))))))))'

There have been some changes made to the modules used in the benchmark code

above. Here's an updated version:

Toggle snippet (27 lines)time ./pre-inst-env guix system build -e \
'(begin
   (use-modules ((gnu ci) #:select (channel-source->package))
		(guix monads)
		(guix channels)
		(guix store)
		((guix status) #:select (with-status-verbosity))
		((guix git-download) #:select (git-predicate))
		(guix utils)
		(gnu packages package-management))
   (with-store store
     (with-status-verbosity 2
       (run-with-store store
	 (mlet* %store-monad
	     ((source-dir -> "/home/mcournoyer/src/guix")
	      (source (interned-file source-dir
				     "guix-source"
				     #:recursive? #t
				     #:select? (or (git-predicate source-dir)
						   (const #t))))
	      (instance -> (checkout->channel-instance source))
	      (new-guix-current -> (channel-source->package instance)))
	   (return  (parameterize ((current-guix-package new-guix-current))
		      ((@@ (gnu tests install) operating-system-with-current-guix)
		       (@@ (gnu tests install) %btrfs-root-os)))))))))'

Time taken:

real 7m39.912s

user 0m57.129s

sys 0m0.499s

Compared to recycling the Guix development copy from the work tree:

real 0m48.297s

user 0m2.516s

sys 0m0.229s

Maxim

maxim.cournoyer wrote on 18 Mar 2020 16:27

Re: [bug#37305] Allow booting from a Btrfs subvolume [review part 2]

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)

Message-ID:87v9n1pts1.fsf@raisin.i-did-not-set--mail-host-address--so-tickle-me

Hi Ludovic,

Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (26 lines)> Hi Maxim,
>
> Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:
>
>>>> +              (error "The store is on a Btrfs subvolume, but the \
>>>> +subvolume name is unknown.
>>>> +Hint: Define the \"btrfs-subvolume-path\" file system property or
>>>> +use the \"subvol\" Btrfs file system option."))))
>>
>>> Rather use ‘raise’ with ‘&message’ and ‘&fix-hint’ conditions.
>>
>> I tried this, but importing (guix utils) to acces &fix-hint caused the init
>> RAM disk to fail mysteriously:
>
> Oh, my bad.  We should move ‘&fix-hint’ to (guix diagnostics)
> eventually.
>
> In the meantime, I’d say just raise a ‘&message’ and leave the hint as a
> comment (it’s not supposed to be a user-facing interface).  Or maybe you
> could define a specific error condition type for this?
>
> Thanks,
> Ludo’.
>
> PS: I’ll comment on the other bits ASAP!

The remaining, reworked patches for this series are attached below. If

you don't want to merge the 0001 for now (speeding up tests), I don't

mind too much (though it still provides value to me, when using old

hardware).

Thanks for your patience :-)

Maxim

Attachment: 0001-gnu-tests-Reduce-the-time-required-to-run-the-system.patch

Attachment: 0002-linux-boot-Refactor-boot-system.patch

Attachment: 0003-file-systems-Add-helpers-for-parsing-the-options-str.patch

Attachment: 0004-bootloader-grub-Allow-booting-from-a-Btrfs-subvolume.patch

Maxim Cournoyer wrote on 31 Mar 2020 03:26

control message for bug #40236

Recipients:(address . control@debbugs.gnu.org)

Message-ID:87lfnhz544.fsf@gmail.com

block 40236 by 37305

quit

Pierre Neidhardt wrote on 17 May 2020 15:29

Re: [bug#37305] Allow booting from a Btrfs subvolume [review part 2]

Recipients:(address . maxim.cournoyer@gmail.com)

Message-ID:87o8qmheqm.fsf@ambrevar.xyz

I'll try to test this patch, possibly today.

Is the patch from the 18th of March the right one?

The patch says it has 5 components but the last patch is

Toggle snippet (3 lines)

[PATCH 4/5] bootloader: grub: Allow booting from a Btrfs subvolume.

Is the fifth patch missing?

Pierre Neidhardt

https://ambrevar.xyz/

Pierre Neidhardt wrote on 17 May 2020 16:03

Recipients:(address . maxim.cournoyer@gmail.com)

Message-ID:87lflqhd4y.fsf@ambrevar.xyz

Hi Maxim!

maxim.cournoyer@gmail.com writes:

Toggle quote (8 lines)> +@lisp
> +(file-system
> +  (device (file-system-label "btrfs-pool-1"))
> +  (mount-point "/gnu/store")
> +  (type "btrfs")
> +  (options "subvol=root-snapshots/root-current/guix-store"
> +  ("compress" . "zstd")))
> +@end lisp

Is this ("compress" . "zstd") intentional? It does not match the first example.

Pierre Neidhardt

https://ambrevar.xyz/

Maxim Cournoyer wrote on 17 May 2020 18:13

Re: [PATCH v3] Allow booting from a Btrfs subvolume

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)

Message-ID:874ksebkup.fsf_-_@gmail.com

Hi Pierre!

Pierre Neidhardt <mail@ambrevar.xyz> writes:

Toggle quote (9 lines)

> I'll try to test this patch, possibly today.

> Is the patch from the 18th of March the right one?

> The patch says it has 5 components but the last patch is

> [PATCH 4/5] bootloader: grub: Allow booting from a Btrfs subvolume.

> Is the fifth patch missing?

As I send my patches as attachments rather than with git send-email, I
simply forgot to update the subject line counter.  There are four
patches in total, although the hack in 0001 was controversial so is
included as a curiosity only. I've rebased the 4 patches on current
master, and fixed some conflict that arose in the initrd code.

Attachment: 0001-gnu-tests-Reduce-the-time-required-to-run-the-system.patch

Attachment: 0002-linux-boot-Refactor-boot-system.patch

Attachment: 0003-file-systems-Add-helpers-for-parsing-the-options-str.patch

Attachment: 0004-bootloader-grub-Allow-booting-from-a-Btrfs-subvolume.patch

I'll test those rebased patches as well. The automated system tests no

longer pass -- although it might not be specific to this change (I'll

try running the "installed-os" test to see if all the install tests are

broken or just this new "btrfs-root-on-subvolume-os" one.). In case you

are not familiar with system tests, a single system test can be run

with:

make check-system TESTS="btrfs-root-on-subvolume-os"

The install tests are defined under (gnu tests install) and are *very*

expensive to run (mostly in time, but they require some disk space as

well). Hence the hack in 0001, but I'm not sure if it brings as much

benefits as when I made it (given Ludovic keeps improving the way Guix

gets built :-)).

A fresh benchmark could be interesting if you have lots of time on your

hands.

Thank you for looking at this!

Maxim

Maxim Cournoyer wrote on 17 May 2020 18:16

Re: [bug#37305] Allow booting from a Btrfs subvolume [review part 2]

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)

Message-ID:87zha6a660.fsf@gmail.com

Hello Pierre!

Pierre Neidhardt <mail@ambrevar.xyz> writes:

Toggle quote (16 lines)> Hi Maxim!
>
> maxim.cournoyer@gmail.com writes:
>
>> +@lisp
>> +(file-system
>> +  (device (file-system-label "btrfs-pool-1"))
>> +  (mount-point "/gnu/store")
>> +  (type "btrfs")
>> +  (options "subvol=root-snapshots/root-current/guix-store"
>> +  ("compress" . "zstd")))
>> +@end lisp
>
>
> Is this ("compress" . "zstd") intentional?  It does not match the first example.

Good catch! This pair syntax was removed, based on feedback from

previous reviews. It can still be used with the helper procedures, but

not directly like that. It should be fixed in the patch series I just

sent in the previous email.

Thank you,

Maxim

Pierre Neidhardt wrote on 17 May 2020 18:37

Re: [PATCH v3] Allow booting from a Btrfs subvolume

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)

Message-ID:878shqh61b.fsf@ambrevar.xyz

Maxim Cournoyer <maxim.cournoyer@gmail.com> writes:

Toggle quote (9 lines)

> I'll test those rebased patches as well. The automated system tests no

> longer pass -- although it might not be specific to this change (I'll

> try running the "installed-os" test to see if all the install tests are

> broken or just this new "btrfs-root-on-subvolume-os" one.). In case you

> are not familiar with system tests, a single system test can be run

> with:

> make check-system TESTS="btrfs-root-on-subvolume-os"

OK. Do you want me to try something here?

Toggle quote (9 lines)

> The install tests are defined under (gnu tests install) and are *very*

> expensive to run (mostly in time, but they require some disk space as

> well). Hence the hack in 0001, but I'm not sure if it brings as much

> benefits as when I made it (given Ludovic keeps improving the way Guix

> gets built :-)).

> A fresh benchmark could be interesting if you have lots of time on your

> hands.

Sorry, I don't think I'll be able to do that :)

Cheers!

Pierre Neidhardt

https://ambrevar.xyz/

Pierre Neidhardt wrote on 17 May 2020 21:05

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)

Message-ID:87367ygz5c.fsf@ambrevar.xyz

Finally "system reconfigure"d with your patch.

The first entry fails to boot with:

Toggle snippet (6 lines)

loading kernel modules...

ice-9/boot-9.scm:1669:16: In procedure raise-exception:

In procedure string-prefix?: Wrong type argument in position 2

(expecting string): #<file-system-label "guix">

Funny thing is, the previous generation boots properly!

My config.scm's filesystem layout:

Toggle snippet (17 lines)(file-systems (cons* (file-system
                         (device (file-system-label "guix"))
                         (mount-point "/")
                         (type "btrfs")
                         (options "subvol=rootfs,compress=zstd"))
                       (file-system
                         (device (file-system-label "data"))
                         (mount-point "/media/data")
                         (type "ext4"))
                       (file-system
                         (mount-point "/tmp")
                         (device "none")
                         (type "tmpfs")
                         (check? #f))
                       %base-file-systems))

-- 
Pierre Neidhardt
https://ambrevar.xyz/

Pierre Neidhardt wrote on 17 May 2020 21:09

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)

Message-ID:87y2pqfkfa.fsf@ambrevar.xyz

The first (failing) generation was generated from today (or yesterday) Guix

master.

The second (working) generation was generated from a 2 month old Guix.

Let me know if you want the precise commits.

Pierre Neidhardt

https://ambrevar.xyz/

Pierre Neidhardt wrote on 17 May 2020 21:48

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)

Message-ID:87v9kufim7.fsf@ambrevar.xyz

False alarm, I had mis-merged your (old) patch.

After applying your patch from today, everything works.

I won't have access to this computer in 24 hours. Let me know if there

is anything else you'd like me to test before then! :)

Cheers!

Pierre Neidhardt

https://ambrevar.xyz/

Pierre Neidhardt wrote on 17 May 2020 22:22

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)

Message-ID:87sgfyfh1m.fsf@ambrevar.xyz

About the patch: maybe some of it could be generalized to other

filesystems that support subvolumes (like ZFS). But this is probably

premature and it could be long before someone tests all this with ZFS,

thus it's probably wiser to first merge this and generalize later.

By the way, can Guix with installed on a ZFS root?

Pierre Neidhardt

https://ambrevar.xyz/

Maxim Cournoyer wrote on 18 May 2020 02:49

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)

Message-ID:87v9ku9ie5.fsf@gmail.com

Hi Pierre,

Pierre Neidhardt <mail@ambrevar.xyz> writes:

Toggle quote (7 lines)

> About the patch: maybe some of it could be generalized to other

> filesystems that support subvolumes (like ZFS). But this is probably

> premature and it could be long before someone tests all this with ZFS,

> thus it's probably wiser to first merge this and generalize later.

> By the way, can Guix with installed on a ZFS root?

I'm sorry, but I know nothing about ZFS, except that its license is

incompatible with the GPL.

Maxim

Maxim Cournoyer wrote on 18 May 2020 03:16

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)

Message-ID:87r1vi9h5q.fsf@gmail.com

Hello Pierre!

Pierre Neidhardt <mail@ambrevar.xyz> writes:

Toggle quote (3 lines)

> False alarm, I had mis-merged your (old) patch.

> After applying your patch from today, everything works.

Woohoo! Thank you for testing!

Toggle quote (3 lines)

> I won't have access to this computer in 24 hours. Let me know if there

> is anything else you'd like me to test before then! :)

Any comments above the implementation (code) or documentation? Was

something unclear/missing?

If everything looks OK, I'll give this another 2 weeks period from now

to allow for more comments then merge the three non-controversial (i.e.,

all except the hack to speed installation tests) to the master branch.

Thank you!

Maxim

Pierre Neidhardt wrote on 18 May 2020 10:54

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)

Message-ID:87k119fwrt.fsf@ambrevar.xyz

Hi Maxim!

Toggle quote (6 lines)

>> I won't have access to this computer in 24 hours. Let me know if there

>> is anything else you'd like me to test before then! :)

> Any comments above the implementation (code) or documentation? Was

> something unclear/missing?

I had a cursory look. The code looks good to me as far as I can tell

(which is little, considering my limited knowledge of the file system

code).

The documentation examples are good. The paragraphs are a little bit

harder to grasp for me, but I am not sure how to make it clearer;

besides, the examples really help clarifying everything, so it's good enough.

Thanks!

Pierre Neidhardt

https://ambrevar.xyz/

Ludovic Courtès wrote on 18 May 2020 23:55

Re: [bug#37305] [PATCH v3] Allow booting from a Btrfs subvolume

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)

Message-ID:87y2pohpqx.fsf@gnu.org

Hi Maxim,

Sorry for dropping the ball for sooo long.

Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:

Toggle quote (17 lines)>>From b03a574ad565b34bbe8a7d3d0322591850984dc6 Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Tue, 11 Feb 2020 14:00:06 -0500
> Subject: [PATCH 2/4] linux-boot: Refactor boot-system.
>
> The --root option can now be omitted, and inferred from the root file system
> declaration instead.
>
> * gnu/build/file-systems.scm (canonicalize-device-spec): Extend to support NFS
> directly, and...
> * gnu/build/linux-boot.scm (boot-system): ...remove NFS special casing from
> here.  Remove nested definitions for root-fs-type, root-fs-flags and
> root-fs-options, and bind those inside the let* instead.  Make "--root" take
> precedence over the device field string representation of the root file
> system.
> * doc/guix.texi (Initial RAM Disk): Document that "--root" can be left

[...]

Toggle quote (25 lines)> +++ b/gnu/build/linux-boot.scm
> @@ -498,25 +498,13 @@ upon error."
>    (define (root-mount-point? fs)
>      (string=? (file-system-mount-point fs) "/"))
>  
> -  (define root-fs-type
> -    (or (any (lambda (fs)
> -               (and (root-mount-point? fs)
> -                    (file-system-type fs)))
> -             mounts)
> -        "ext4"))
> -
> -  (define root-fs-flags
> -    (mount-flags->bit-mask (or (any (lambda (fs)
> -                                      (and (root-mount-point? fs)
> -                                           (file-system-flags fs)))
> -                                    mounts)
> -                               '())))
> -
> -  (define root-fs-options
> -    (any (lambda (fs)
> -           (and (root-mount-point? fs)
> -                (file-system-options fs)))
> -         mounts))

[...]

Toggle quote (11 lines)> +             (root-fs (find root-mount-point? mounts))
> +             (root-fs-type (or (and=> root-fs file-system-type)
> +                               "ext4"))
> +             (root-fs-device (and=> root-fs file-system-device))
> +             (root-fs-flags (mount-flags->bit-mask
> +                             (or (and=> root-fs file-system-flags)
> +                                 '())))
> +             (root-options (if root-fs
> +                               (file-system-options root-fs)
> +                               #f))

I would tend to keep these as defines to make the ‘let*’ less

intimidating, but it’s a detail.

Toggle quote (5 lines)

> + ;; XXX: Importing (guix utils) and using &fix-hint causes the

> + ;; following error when booting the init RAM disk: "ERROR: In

> + ;; procedure dynamic-func:\nIn procedure dynamic-pointer: Symbol not

> + ;; found: strverscmp", so we just embed the hint in the message.

I think it should just be “FIXME: Use &fix-hint once it no longer pulls

in (guix utils).”

Toggle quote (26 lines)>>From 082934db68964890ebd2a2118fb44d66911844d3 Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Sun, 14 Jul 2019 20:50:23 +0900
> Subject: [PATCH 4/4] bootloader: grub: Allow booting from a Btrfs subvolume.
>
> * gnu/bootloader/grub.scm (strip-mount-point): Remove procedure.
> (normalize-file): Add procedure.
> (grub-configuration-file): New BTRFS-SUBVOLUME-FILE-NAME parameter.  When
> defined, prepend its value to the kernel and initrd file names, using the
> NORMALIZE-FILE procedure.  Adjust the call to EYE-CANDY to pass the
> BTRFS-SUBVOLUME-FILE-NAME argument.  Normalize the KEYMAP file as well.
> (eye-candy): Add a BTRFS-SUBVOLUME-FILE-NAME parameter, and use it, along with
> the NORMALIZE-FILE procedure, to normalize the FONT-FILE and IMAGE nested
> variables.  Adjust doc.
> * gnu/bootloader/depthcharge.scm (depthcharge-configuration-file): Adapt.
> * gnu/bootloader/extlinux.scm (extlinux-configuration-file): Likewise.
> * gnu/system/file-systems.scm (btrfs-subvolume?)
> (btrfs-store-subvolume-file-name): New procedures.
> * gnu/system.scm (operating-system-bootcfg): Specify the Btrfs
> subvolume file name the store resides on to the
> `operating-system-bootcfg' procedure, using the new
> BTRFS-SUBVOLUME-FILE-NAME argument.
> * doc/guix.texi (File Systems): Add a Btrfs subsection to document the use of
> subvolumes.
> * gnu/tests/install.scm: Add test "btrfs-root-on-subvolume-os".

Please list the entities added to ‘install.scm’.

Toggle quote (4 lines)

> (define* (eye-candy config store-device store-mount-point

> + btrfs-store-subvolume-file-name

> #:key system port)

I think ‘btrfs-store-subvolume-file-name’ should be a keyword argument.

Toggle quote (7 lines)

> (define* (grub-configuration-file config entries

> #:key

> (system (%current-system))

> - (old-entries '()))

> + (old-entries '())

> + btrfs-subvolume-file-name)

I wonder if we should just call it ‘store-directory-prefix’ or similar

since, after all, it’s just about prepending a prefix, which could

perhaps be useful for file systems other than Btrfs.

Thoughts?

Anyway, that’s great work, so I’ll be happy to finally see it committed

in the coming days!

Ludo’.

Maxim Cournoyer wrote on 20 May 2020 14:44

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)

Message-ID:87wo56pyi5.fsf@gmail.com

Hi Ludovic!

Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (4 lines)

> Hi Maxim,

> Sorry for dropping the ball for sooo long.

No worries :-)

Toggle quote (2 lines)

> Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:

[...]

Toggle quote (41 lines)>> +++ b/gnu/build/linux-boot.scm
>> @@ -498,25 +498,13 @@ upon error."
>>    (define (root-mount-point? fs)
>>      (string=? (file-system-mount-point fs) "/"))
>>  
>> -  (define root-fs-type
>> -    (or (any (lambda (fs)
>> -               (and (root-mount-point? fs)
>> -                    (file-system-type fs)))
>> -             mounts)
>> -        "ext4"))
>> -
>> -  (define root-fs-flags
>> -    (mount-flags->bit-mask (or (any (lambda (fs)
>> -                                      (and (root-mount-point? fs)
>> -                                           (file-system-flags fs)))
>> -                                    mounts)
>> -                               '())))
>> -
>> -  (define root-fs-options
>> -    (any (lambda (fs)
>> -           (and (root-mount-point? fs)
>> -                (file-system-options fs)))
>> -         mounts))
>
> [...]
>
>> +             (root-fs (find root-mount-point? mounts))
>> +             (root-fs-type (or (and=> root-fs file-system-type)
>> +                               "ext4"))
>> +             (root-fs-device (and=> root-fs file-system-device))
>> +             (root-fs-flags (mount-flags->bit-mask
>> +                             (or (and=> root-fs file-system-flags)
>> +                                 '())))
>> +             (root-options (if root-fs
>> +                               (file-system-options root-fs)
>> +                               #f))
>
> I would tend to keep these as defines to make the ‘let*’ less
> intimidating, but it’s a detail.

It would only *appear* less intimidating ;-). I personally prefer the

let* versions as the logic is more succinctly expressed (there is no

need for 'any' + lambdas, for example).

Toggle quote (8 lines)

>> + ;; XXX: Importing (guix utils) and using &fix-hint causes the

>> + ;; following error when booting the init RAM disk: "ERROR: In

>> + ;; procedure dynamic-func:\nIn procedure dynamic-pointer: Symbol not

>> + ;; found: strverscmp", so we just embed the hint in the message.

> I think it should just be “FIXME: Use &fix-hint once it no longer pulls

> in (guix utils).”

Done!

Toggle quote (28 lines)>>>From 082934db68964890ebd2a2118fb44d66911844d3 Mon Sep 17 00:00:00 2001
>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>> Date: Sun, 14 Jul 2019 20:50:23 +0900
>> Subject: [PATCH 4/4] bootloader: grub: Allow booting from a Btrfs subvolume.
>>
>> * gnu/bootloader/grub.scm (strip-mount-point): Remove procedure.
>> (normalize-file): Add procedure.
>> (grub-configuration-file): New BTRFS-SUBVOLUME-FILE-NAME parameter.  When
>> defined, prepend its value to the kernel and initrd file names, using the
>> NORMALIZE-FILE procedure.  Adjust the call to EYE-CANDY to pass the
>> BTRFS-SUBVOLUME-FILE-NAME argument.  Normalize the KEYMAP file as well.
>> (eye-candy): Add a BTRFS-SUBVOLUME-FILE-NAME parameter, and use it, along with
>> the NORMALIZE-FILE procedure, to normalize the FONT-FILE and IMAGE nested
>> variables.  Adjust doc.
>> * gnu/bootloader/depthcharge.scm (depthcharge-configuration-file): Adapt.
>> * gnu/bootloader/extlinux.scm (extlinux-configuration-file): Likewise.
>> * gnu/system/file-systems.scm (btrfs-subvolume?)
>> (btrfs-store-subvolume-file-name): New procedures.
>> * gnu/system.scm (operating-system-bootcfg): Specify the Btrfs
>> subvolume file name the store resides on to the
>> `operating-system-bootcfg' procedure, using the new
>> BTRFS-SUBVOLUME-FILE-NAME argument.
>> * doc/guix.texi (File Systems): Add a Btrfs subsection to document the use of
>> subvolumes.
>> * gnu/tests/install.scm: Add test "btrfs-root-on-subvolume-os".
>
> Please list the entities added to ‘install.scm’.

Done!

Toggle quote (6 lines)

>> (define* (eye-candy config store-device store-mount-point

>> + btrfs-store-subvolume-file-name

>> #:key system port)

> I think ‘btrfs-store-subvolume-file-name’ should be a keyword argument.

Done!

Toggle quote (13 lines)>>  (define* (grub-configuration-file config entries
>>                                    #:key
>>                                    (system (%current-system))
>> -                                  (old-entries '()))
>> +                                  (old-entries '())
>> +                                  btrfs-subvolume-file-name)
>
> I wonder if we should just call it ‘store-directory-prefix’ or similar
> since, after all, it’s just about prepending a prefix, which could
> perhaps be useful for file systems other than Btrfs.
>
> Thoughts?

Perhaps, but given it's not yet clear whether another file system will

require similar support from GRUB, I'd rather keep things as explicit as

possible for now.

Toggle quote (3 lines)

> Anyway, that’s great work, so I’ll be happy to finally see it committed

> in the coming days!

Thanks for the great words and for having a last look :-).

I've added a news entry and pushed to master as:

Toggle snippet (6 lines)

489699c456 allow-booting-from-btrfs-subvolume news: Add entry for Btrfs subvolume boot support.

b460ba7992 bootloader: grub: Allow booting from a Btrfs subvolume.

fa35fb58c8 file-systems: Add helpers for parsing the options string into an alist.

281d80d8e5 linux-boot: Refactor boot-system.

Closing! Thanks to Pierre and Ludovic for testing and reviewing.

Maxim

Maxim Cournoyer wrote on 20 May 2020 14:44

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)

Message-ID:87v9kqpyhu.fsf@gmail.com

Hi Ludovic!

Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (4 lines)

> Hi Maxim,

> Sorry for dropping the ball for sooo long.

No worries :-)

Toggle quote (2 lines)

> Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:

[...]

Toggle quote (41 lines)>> +++ b/gnu/build/linux-boot.scm
>> @@ -498,25 +498,13 @@ upon error."
>>    (define (root-mount-point? fs)
>>      (string=? (file-system-mount-point fs) "/"))
>>  
>> -  (define root-fs-type
>> -    (or (any (lambda (fs)
>> -               (and (root-mount-point? fs)
>> -                    (file-system-type fs)))
>> -             mounts)
>> -        "ext4"))
>> -
>> -  (define root-fs-flags
>> -    (mount-flags->bit-mask (or (any (lambda (fs)
>> -                                      (and (root-mount-point? fs)
>> -                                           (file-system-flags fs)))
>> -                                    mounts)
>> -                               '())))
>> -
>> -  (define root-fs-options
>> -    (any (lambda (fs)
>> -           (and (root-mount-point? fs)
>> -                (file-system-options fs)))
>> -         mounts))
>
> [...]
>
>> +             (root-fs (find root-mount-point? mounts))
>> +             (root-fs-type (or (and=> root-fs file-system-type)
>> +                               "ext4"))
>> +             (root-fs-device (and=> root-fs file-system-device))
>> +             (root-fs-flags (mount-flags->bit-mask
>> +                             (or (and=> root-fs file-system-flags)
>> +                                 '())))
>> +             (root-options (if root-fs
>> +                               (file-system-options root-fs)
>> +                               #f))
>
> I would tend to keep these as defines to make the ‘let*’ less
> intimidating, but it’s a detail.

It would only *appear* less intimidating ;-). I personally prefer the

let* versions as the logic is more succinctly expressed (there is no

need for 'any' + lambdas, for example).

Toggle quote (8 lines)

>> + ;; XXX: Importing (guix utils) and using &fix-hint causes the

>> + ;; following error when booting the init RAM disk: "ERROR: In

>> + ;; procedure dynamic-func:\nIn procedure dynamic-pointer: Symbol not

>> + ;; found: strverscmp", so we just embed the hint in the message.

> I think it should just be “FIXME: Use &fix-hint once it no longer pulls

> in (guix utils).”

Done!

Toggle quote (28 lines)>>>From 082934db68964890ebd2a2118fb44d66911844d3 Mon Sep 17 00:00:00 2001
>> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
>> Date: Sun, 14 Jul 2019 20:50:23 +0900
>> Subject: [PATCH 4/4] bootloader: grub: Allow booting from a Btrfs subvolume.
>>
>> * gnu/bootloader/grub.scm (strip-mount-point): Remove procedure.
>> (normalize-file): Add procedure.
>> (grub-configuration-file): New BTRFS-SUBVOLUME-FILE-NAME parameter.  When
>> defined, prepend its value to the kernel and initrd file names, using the
>> NORMALIZE-FILE procedure.  Adjust the call to EYE-CANDY to pass the
>> BTRFS-SUBVOLUME-FILE-NAME argument.  Normalize the KEYMAP file as well.
>> (eye-candy): Add a BTRFS-SUBVOLUME-FILE-NAME parameter, and use it, along with
>> the NORMALIZE-FILE procedure, to normalize the FONT-FILE and IMAGE nested
>> variables.  Adjust doc.
>> * gnu/bootloader/depthcharge.scm (depthcharge-configuration-file): Adapt.
>> * gnu/bootloader/extlinux.scm (extlinux-configuration-file): Likewise.
>> * gnu/system/file-systems.scm (btrfs-subvolume?)
>> (btrfs-store-subvolume-file-name): New procedures.
>> * gnu/system.scm (operating-system-bootcfg): Specify the Btrfs
>> subvolume file name the store resides on to the
>> `operating-system-bootcfg' procedure, using the new
>> BTRFS-SUBVOLUME-FILE-NAME argument.
>> * doc/guix.texi (File Systems): Add a Btrfs subsection to document the use of
>> subvolumes.
>> * gnu/tests/install.scm: Add test "btrfs-root-on-subvolume-os".
>
> Please list the entities added to ‘install.scm’.

Done!

Toggle quote (6 lines)

>> (define* (eye-candy config store-device store-mount-point

>> + btrfs-store-subvolume-file-name

>> #:key system port)

> I think ‘btrfs-store-subvolume-file-name’ should be a keyword argument.

Done!

Toggle quote (13 lines)>>  (define* (grub-configuration-file config entries
>>                                    #:key
>>                                    (system (%current-system))
>> -                                  (old-entries '()))
>> +                                  (old-entries '())
>> +                                  btrfs-subvolume-file-name)
>
> I wonder if we should just call it ‘store-directory-prefix’ or similar
> since, after all, it’s just about prepending a prefix, which could
> perhaps be useful for file systems other than Btrfs.
>
> Thoughts?

Perhaps, but given it's not yet clear whether another file system will

require similar support from GRUB, I'd rather keep things as explicit as

possible for now.

Toggle quote (3 lines)

> Anyway, that’s great work, so I’ll be happy to finally see it committed

> in the coming days!

Thanks for the great words and for having a last look :-).

I've added a news entry and pushed to master as:

Toggle snippet (6 lines)

489699c456 allow-booting-from-btrfs-subvolume news: Add entry for Btrfs subvolume boot support.

b460ba7992 bootloader: grub: Allow booting from a Btrfs subvolume.

fa35fb58c8 file-systems: Add helpers for parsing the options string into an alist.

281d80d8e5 linux-boot: Refactor boot-system.

Closing! Thanks to Pierre and Ludovic for testing and reviewing.

Maxim

Closed

Pierre Neidhardt wrote on 20 May 2020 15:29

Recipients:(address . 37305-done@debbugs.gnu.org)

Message-ID:87imgq68g1.fsf@ambrevar.xyz

Thanks a lot!
This is a huge improvement to Guix in my opinion :)

-- 
Pierre Neidhardt
https://ambrevar.xyz/

Closed

Ludovic Courtès wrote on 21 May 2020 00:03

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)

Message-ID:87sgfuclh8.fsf@gnu.org

Hi Maxim,

Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:

[...]

Toggle quote (17 lines)>>>  (define* (grub-configuration-file config entries
>>>                                    #:key
>>>                                    (system (%current-system))
>>> -                                  (old-entries '()))
>>> +                                  (old-entries '())
>>> +                                  btrfs-subvolume-file-name)
>>
>> I wonder if we should just call it ‘store-directory-prefix’ or similar
>> since, after all, it’s just about prepending a prefix, which could
>> perhaps be useful for file systems other than Btrfs.
>>
>> Thoughts?
>
> Perhaps, but given it's not yet clear whether another file system will
> require similar support from GRUB, I'd rather keep things as explicit as
> possible for now.

To me, another consideration is familiarity with Btrfs for those who’ll
touch the code: to someone not familiar with it, the code may be viewed
as “read-only” because it says “btrfs”.  Whereas if it clearly states
that it’s just about prepending a directory name or similar, it’s easy
to reason about it.

Toggle quote (14 lines)>> Anyway, that’s great work, so I’ll be happy to finally see it committed
>> in the coming days!
>
> Thanks for the great words and for having a last look :-).
>
> I've added a news entry and pushed to master as:
>
> 489699c456 allow-booting-from-btrfs-subvolume news: Add entry for Btrfs subvolume boot support.
> b460ba7992 bootloader: grub: Allow booting from a Btrfs subvolume.
> fa35fb58c8 file-systems: Add helpers for parsing the options string into an alist.
> 281d80d8e5 linux-boot: Refactor boot-system.
>
> Closing!  Thanks to Pierre and Ludovic for testing and reviewing.

That’s great news, thanks a lot for your work… and for your patience!

Ludo’.

Closed

Pierre Neidhardt wrote on 21 May 2020 08:58

Recipients:(address . 37305-done@debbugs.gnu.org)

Message-ID:87tv09epux.fsf@ambrevar.xyz

Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (6 lines)

> To me, another consideration is familiarity with Btrfs for those who’ll

> touch the code: to someone not familiar with it, the code may be viewed

> as “read-only” because it says “btrfs”. Whereas if it clearly states

> that it’s just about prepending a directory name or similar, it’s easy

> to reason about it.

Agreed, this is where I was going to with my comment on ZFS.
Maybe the "btrfs" part of the symbols can be left out to make it more
general and understandable.

-- 
Pierre Neidhardt
https://ambrevar.xyz/

Closed

Maxim Cournoyer wrote on 28 May 2020 06:30

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)

Message-ID:87lflc65ra.fsf@gmail.com

Hello,

Pierre Neidhardt <mail@ambrevar.xyz> writes:

Toggle quote (12 lines)> Ludovic Courtès <ludo@gnu.org> writes:
>
>> To me, another consideration is familiarity with Btrfs for those who’ll
>> touch the code: to someone not familiar with it, the code may be viewed
>> as “read-only” because it says “btrfs”.  Whereas if it clearly states
>> that it’s just about prepending a directory name or similar, it’s easy
>> to reason about it.
>
> Agreed, this is where I was going to with my comment on ZFS.
> Maybe the "btrfs" part of the symbols can be left out to make it more
> general and understandable.

I've adapted with the naming suggested earlier by Ludovic. Does the

patch below fit the bill?

Thanks,

Maxim

Attachment: 0001-bootloader-grub-Rename-the-btrfs-subvolume-file-name.patch

Closed

Pierre Neidhardt wrote on 28 May 2020 10:26

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)

Message-ID:87eer44g8m.fsf@ambrevar.xyz

Looks good to me, but:

Toggle quote (2 lines)

> (btrfs-store-subvolume-file-name file-systems))))

I haven't looked at the context, but is this procedure really Btrfs-only?

Pierre Neidhardt

https://ambrevar.xyz/

Closed

Ludovic Courtès wrote on 28 May 2020 14:30

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)

Message-ID:87tv00ckdl.fsf@gnu.org

Hello!

Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:

Toggle quote (16 lines)> From ee23cc391cce7b8dcdcb5146d4b84a55881a5cb9 Mon Sep 17 00:00:00 2001
> From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
> Date: Wed, 27 May 2020 22:44:28 -0400
> Subject: [PATCH] bootloader: grub: Rename the btrfs-subvolume-file-name
>  parameter.
>
> Following discussion in <https://issues.guix.gnu.org/37305>, it seems more
> appropriate to give the parameter a more generic name that better describes
> what it does.
>
> * gnu/bootloader/grub.scm (normalize-file): Rename the
> btrfs-subvolume-file-name parameter to store-directory-prefix.
> (eye-candy): Likewise.
> (grub-configuration-file): Likewise.
> * gnu/system.scm (operating-system-bootcfg): Adapt.

[...]

Toggle quote (13 lines)>  (define* (eye-candy config store-device store-mount-point
> -                    #:key btrfs-store-subvolume-file-name system port)
> +                    #:key store-directory-prefix system port)
>    "Return a gexp that writes to PORT (a port-valued gexp) the 'grub.cfg' part
>  concerned with graphics mode, background images, colors, and all that.
>  STORE-DEVICE designates the device holding the store, and STORE-MOUNT-POINT is
>  its mount point; these are used to determine where the background image and
>  fonts must be searched for.  SYSTEM must be the target system string---e.g.,
> -\"x86_64-linux\".  BTRFS-STORE-SUBVOLUME-FILE-NAME is the file name of the
> -Btrfs subvolume, to be prepended to any store path, if any."
> +\"x86_64-linux\".  STORE-DIRECTORY-PREFIX is a directory prefix to prepend to
> +any store path."

s/path/file name/ :-)

Maybe you can have ‘store-directory-prefix’ default to "" and adjust

users accordingly.

Otherwise LGTM.

Thanks for taking the time to work on this patch!

Ludo’.

Closed

Maxim Cournoyer wrote on 29 May 2020 23:14

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)

Message-ID:87sgfixx37.fsf@gmail.com

Pierre Neidhardt <mail@ambrevar.xyz> writes:

Toggle quote (6 lines)

> Looks good to me, but:

>> (btrfs-store-subvolume-file-name file-systems))))

> I haven't looked at the context, but is this procedure really Btrfs-only?

Yes!  It takes a list of file-system objects, filter the Btrfs ones
using the "subvol" option and returns their subvolume file name, if the
store is mounted on such subvolume.

Closed

Maxim Cournoyer wrote on 30 May 2020 04:00

Recipients:(name . Ludovic Courtès)(address . ludo@gnu.org)

Message-ID:87lflaxjv2.fsf@gmail.com

Hello,

Ludovic Courtès <ludo@gnu.org> writes:

[...]

Toggle quote (5 lines)

>> +\"x86_64-linux\". STORE-DIRECTORY-PREFIX is a directory prefix to prepend to

>> +any store path."

> s/path/file name/ :-)

Done!

Toggle quote (3 lines)

> Maybe you can have ‘store-directory-prefix’ default to "" and adjust

> users accordingly.

I tried adapting it, but my test failed, because I had failed to adapt
one of the users.  I find that using #f to denote the absence of a
prefix more accurate than "" (after all, "" is a valid prefix) and
composes better (should we want to 'or' a couple of prefix seeking
procedures together).

Toggle quote (4 lines)

> Otherwise LGTM.

> Thanks for taking the time to work on this patch!

No problem :-) Pushed as commit e7b86a0d88.

Maxim

Closed

Pierre Neidhardt wrote on 30 May 2020 09:32

Recipients:(address . 37305-done@debbugs.gnu.org)

Message-ID:87r1v1dgiw.fsf@ambrevar.xyz

Hooray!

Pierre Neidhardt

https://ambrevar.xyz/

Closed

Pierre Neidhardt wrote on 30 May 2020 09:32

Recipients:(address . 37305-done@debbugs.gnu.org)

Message-ID:87o8q5dgif.fsf@ambrevar.xyz

Maybe this would warrant a blog post! :)

Pierre Neidhardt

https://ambrevar.xyz/

Closed

Maxim Cournoyer wrote on 31 May 2020 04:44

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)

Message-ID:87k10sx1qn.fsf@gmail.com

Hello Pierre!

Pierre Neidhardt <mail@ambrevar.xyz> writes:

Toggle quote (2 lines)

> Maybe this would warrant a blog post! :)

Eh, I'm not sure. It's kind of a basic thing that other systems already

support (through grub-mkconfig). We've just caught up :-).

Maxim

Closed

Pierre Neidhardt wrote on 31 May 2020 09:32

Recipients:(name . Maxim Cournoyer)(address . maxim.cournoyer@gmail.com)

Message-ID:871rn04l04.fsf@ambrevar.xyz

Sure, but I don't know how popular Btrfs is, plus compression is a very

welcome addition to Guix (for all those SSD users with little space :p).

Granted, this is not very Guix specific...

Pierre Neidhardt

https://ambrevar.xyz/

Closed

Your comment

This issue is archived.

To comment on this conversation send an email to 37305@debbugs.gnu.org

To respond to this issue using the mumi CLI, first switch to it

mumi current 37305

Then, you may apply the latest patchset in this issue (with sign off)

mumi am -- -s

Or, compose a reply to this issue

mumi compose

Or, send patches to this issue

mumi send-email *.patch

is:open	open issues
is:done	closed issues
submitter:<who>	search issue submitter
author:<who>	search by message author
date:yesterday..now	search by issue date
mdate:3m..2d	search by message date