Re: [GIT PULL] nolibc for 6.11-rc1

[Paul E. McKenney]

On Sun, Jul 14, 2024 at 10:09:13AM +0200, Thomas Weißschuh wrote:
> On 2024-07-12 20:16:13+0000, Paul E. McKenney wrote:
> > On Sun, Jun 30, 2024 at 09:06:39AM -0700, Paul E. McKenney wrote:
> > > On Sat, Jun 29, 2024 at 01:04:08PM +0200, Thomas Weißschuh wrote:
> > > > Hi Paul,
> > > > 
> > > > The following changes since commit f2661062f16b2de5d7b6a5c42a9a5c96326b8454:
> > > > 
> > > >   Linux 6.10-rc5 (2024-06-23 17:08:54 -0400)
> > > > 
> > > > are available in the Git repository at:
> > > > 
> > > >   https://git.kernel.org/pub/scm/linux/kernel/git/nolibc/linux-nolibc.git tags/nolibc-20240629-for-6.11-1
> > > > 
> > > > for you to fetch changes up to 6ca8f2e20bd1ced8a7cd12b3ae4b1ceca85cfc2b:
> > > > 
> > > >   selftests: kselftest: also use strerror() on nolibc (2024-06-29 09:44:58 +0200)
> > > 
> > > Hearing no objections, I have pulled this in so that it will appear
> > > in the next -next.  Here are the test results:
> > > 
> > > make run:
> > > 195 test(s): 195 passed,   0 skipped,   0 failed => status: success
> > > 
> > > make run-user:
> > > 195 test(s): 193 passed,   2 skipped,   0 failed => status: warning
> > > 
> > > So looks good to me!
> 
> For testing you can use "./run-tests.sh -m [user | system]" to run the
> tests on all supported architectures via QEMU.
> 
> (On the first run you can use "-p" to download the toolchains)

Thank you for the info!

My near-term plan is that I do a smoke test on x86 (or whatever I am
running), and let you guys cover the various architectures.  Longer
term, I might get more into cross-architecture work.

> > And please see below for my proposed signed tag.  Please let me know of
> > any needed adjustments.
> > 
> > 							Thanx, Paul
> > 
> > ----------------------------------------------------------------
> > 
> > tag nolibc.2024.07.12a
> > Tagger: Paul E. McKenney <paulmck@xxxxxxxxxx>
> > Date:   Fri Jul 12 16:56:21 2024 -0700
> > 
> > nolibc updates for v6.11
> > 
> > o	Fix selftest printf format mismatch in expect_str_buf_eq()
> 
> Period at the end.

Good eyes, thank you, fixed.

> > o	Stop using brk() and sbrk() when testing against musl, which
> > 	implements these two functions with ENOMEM.
> > 
> > o	Make tests us -Werror to force failure on compiler warnings.
> > 
> > o	Add limits for the {u,}intmax_t, ulong and {u,}llong types.
> > 
> > o	Implement strtol() and friends.
> > 
> > o	Add facility to skip nolibc-specific tests when running against
> > 	non-nolibc libraries.
> > 
> > o	Implement strerror().
> > 
> > o	Use strerror() unconditionally, instead of only when running
> > 	against non-nolibc libraries.
> 
> Maybe mention that this is about kselftest and not nolibc itself.
> 
> Otherwise looks good, thanks!

Thank you for looking this over, and does the updated version below
cover this?  If I am still off, please suggest updates.

						Thanx, Paul

------------------------------------------------------------------------

tag nolibc.2024.07.14a
Tagger: Paul E. McKenney <paulmck@xxxxxxxxxx>
Date:   Sun Jul 14 09:07:29 2024 -0700

nolibc updates for v6.11

o	Fix selftest printf format mismatch in expect_str_buf_eq().

o	Stop using brk() and sbrk() when testing against musl, which
	implements these two functions with ENOMEM.

o	Make tests use -Werror to force failure on compiler warnings.

o	Add limits for the {u,}intmax_t, ulong and {u,}llong types.

o	Implement strtol() and friends.

o	Add facility to skip nolibc-specific tests when running against
	non-nolibc libraries.

o	Implement strerror().

o	Use strerror() unconditionally, instead of only when running
	kselftest.
-----BEGIN PGP SIGNATURE-----
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=yaww
-----END PGP SIGNATURE-----

commit 6ca8f2e20bd1ced8a7cd12b3ae4b1ceca85cfc2b
Author: Thomas Weißschuh <linux@xxxxxxxxxxxxxx>
Date:   Fri Apr 26 13:08:58 2024 +0200

    selftests: kselftest: also use strerror() on nolibc
    
    nolibc gained an implementation of strerror() recently.
    Use it and drop the ifdeffery.
    
    Acked-by: Shuah Khan <skhan@xxxxxxxxxxxxxxxxxxx>
    Signed-off-by: Thomas Weißschuh <linux@xxxxxxxxxxxxxx>

diff --git a/tools/testing/selftests/kselftest.h b/tools/testing/selftests/kselftest.h
index 76c2a6945d3e..b8967b6e29d5 100644
--- a/tools/testing/selftests/kselftest.h
+++ b/tools/testing/selftests/kselftest.h
@@ -168,15 +168,7 @@ static inline __printf(1, 2) void ksft_print_msg(const char *msg, ...)
 
 static inline void ksft_perror(const char *msg)
 {
-#ifndef NOLIBC
 	ksft_print_msg("%s: %s (%d)\n", msg, strerror(errno), errno);
-#else
-	/*
-	 * nolibc doesn't provide strerror() and it seems
-	 * inappropriate to add one, just print the errno.
-	 */
-	ksft_print_msg("%s: %d)\n", msg, errno);
-#endif
 }
 
 static inline __printf(1, 2) void ksft_test_result_pass(const char *msg, ...)