linux-next: build failure after merge of the tip-fixes tree

linux-next: build failure after merge of the tip-fixes tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1070 bytes --]

Hi all,

After merging the tip-fixes tree, today's linux-next build (x86_64
allmodconfig) failed like this:

In file included from builtin-check.c:16:
In function 'save_argv',
    inlined from 'objtool_run' at builtin-check.c:296:2:
tools/objtool/include/objtool/warn.h:47:17: error: '%s' directive argument is null [-Werror=format-overflow=]
   47 |                 "%s%s%s: objtool: " format "\n",        \
      |                 ^~~~~~~~~~~~~~~~~~~
tools/objtool/include/objtool/warn.h:92:9: note: in expansion of macro 'WARN'
   92 |         WARN("%s: " format " failed: %s", __func__, ##__VA_ARGS__, strerror(errno))
      |         ^~~~
builtin-check.c:241:25: note: in expansion of macro 'WARN_GLIBC'
  241 |                         WARN_GLIBC("strdup(%s)", orig_argv[i]);
      |                         ^~~~~~~~~~
cc1: all warnings being treated as errors

Caused (or exposed?) by commit

  c5995abe1547 ("objtool: Improve error handling")

I have used the tip-fixes tree from next-20250325 for today.

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: build failure after merge of the tip-fixes tree

[Josh Poimboeuf]

On Wed, Mar 26, 2025 at 10:38:54AM +1100, Stephen Rothwell wrote:
> Hi all,
> 
> After merging the tip-fixes tree, today's linux-next build (x86_64
> allmodconfig) failed like this:
> 
> In file included from builtin-check.c:16:
> In function 'save_argv',
>     inlined from 'objtool_run' at builtin-check.c:296:2:
> tools/objtool/include/objtool/warn.h:47:17: error: '%s' directive argument is null [-Werror=format-overflow=]
>    47 |                 "%s%s%s: objtool: " format "\n",        \
>       |                 ^~~~~~~~~~~~~~~~~~~
> tools/objtool/include/objtool/warn.h:92:9: note: in expansion of macro 'WARN'
>    92 |         WARN("%s: " format " failed: %s", __func__, ##__VA_ARGS__, strerror(errno))
>       |         ^~~~
> builtin-check.c:241:25: note: in expansion of macro 'WARN_GLIBC'
>   241 |                         WARN_GLIBC("strdup(%s)", orig_argv[i]);
>       |                         ^~~~~~~~~~
> cc1: all warnings being treated as errors
> 
> Caused (or exposed?) by commit
> 
>   c5995abe1547 ("objtool: Improve error handling")
> 
> I have used the tip-fixes tree from next-20250325 for today.

Thanks, I'll fix this up.  With my GCC 14.2.1 I get the same error with
that commit, but for some reason the error goes away with the next
commit.

-- 
Josh

linux-next: build failure after merge of the tip-fixes tree

[Mark Brown]

Hi all,

After merging the tip-fixes tree, today's linux-next build (arm64 kunit)
failed like this:

[12:34:18] Configuring KUnit Kernel ...
[12:34:18] Building KUnit Kernel ...
Populating config with:
$ make ARCH=arm64 O=/tmp/next/arm64_kunit olddefconfig CROSS_COMPILE=aarch64-linux-gnu-
Building with:
$ make all compile_commands.json scripts_gdb ARCH=arm64 O=/tmp/next/arm64_kunit --jobs=192 CROSS_COMPILE=aarch64-linux-gnu-
[12:34:34] Starting KUnit Kernel (1/1)...
Running tests with:
$ qemu-system-aarch64 -nodefaults -m 1024 -kernel /tmp/next/arm64_kunit/arch/arm64/boot/Image.gz -append 'kunit.enable=1 console=ttyAMA0 kunit_shutdown=reboot' -no-reboot -nographic -accel kvm -accel hvf -accel tcg -serial stdio -machine virt -cpu max
Command '['qemu-system-aarch64', '-nodefaults', '-m', '1024', '-kernel', '/tmp/next/arm64_kunit/arch/arm64/boot/Image.gz', '-append', 'kunit.enable=1 console=ttyAMA0 kunit_shutdown=reboot', '-no-reboot', '-nographic', '-accel', 'kvm', '-accel', 'hvf', '-accel', 'tcg', '-serial', 'stdio', '-machine', 'virt', '-cpu', 'max']' timed out after 300 seconds
[12:39:34] [FAILED] <missing>
[12:39:34] Testing complete. Ran 0 tests: errors: 1
[12:39:34] Test results stored in /tmp/next/arm64_kunit.json
[12:39:34] Elapsed time: 316.083s total, 0.002s configuring, 16.074s building, 300.006s running

which isn't super helpful.  I haven't isolated which commit but it'll be
one of the three:

  b81207470caf ("clockevents: Prevent timer interrupt starvation")
  82b915051d32 ("tick/nohz: Fix inverted return value in check_tick_dependency() fast path")
  105c42566a55 ("ata: ahci: force 32-bit DMA for JMicron JMB582/JMB585")

which are new since yesterday, I'm guessing not the latter.  I have used
the tree from next-20260407 instead.

Re: linux-next: build failure after merge of the tip-fixes tree

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 604 bytes --]

On Wed, Apr 08, 2026 at 01:44:08PM +0100, Mark Brown wrote:

> After merging the tip-fixes tree, today's linux-next build (arm64 kunit)
> failed like this:

> [12:39:34] [FAILED] <missing>
> [12:39:34] Testing complete. Ran 0 tests: errors: 1
> [12:39:34] Test results stored in /tmp/next/arm64_kunit.json
> [12:39:34] Elapsed time: 316.083s total, 0.002s configuring, 16.074s building, 300.006s running

> which isn't super helpful.  I haven't isolated which commit but it'll be
> one of the three:
> 
>   b81207470caf ("clockevents: Prevent timer interrupt starvation")

It's this one.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: build failure after merge of the tip-fixes tree

[Thomas Weißschuh]

On Wed, Apr 08, 2026 at 03:50:44PM +0100, Mark Brown wrote:
> On Wed, Apr 08, 2026 at 01:44:08PM +0100, Mark Brown wrote:
> 
> > After merging the tip-fixes tree, today's linux-next build (arm64 kunit)
> > failed like this:
> 
> > [12:39:34] [FAILED] <missing>
> > [12:39:34] Testing complete. Ran 0 tests: errors: 1
> > [12:39:34] Test results stored in /tmp/next/arm64_kunit.json
> > [12:39:34] Elapsed time: 316.083s total, 0.002s configuring, 16.074s building, 300.006s running
> 
> > which isn't super helpful.  I haven't isolated which commit but it'll be
> > one of the three:
> > 
> >   b81207470caf ("clockevents: Prevent timer interrupt starvation")

Should be fixed by:

--- a/kernel/time/clockevents.c
+++ b/kernel/time/clockevents.c
@@ -369,7 +369,7 @@ int clockevents_program_event(struct clock_event_device *dev, ktime_t expires, b
        if (dev->next_event_forced)
                return 0;
 
-       if (dev->set_next_event(dev->min_delta_ticks, dev)) {
+       if (dev->set_next_event(dev->min_delta_ns, dev)) {
                if (!force || clockevents_program_min_delta(dev))
                        return -ETIME;
        }


https://lore.kernel.org/lkml/20260408155353-42aeefa4-db66-48aa-ab07-0538a8cfdbf0@linutronix.de/


Thomas

Re: linux-next: build failure after merge of the tip-fixes tree

[Thomas Gleixner]

On Wed, Apr 08 2026 at 13:44, Mark Brown wrote:

It's a boot not a build fail, right?

> After merging the tip-fixes tree, today's linux-next build (arm64 kunit)
> failed like this:
> which isn't super helpful.  I haven't isolated which commit but it'll be
> one of the three:
>
>   b81207470caf ("clockevents: Prevent timer interrupt starvation")
>   82b915051d32 ("tick/nohz: Fix inverted return value in check_tick_dependency() fast path")
>   105c42566a55 ("ata: ahci: force 32-bit DMA for JMicron JMB582/JMB585")

Found it with qemu and pushed out an updated version. Delta patch
below. It restores the behaviour for !force and delta <= 0 which is
required under certain circumstances to get through the tick_periodic()
boot stage.

Thanks,

        tglx
---
--- a/kernel/time/clockevents.c
+++ b/kernel/time/clockevents.c
@@ -324,6 +324,8 @@ int clockevents_program_event(struct clo
 		return dev->set_next_ktime(expires, dev);
 
 	delta = ktime_to_ns(ktime_sub(expires, ktime_get()));
+	if (delta <= 0 && !force)
+		return -ETIME;
 
 	if (delta > (int64_t)dev->min_delta_ns) {
 		delta = min(delta, (int64_t) dev->max_delta_ns);

Re: linux-next: build failure after merge of the tip-fixes tree

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 769 bytes --]

On Wed, Apr 08, 2026 at 05:21:29PM +0200, Thomas Gleixner wrote:
> On Wed, Apr 08 2026 at 13:44, Mark Brown wrote:
> 
> It's a boot not a build fail, right?

Yes, sorry - the -next tooling understands running kunit as a build
command and I need to remember to edit that bit of the message.

> >   105c42566a55 ("ata: ahci: force 32-bit DMA for JMicron JMB582/JMB585")

> Found it with qemu and pushed out an updated version. Delta patch
> below. It restores the behaviour for !force and delta <= 0 which is
> required under certain circumstances to get through the tick_periodic()
> boot stage.

Thanks, hopefully everything will go through smoothly tomorrow.  As
mentioned on IRC I'm also holding tip at yesterday's version since it's
also affected.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]