[PATCH 00/21] Add riscv64 support

> * guix/utils.scm (target-riscv?): New predicate.

> ---

> guix/utils.scm | 6 ++++++

> 1 file changed, 6 insertions(+)

>

> diff --git a/guix/utils.scm b/guix/utils.scm

> index 32fcff72ea..134879feb1 100644

> --- a/guix/utils.scm

> +++ b/guix/utils.scm

> @@ -96,6 +96,7 @@

> target-arm?

> target-ppc32?

> target-powerpc?

> + target-riscv?

> target-64bit?

> cc-for-target

> cxx-for-target

> @@ -699,6 +700,11 @@ architecture (x86_64)?"

> (%current-system))))

> (string-prefix? "powerpc" target))

>

> +(define* (target-riscv? #:optional (target (or (%current-target-system)

> + (%current-system))))

> + "Is the architecture of TARGET a 'riscv' architecture variant?"

> + (string-prefix? "riscv" target))

> +

> (define* (target-64bit? #:optional (system (or (%current-target-system)

> (%current-system))))

> (any (cut string-prefix? <> system) '("x86_64" "aarch64" "mips64" "powerpc64")))

> > +(define* (target-riscv? #:optional (target (or (%current-target-system)

> > + (%current-system))))

> > + "Is the architecture of TARGET a 'riscv' architecture variant?"

> > + (string-prefix? "riscv" target))

> > +

> > (define* (target-64bit? #:optional (system (or (%current-target-system)

> > (%current-system))))

> > (any (cut string-prefix? <> system) '("x86_64" "aarch64" "mips64" "powerpc64")))

>

> The next patch adds a ‘riscv64-linux-gnu’ target. riscv64 seems 64-bit to me.

> It would seem riscv64 needs to be added to target-64bit?.

> Greetings,

> Maxime.

> Efraim Flashner schreef op di 17-08-2021 om 13:18 [+0300]:

> > * guix/utils.scm (target-riscv?): New predicate.

> > ---

> > guix/utils.scm | 6 ++++++

> > 1 file changed, 6 insertions(+)

> >

> > diff --git a/guix/utils.scm b/guix/utils.scm

> > index 32fcff72ea..134879feb1 100644

> > --- a/guix/utils.scm

> > +++ b/guix/utils.scm

> > @@ -96,6 +96,7 @@

> > target-arm?

> > target-ppc32?

> > target-powerpc?

> > + target-riscv?

> > target-64bit?

> > cc-for-target

> > cxx-for-target

> > @@ -699,6 +700,11 @@ architecture (x86_64)?"

> > (%current-system))))

> > (string-prefix? "powerpc" target))

> >

> > +(define* (target-riscv? #:optional (target (or (%current-target-system)

> > + (%current-system))))

> > + "Is the architecture of TARGET a 'riscv' architecture variant?"

> > + (string-prefix? "riscv" target))

> > +

> > (define* (target-64bit? #:optional (system (or (%current-target-system)

> > (%current-system))))

> > (any (cut string-prefix? <> system) '("x86_64" "aarch64" "mips64" "powerpc64")))

>

> The next patch adds a ‘riscv64-linux-gnu’ target. riscv64 seems 64-bit to me.

> It would seem riscv64 needs to be added to target-64bit?.

>

> Greetings,

> Maxime.

> * gnu/packages/commencement.scm (%boot3-inputs): Add gcc:lib.

> ---

> gnu/packages/commencement.scm | 3 +++

> 1 file changed, 3 insertions(+)

>

> diff --git a/gnu/packages/commencement.scm b/gnu/packages/commencement.scm

> index 69cf30483d..e2339170e5 100644

> --- a/gnu/packages/commencement.scm

> +++ b/gnu/packages/commencement.scm

> @@ -3579,6 +3579,9 @@ exec ~a/bin/~a-~a -B~a/lib -Wl,-dynamic-linker -Wl,~a/~a \"$@\"~%"

> (define (%boot3-inputs)

> ;; 4th stage inputs.

> `(("gcc" ,gcc-final)

> + ,@(if (target-riscv?)

> + `(("gcc:lib" ,gcc-final "lib"))

> + '())

> ("ld-wrapper" ,ld-wrapper-boot3)

> ,@(alist-delete "gcc" (%boot2-inputs))))

>

> --

> 2.32.0

>

> * gnu/packages/guile.scm (guile-3.0)[arguments]: On riscv64-linux add

> a phase to skip a failing test.

> ---

> gnu/packages/guile.scm | 22 +++++++++++++---------

> 1 file changed, 13 insertions(+), 9 deletions(-)

>

> diff --git a/gnu/packages/guile.scm b/gnu/packages/guile.scm

> index fffc30f913..f9a7125f83 100644

> --- a/gnu/packages/guile.scm

> +++ b/gnu/packages/guile.scm

> @@ -65,7 +65,8 @@

> #:use-module (guix deprecation)

> #:use-module (guix utils)

> #:use-module (ice-9 match)

> - #:use-module ((srfi srfi-1) #:prefix srfi-1:))

> + #:use-module ((srfi srfi-1) #:prefix srfi-1:)

> + #:use-module (srfi srfi-26))

>

> ;;; Commentary:

> ;;;

> @@ -346,14 +347,17 @@ without requiring the source code to be rewritten.")

> ;; https://debbugs.gnu.org/cgi/bugreport.cgi?bug=45214

> (substitute* "bootstrap/Makefile.in"

> (("^GUILE_OPTIMIZATIONS.*")

> - "GUILE_OPTIMIZATIONS = -O1 -Oresolve-primitives -Ocps\n"))))

> - (add-after 'unpack 'skip-failing-fdes-test

> - (lambda _

> - ;; ERROR: ((system-error "seek" "~A" ("Bad file descriptor") (9)))

> - (substitute* "test-suite/tests/ports.test"

> - (("fdes not closed\"" all) (string-append all "(exit 77)")))

> - #t)))

> - '())))))

> + "GUILE_OPTIMIZATIONS = -O1 -Oresolve-primitives -Ocps\n")))))

> + '())

> + ,@(if (srfi-1:any (cute string-prefix? <> (%current-system))

> + '("powerpc-" "riscv64-"))

> + `((add-after 'unpack 'skip-failing-fdes-test

> + (lambda _

> + ;; ERROR: ((system-error "seek" "~A" ("Bad file descriptor") (9)))

> + (substitute* "test-suite/tests/ports.test"

> + (("fdes not closed\"" all) (string-append all "(exit 77)")))

> + #t)))

> + '())))))

>

> (native-search-paths

> (list (search-path-specification

> --

> 2.32.0

>

> * gnu/packages/patches/libtool-skip-tests2.patch: Also skip pie tests

> on riscv architectures.

> ---

> gnu/packages/patches/libtool-skip-tests2.patch | 8 ++++----

> 1 file changed, 4 insertions(+), 4 deletions(-)

>

> diff --git a/gnu/packages/patches/libtool-skip-tests2.patch b/gnu/packages/patches/libtool-skip-tests2.patch

> index c9d61e3b41..b0a479325f 100644

> --- a/gnu/packages/patches/libtool-skip-tests2.patch

> +++ b/gnu/packages/patches/libtool-skip-tests2.patch

> @@ -1,4 +1,4 @@

> -Skip the nopic test on ARM and MIPS systems.

> +Skip the nopic test on ARM, MIPS and RISCV systems.

>

> --- libtool-2.4.6/tests/demo.at.orig 2015-01-16 13:52:04.000000000 -0500

> +++ libtool-2.4.6/tests/demo.at 2015-02-16 10:48:51.435851966 -0500

> @@ -7,7 +7,7 @@ Skip the nopic test on ARM and MIPS systems.

>

> AT_CHECK([case $host in

> -hppa*|x86_64*|s390*)

> -+hppa*|x86_64*|s390*|arm*|mips*)

> ++hppa*|x86_64*|s390*|arm*|mips*|riscv*)

> # These hosts cannot use non-PIC shared libs

> exit 77 ;;

> *-solaris*|*-sunos*)

> @@ -18,7 +18,7 @@ Skip the nopic test on ARM and MIPS systems.

> { set +x

> $as_echo "$at_srcdir/demo.at:535: case \$host in

> -hppa*|x86_64*|s390*)

> -+hppa*|x86_64*|s390*|arm*|mips*)

> ++hppa*|x86_64*|s390*|arm*|mips*|riscv*)

> # These hosts cannot use non-PIC shared libs

> exit 77 ;;

> *-solaris*|*-sunos*)

> @@ -27,7 +27,7 @@ Skip the nopic test on ARM and MIPS systems.

> at_fn_check_prepare_notrace 'a `...` command substitution' "demo.at:535"

> ( $at_check_trace; case $host in

> -hppa*|x86_64*|s390*)

> -+hppa*|x86_64*|s390*|arm*|mips*)

> ++hppa*|x86_64*|s390*|arm*|mips*|riscv*)

> # These hosts cannot use non-PIC shared libs

> exit 77 ;;

> *-solaris*|*-sunos*)

> --

> 2.32.0

>

> * gnu/packages/tex.scm (texlive-bin)[arguments]: Adjust configure-flags

> for riscv64-linux to skip luajit based binaries. Add phase to skip some

> tests.

> ---

> gnu/packages/tex.scm | 18 ++++++++++--------

> 1 file changed, 10 insertions(+), 8 deletions(-)

>

> diff --git a/gnu/packages/tex.scm b/gnu/packages/tex.scm

> index c00c2ae177..b96aadc357 100644

> --- a/gnu/packages/tex.scm

> +++ b/gnu/packages/tex.scm

> @@ -338,13 +338,14 @@ files from LOCATIONS with expected checksum HASH. CODE is not currently in use.

> "--with-system-teckit"

> "--with-system-zlib"

> "--with-system-zziplib"

> - ;; LuaJIT is not ported to powerpc64le* yet.

> - ,@(if (string-prefix? "powerpc64le" (or (%current-target-system)

> - (%current-system)))

> - '("--disable-luajittex"

> - "--disable-luajithbtex"

> - "--disable-mfluajit")

> - '()))

> + ;; LuaJIT is not ported to some architectures yet.

> + ,@(if (or (string-prefix? "powerpc64le" (or (%current-target-system)

> + (%current-system)))

> + (target-riscv?))

> + '("--disable-luajittex"

> + "--disable-luajithbtex"

> + "--disable-mfluajit")

> + '()))

>

> ;; Disable tests on some architectures to cope with a failure of

> ;; luajiterr.test.

> @@ -385,7 +386,8 @@ files from LOCATIONS with expected checksum HASH. CODE is not currently in use.

> (("^\\./omfonts -ofm2opl \\$srcdir/tests/check tests/xcheck \\|\\| exit 1")

> "./omfonts -ofm2opl $srcdir/tests/check tests/xcheck || exit 77"))

> #t))

> - ,@(if (target-ppc32?)

> + ,@(if (or (target-ppc32?)

> + (target-riscv?))

> ;; Some mendex tests fail on some architectures.

> `((add-after 'unpack 'skip-mendex-tests

> (lambda _

> --

> 2.32.0

>

> * gnu/packages/compression.scm (lz4)[native-inputs]: Don't use valgrind

> when building for riscv64-linux.

> ---

> gnu/packages/compression.scm | 4 +++-

> 1 file changed, 3 insertions(+), 1 deletion(-)

>

> diff --git a/gnu/packages/compression.scm b/gnu/packages/compression.scm

> index 93d517abd5..1803237e6e 100644

> --- a/gnu/packages/compression.scm

> +++ b/gnu/packages/compression.scm

> @@ -819,7 +819,9 @@ decompression of some loosely related file formats used by Microsoft.")

> (native-inputs

> `(;; For tests.

> ("python" ,python)

> - ("valgrind" ,valgrind)))

> + ,@(if (target-riscv?)

> + '()

> + `(("valgrind" ,valgrind)))))

> (arguments

> `(;; Not designed for parallel testing.

> ;; See https://github.com/lz4/lz4/issues/957#issuecomment-737419821

> --

> 2.32.0

>

> ;;; Commentary:

> ;;;

> @@ -346,14 +347,17 @@ without requiring the source code to be rewritten.")

> ;; https://debbugs.gnu.org/cgi/bugreport.cgi?bug=45214

> (substitute* "bootstrap/Makefile.in"

> (("^GUILE_OPTIMIZATIONS.*")

> - "GUILE_OPTIMIZATIONS = -O1 -Oresolve-primitives -Ocps\n"))))

> - (add-after 'unpack 'skip-failing-fdes-test

> - (lambda _

> - ;; ERROR: ((system-error "seek" "~A" ("Bad file descriptor") (9)))

> - (substitute* "test-suite/tests/ports.test"

> - (("fdes not closed\"" all) (string-append all "(exit 77)")))

> - #t)))

> - '())))))

> + "GUILE_OPTIMIZATIONS = -O1 -Oresolve-primitives -Ocps\n")))))

> + '())

> + ,@(if (srfi-1:any (cute string-prefix? <> (%current-system))

> + '("powerpc-" "riscv64-"))

> + `((add-after 'unpack 'skip-failing-fdes-test

> + (lambda _

> + ;; ERROR: ((system-error "seek" "~A" ("Bad file descriptor") (9)))

> + (substitute* "test-suite/tests/ports.test"

> + (("fdes not closed\"" all) (string-append all "(exit 77)")))

> + #t)))

> + '())))))

> * gnu/packages/patches/libtool-skip-tests2.patch: Also skip pie tests

> on riscv architectures.

> ---

> gnu/packages/patches/libtool-skip-tests2.patch | 8 ++++----

> 1 file changed, 4 insertions(+), 4 deletions(-)

>

> > ;;; Commentary:

> > ;;;

> > @@ -346,14 +347,17 @@ without requiring the source code to be rewritten.")

> > ;; https://debbugs.gnu.org/cgi/bugreport.cgi?bug=45214

> > (substitute* "bootstrap/Makefile.in"

> > (("^GUILE_OPTIMIZATIONS.*")

> > - "GUILE_OPTIMIZATIONS = -O1 -Oresolve-primitives -Ocps\n"))))

> > - (add-after 'unpack 'skip-failing-fdes-test

> > - (lambda _

> > - ;; ERROR: ((system-error "seek" "~A" ("Bad file descriptor") (9)))

> > - (substitute* "test-suite/tests/ports.test"

> > - (("fdes not closed\"" all) (string-append all "(exit 77)")))

> > - #t)))

> > - '())))))

> > + "GUILE_OPTIMIZATIONS = -O1 -Oresolve-primitives -Ocps\n")))))

> > + '())

> > + ,@(if (srfi-1:any (cute string-prefix? <> (%current-system))

> > + '("powerpc-" "riscv64-"))

> > + `((add-after 'unpack 'skip-failing-fdes-test

> > + (lambda _

> > + ;; ERROR: ((system-error "seek" "~A" ("Bad file descriptor") (9)))

> > + (substitute* "test-suite/tests/ports.test"

> > + (("fdes not closed\"" all) (string-append all "(exit 77)")))

> > + #t)))

> > + '())))))

>

> This is not powerpc- or riscv64-specific -- iirc, I sometimes encounter it on a x86_64.

> It's non-determenistic though. The debian patch has a nice explanation:

> <https://salsa.debian.org/rlb/deb-guile/-/blob/f24ab0150132d906b9724128576c36c39361cab7/debian/patches/0007-Fix-non-revealed-port-is-closed-ports.test.patch>.

> I've also linked some of the public reports on the issue together:

> <https://issues.guix.gnu.org/48389#11>.

>

> Could the patch (or substitution) be applied, independently of the architecture?

> Greetings,

> Maxime.

> Efraim Flashner schreef op di 17-08-2021 om 13:19 [+0300]:

> > * gnu/packages/patches/libtool-skip-tests2.patch: Also skip pie tests

> > on riscv architectures.

> > ---

> > gnu/packages/patches/libtool-skip-tests2.patch | 8 ++++----

> > 1 file changed, 4 insertions(+), 4 deletions(-)

> >

>

> Does upstream know about the issue?

> I would recommend including a link to the upstream bug report.

> (in libtool-skip-tests2.patch or in a comment in the 'libtool'

> package definition).

>

> Greetings,

> Maxime.

> * gnu/packages/dbm.scm (bdb)[arguments]: Modify configure to build for

> the currect architecture when building for riscv64-linux.

> ---

> gnu/packages/dbm.scm | 23 ++++++++++++-----------

> 1 file changed, 12 insertions(+), 11 deletions(-)

>

> diff --git a/gnu/packages/dbm.scm b/gnu/packages/dbm.scm

> index ce84ef9eda..c2bdaf782a 100644

> --- a/gnu/packages/dbm.scm

> +++ b/gnu/packages/dbm.scm

> @@ -1,7 +1,7 @@

> ;;; GNU Guix --- Functional package management for GNU

> ;;; Copyright © 2012, 2013, 2014, 2016, 2020 Ludovic Courtès

> <ludo@gnu.org> ;;; Copyright © 2013, 2015 Andreas Enge <andreas@enge.fr>

> -;;; Copyright © 2016, 2017, 2018, 2020 Efraim Flashner

> <efraim@flashner.co.il> +;;; Copyright © 2016, 2017, 2018, 2020, 2021

> Efraim Flashner <efraim@flashner.co.il> ;;; Copyright © 2017, 2018

> Marius Bakke <mbakke@fastmail.com>

> ;;; Copyright © 2018 Mark H Weaver <mhw@netris.org>

> ;;; Copyright © 2020 Jakub Kądziołka <kuba@kadziolka.net>

> @@ -28,7 +28,8 @@

> #:use-module (guix packages)

> #:use-module (guix download)

> #:use-module (guix build-system gnu)

> - #:use-module (guix utils))

> + #:use-module (guix utils)

> + #:use-module (ice-9 match))

>

> ;;; Commentary:

> ;;;

> @@ -72,15 +73,15 @@

> (string-append "CONFIG_SHELL=" (which "bash"))

> (string-append "SHELL=" (which "bash"))

>

> - ;; Bdb doesn't recognize aarch64 as an

> architecture. - ,@(if (string=? "aarch64-linux"

> (%current-system)) -

> '("--build=aarch64-unknown-linux-gnu") -

> '())

> -

> - ;; Bdb doesn't recognize powerpc64le as an

> architecture. - ,@(if (string=?

> "powerpc64le-linux" (%current-system)) -

> '("--build=powerpc64le-unknown-linux-gnu") -

> '())

> + ;; Bdb doesn't recognize very many architectures.

> + ,@(match (%current-system)

> + ("aarch64-linux"

> + '("--build=aarch64-unknown-linux-gnu"))

> + ("powerpc64le-linux"

> + '("--build=powerpc64le-unknown-linux-gnu"))

> + ("riscv64-linux"

> + '("--build=riscv64-unknown-linux-gnu"))

> + (_ '()))

>

> ,@(if (%current-target-system) ; cross

> building '((string-append "--host=" target))

> On Tue, Aug 17, 2021 at 01:19:16PM +0300, Efraim Flashner wrote:

> > * gnu/packages/tex.scm (texlive-bin)[arguments]: Adjust configure-flags

> > for riscv64-linux to skip luajit based binaries. Add phase to skip some

> > tests.

> > ---

> >

> > gnu/packages/tex.scm | 18 ++++++++++--------

> > 1 file changed, 10 insertions(+), 8 deletions(-)

> >

> > diff --git a/gnu/packages/tex.scm b/gnu/packages/tex.scm

> > index c00c2ae177..b96aadc357 100644

> > --- a/gnu/packages/tex.scm

> > +++ b/gnu/packages/tex.scm

> > @@ -338,13 +338,14 @@ files from LOCATIONS with expected checksum HASH.

> > CODE is not currently in use.>

> > "--with-system-teckit"

> > "--with-system-zlib"

> > "--with-system-zziplib"

> >

> > - ;; LuaJIT is not ported to powerpc64le* yet.

> > - ,@(if (string-prefix? "powerpc64le" (or

> > (%current-target-system) -

> > (%current-system))) - '("--disable-luajittex"

> > - "--disable-luajithbtex"

> > - "--disable-mfluajit")

> > - '()))

> > + ;; LuaJIT is not ported to some architectures yet.

> > + ,@(if (or (string-prefix? "powerpc64le" (or

> > (%current-target-system) +

> > (%current-system))) + (target-riscv?))

>

> I don't think I pushed the patch yet to add target-ppc64le? as an option

> yet, but I'll rewrite this one when I do.

> The hash seems to have changed.

> On Tue, Aug 17, 2021 at 01:19:18PM +0300, Efraim Flashner wrote:

> > * gnu/packages/compression.scm (lz4)[native-inputs]: Don't use valgrind

> > when building for riscv64-linux.

> > ---

> >

> > gnu/packages/compression.scm | 4 +++-

> > 1 file changed, 3 insertions(+), 1 deletion(-)

> >

> > diff --git a/gnu/packages/compression.scm

> > b/gnu/packages/compression.scm

> > index 93d517abd5..1803237e6e 100644

> > --- a/gnu/packages/compression.scm

> > +++ b/gnu/packages/compression.scm

> > @@ -819,7 +819,9 @@ decompression of some loosely related file formats

> > used by Microsoft.")>

> > (native-inputs

> >

> > `(;; For tests.

> >

> > ("python" ,python)

> >

> > - ("valgrind" ,valgrind)))

> > + ,@(if (target-riscv?)

> > + '()

> > + `(("valgrind" ,valgrind)))))

> >

> > (arguments

> >

> > `(;; Not designed for parallel testing.

> >

> > ;; See

> > https://github.com/lz4/lz4/issues/957#issuecomment-737419821

>

> Valgrind added support for riscv64-linux after the last release, but the

> patch isn't easy to backport. I've thought about adding a valgrind-next

> and pulling from a recent commit but I'm not sure it's actually worth

> it.

> * gnu/packages/tls.scm (openssl)[arguments]: Add phase for riscv64-linux

> to skip failing test.

> (target->openssl-target): Add case for riscv64-linux.

> ---

> gnu/packages/tls.scm | 11 ++++++++++-

> 1 file changed, 10 insertions(+), 1 deletion(-)

>

> diff --git a/gnu/packages/tls.scm b/gnu/packages/tls.scm

> index 4b8df04f17..f6cdeac363 100644

> --- a/gnu/packages/tls.scm

> +++ b/gnu/packages/tls.scm

> @@ -337,7 +337,10 @@ OpenSSL for TARGET."

> ((string-prefix? "powerpc64" target)

> "linux-ppc64")

> ((string-prefix? "powerpc" target)

> - "linux-ppc")))

> + "linux-ppc")

> + ((string-prefix? "riscv64" target)

> + ;; linux64-riscv64 isn't recognized until 3.0.0.

> + "linux-generic64")))

>

> (define-public openssl

> (package

> @@ -380,6 +383,12 @@ OpenSSL for TARGET."

> #$(target->openssl-target

> (%current-target-system))))))

> #~())

> + ;; Unclear why this test is failing.

> + #$@(if (target-riscv?)

> + #~((add-after 'unpack 'skip-failing-test

> + (lambda _

> + (delete-file "test/recipes/30-test_afalg.t"))))

> + #~())

> (replace 'configure

> (lambda* (#:key configure-flags #:allow-other-keys)

> (let* ((out #$output)

> Hi Efraim,

>

> Efraim Flashner <efraim@flashner.co.il> writes:

>

> > * gnu/packages/tls.scm (openssl)[arguments]: Add phase for riscv64-linux

> > to skip failing test.

> > (target->openssl-target): Add case for riscv64-linux.

> > ---

> > gnu/packages/tls.scm | 11 ++++++++++-

> > 1 file changed, 10 insertions(+), 1 deletion(-)

> >

> > diff --git a/gnu/packages/tls.scm b/gnu/packages/tls.scm

> > index 4b8df04f17..f6cdeac363 100644

> > --- a/gnu/packages/tls.scm

> > +++ b/gnu/packages/tls.scm

> > @@ -337,7 +337,10 @@ OpenSSL for TARGET."

> > ((string-prefix? "powerpc64" target)

> > "linux-ppc64")

> > ((string-prefix? "powerpc" target)

> > - "linux-ppc")))

> > + "linux-ppc")

> > + ((string-prefix? "riscv64" target)

> > + ;; linux64-riscv64 isn't recognized until 3.0.0.

> > + "linux-generic64")))

> >

> > (define-public openssl

> > (package

> > @@ -380,6 +383,12 @@ OpenSSL for TARGET."

> > #$(target->openssl-target

> > (%current-target-system))))))

> > #~())

> > + ;; Unclear why this test is failing.

> > + #$@(if (target-riscv?)

> > + #~((add-after 'unpack 'skip-failing-test

> > + (lambda _

> > + (delete-file "test/recipes/30-test_afalg.t"))))

> > + #~())

> > (replace 'configure

> > (lambda* (#:key configure-flags #:allow-other-keys)

> > (let* ((out #$output)

>

> This test is also failing on aarch64 [0] and has been reported upstream

> for some time [1] with no action.

> Looks like it can also be disabled with

>

> make TESTS=-test_afalg test

>

> if that's more desirable.

> [0] https://ci.guix.gnu.org/build/29270/details

> [1] https://github.com/openssl/openssl/issues/12242

>

> --

> Sarah

> Hi,

>

> Em terça-feira, 17 de agosto de 2021, às 07:19:17 -03, Efraim Flashner

> escreveu:

> > The hash seems to have changed.

>

> That’s true. I get the same hash as the new one in this patch when I

> manually download ‘updmap.cfg’.

>

> How did you notice it? I didn’t get an error in my (admittedly basic) tests

> of the TeX Live packages.

> Thanks for fixing it.

> --

> Thanks,

> Thiago

>

>

> Hello,

>

> Em terça-feira, 17 de agosto de 2021, às 07:35:35 -03, Efraim Flashner

> escreveu:

> > On Tue, Aug 17, 2021 at 01:19:18PM +0300, Efraim Flashner wrote:

> > > * gnu/packages/compression.scm (lz4)[native-inputs]: Don't use valgrind

> > > when building for riscv64-linux.

> > > ---

> > >

> > > gnu/packages/compression.scm | 4 +++-

> > > 1 file changed, 3 insertions(+), 1 deletion(-)

> > >

> > > diff --git a/gnu/packages/compression.scm

> > > b/gnu/packages/compression.scm

> > > index 93d517abd5..1803237e6e 100644

> > > --- a/gnu/packages/compression.scm

> > > +++ b/gnu/packages/compression.scm

> > > @@ -819,7 +819,9 @@ decompression of some loosely related file formats

> > > used by Microsoft.")>

> > > (native-inputs

> > >

> > > `(;; For tests.

> > >

> > > ("python" ,python)

> > >

> > > - ("valgrind" ,valgrind)))

> > > + ,@(if (target-riscv?)

> > > + '()

> > > + `(("valgrind" ,valgrind)))))

> > >

> > > (arguments

> > >

> > > `(;; Not designed for parallel testing.

> > >

> > > ;; See

> > > https://github.com/lz4/lz4/issues/957#issuecomment-737419821

> >

> > Valgrind added support for riscv64-linux after the last release, but the

> > patch isn't easy to backport. I've thought about adding a valgrind-next

> > and pulling from a recent commit but I'm not sure it's actually worth

> > it.

>

> IMHO, either the commit message or the code would benefit from a note

> mentioning that the current version of Valgrind doesn’t support

> riscv64-linux.

>

> On Tue, Aug 17, 2021 at 12:23:27PM -0300, Thiago Jung Bauermann wrote:

> > Hi,

> >

> > Em terça-feira, 17 de agosto de 2021, às 07:19:17 -03, Efraim Flashner

> > escreveu:

> > > The hash seems to have changed.

> >

> > That’s true. I get the same hash as the new one in this patch when I

> > manually download ‘updmap.cfg’.

> >

> > How did you notice it? I didn’t get an error in my (admittedly basic) tests

> > of the TeX Live packages.

>

> On some of my machines I build without substitutes, so when this one

> changed upstream it showed up.

>

> > Thanks for fixing it.

>

> Don't thank me just yet, I still need to see what changed between the

> two versions.

>

> On Tue, Aug 17, 2021 at 12:33:45PM -0700, Sarah Morgensen wrote:

> > Hi Efraim,

> >

> > Efraim Flashner <efraim@flashner.co.il> writes:

> >

> > > * gnu/packages/tls.scm (openssl)[arguments]: Add phase for riscv64-linux

> > > to skip failing test.

> > > (target->openssl-target): Add case for riscv64-linux.

> > > ---

> > > gnu/packages/tls.scm | 11 ++++++++++-

> > > 1 file changed, 10 insertions(+), 1 deletion(-)

> > >

> > > diff --git a/gnu/packages/tls.scm b/gnu/packages/tls.scm

> > > index 4b8df04f17..f6cdeac363 100644

> > > --- a/gnu/packages/tls.scm

> > > +++ b/gnu/packages/tls.scm

> > > @@ -337,7 +337,10 @@ OpenSSL for TARGET."

> > > ((string-prefix? "powerpc64" target)

> > > "linux-ppc64")

> > > ((string-prefix? "powerpc" target)

> > > - "linux-ppc")))

> > > + "linux-ppc")

> > > + ((string-prefix? "riscv64" target)

> > > + ;; linux64-riscv64 isn't recognized until 3.0.0.

> > > + "linux-generic64")))

> > >

> > > (define-public openssl

> > > (package

> > > @@ -380,6 +383,12 @@ OpenSSL for TARGET."

> > > #$(target->openssl-target

> > > (%current-target-system))))))

> > > #~())

> > > + ;; Unclear why this test is failing.

> > > + #$@(if (target-riscv?)

> > > + #~((add-after 'unpack 'skip-failing-test

> > > + (lambda _

> > > + (delete-file "test/recipes/30-test_afalg.t"))))

> > > + #~())

> > > (replace 'configure

> > > (lambda* (#:key configure-flags #:allow-other-keys)

> > > (let* ((out #$output)

> >

> > This test is also failing on aarch64 [0] and has been reported upstream

> > for some time [1] with no action.

>

> I was able to build openssl without any problems on my pine64, I'll run

> it again and see what I come up with.

>

> Looking at the bug report, it looks like it also might be kernel

> dependant and/or related to cross compiling. I'm pretty sure the aarch64

> builders are running Guix System, and my pine64 is running Armbian.

>

> I'll go ahead and disable it for all the arm* architectures.

>

> > Looks like it can also be disabled with

> >

> > make TESTS=-test_afalg test

> >

> > if that's more desirable.

>

> I do like that better.

>

> > [0] https://ci.guix.gnu.org/build/29270/details

> > [1] https://github.com/openssl/openssl/issues/12242

> >

> On 7d93b21ab1c132990054372a9677c1639d54e631

> gnu: glibc-for-bootstrap: Update patch.

>

> Run

> ./pre-inst-env guix build --target=riscv64-linux-gnu bootstrap-tarballs

>

> Producing

> /gnu/store/4hdzva9i0wyyfbgj1lmqc1wkk644pv07-bootstrap-tarballs-0

>

> With guix hash -rx

> 1nj0fdgj08bbmfny01mp2blv7c3p2iciqh31zmf04ap5s7ygsqlp

> Hello Efraim,

>

> Em terça-feira, 17 de agosto de 2021, às 07:19:05 -03, Efraim Flashner

> escreveu:

> > * gnu/packages/dbm.scm (bdb)[arguments]: Modify configure to build for

> > the currect architecture when building for riscv64-linux.

> > ---

> > gnu/packages/dbm.scm | 23 ++++++++++++-----------

> > 1 file changed, 12 insertions(+), 11 deletions(-)

> >

> > diff --git a/gnu/packages/dbm.scm b/gnu/packages/dbm.scm

> > index ce84ef9eda..c2bdaf782a 100644

> > --- a/gnu/packages/dbm.scm

> > +++ b/gnu/packages/dbm.scm

> > @@ -1,7 +1,7 @@

> > ;;; GNU Guix --- Functional package management for GNU

> > ;;; Copyright © 2012, 2013, 2014, 2016, 2020 Ludovic Courtès

> > <ludo@gnu.org> ;;; Copyright © 2013, 2015 Andreas Enge <andreas@enge.fr>

> > -;;; Copyright © 2016, 2017, 2018, 2020 Efraim Flashner

> > <efraim@flashner.co.il> +;;; Copyright © 2016, 2017, 2018, 2020, 2021

> > Efraim Flashner <efraim@flashner.co.il> ;;; Copyright © 2017, 2018

> > Marius Bakke <mbakke@fastmail.com>

> > ;;; Copyright © 2018 Mark H Weaver <mhw@netris.org>

> > ;;; Copyright © 2020 Jakub Kądziołka <kuba@kadziolka.net>

> > @@ -28,7 +28,8 @@

> > #:use-module (guix packages)

> > #:use-module (guix download)

> > #:use-module (guix build-system gnu)

> > - #:use-module (guix utils))

> > + #:use-module (guix utils)

> > + #:use-module (ice-9 match))

> >

> > ;;; Commentary:

> > ;;;

> > @@ -72,15 +73,15 @@

> > (string-append "CONFIG_SHELL=" (which "bash"))

> > (string-append "SHELL=" (which "bash"))

> >

> > - ;; Bdb doesn't recognize aarch64 as an

> > architecture. - ,@(if (string=? "aarch64-linux"

> > (%current-system)) -

> > '("--build=aarch64-unknown-linux-gnu") -

> > '())

> > -

> > - ;; Bdb doesn't recognize powerpc64le as an

> > architecture. - ,@(if (string=?

> > "powerpc64le-linux" (%current-system)) -

> > '("--build=powerpc64le-unknown-linux-gnu") -

> > '())

> > + ;; Bdb doesn't recognize very many architectures.

> > + ,@(match (%current-system)

> > + ("aarch64-linux"

> > + '("--build=aarch64-unknown-linux-gnu"))

> > + ("powerpc64le-linux"

> > + '("--build=powerpc64le-unknown-linux-gnu"))

> > + ("riscv64-linux"

> > + '("--build=riscv64-unknown-linux-gnu"))

> > + (_ '()))

> >

> > ,@(if (%current-target-system) ; cross

> > building '((string-append "--host=" target))

>

> Can this be fixed instead by updating ‘config.guess’ and ‘config.sub’ as

> done in https://issues.guix.gnu.org/50086#1 ?

>

> That could possibly even fix cross-building (as it does for ‘pth’) and

> eliminate the need for all the “,@” forms in the ‘configure’ phase.

>

> On Tue, Aug 17, 2021 at 11:58:44AM -0300, Thiago Jung Bauermann wrote:

> > Hello Efraim,

> >

> > Em terça-feira, 17 de agosto de 2021, às 07:19:05 -03, Efraim Flashner

> >

> > escreveu:

> > > @@ -72,15 +73,15 @@

> > >

> > > (string-append "CONFIG_SHELL=" (which

> > > "bash"))

> > > (string-append "SHELL=" (which "bash"))

> > >

> > > - ;; Bdb doesn't recognize aarch64 as an

> > > architecture. - ,@(if (string=? "aarch64-linux"

> > > (%current-system)) -

> > > '("--build=aarch64-unknown-linux-gnu") -

> > > '())

> > > -

> > > - ;; Bdb doesn't recognize powerpc64le as an

> > > architecture. - ,@(if (string=?

> > > "powerpc64le-linux" (%current-system)) -

> > > '("--build=powerpc64le-unknown-linux-gnu") -

> > > '())

> > > + ;; Bdb doesn't recognize very many

> > > architectures. + ,@(match (%current-system)

> > > + ("aarch64-linux"

> > > + '("--build=aarch64-unknown-linux-gnu"))

> > > + ("powerpc64le-linux"

> > > +

> > > '("--build=powerpc64le-unknown-linux-gnu"))

> > > + ("riscv64-linux"

> > > + '("--build=riscv64-unknown-linux-gnu"))

> > > + (_ '()))

> > >

> > > ,@(if (%current-target-system) ;

> > > cross

> > >

> > > building '((string-append "--host=" target))

> >

> > Can this be fixed instead by updating ‘config.guess’ and ‘config.sub’

> > as

> > done in https://issues.guix.gnu.org/50086#1 ?

> >

> > That could possibly even fix cross-building (as it does for ‘pth’) and

> > eliminate the need for all the “,@” forms in the ‘configure’ phase.

>

> Sorry for taking so long to get back to you.

> I finally tried this but there seems to be a cycle somewhere between bdb

> and config, so using config and dropping the custom --build flag isn't

> currently an option.