[GIT PULL] objtool fixes and updates

[GIT PULL] objtool fixes and updates

[Ingo Molnar]

Linus,

Please pull the latest objtool/urgent Git tree from:

   git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git objtool-urgent-2025-03-28

   # HEAD: ae958b12940bcd4ffa32c44684e4f2878bc5e140 objtool, drm/vmwgfx: Don't ignore vmw_send_msg() for ORC

[ Merge note: not all driver fixes below have maintainer acks. ]

These are objtool fixes and updates centered around the fallout
from the new CONFIG_OBJTOOL_WERROR=y feature, which, despite its
default-off nature, increased the profile/impact of objtool warnings:

 - Improve error handling and the presentation of warnings/errors.

 - In particular reduce the verbosity of objtool warnings if a
   function triggers a check already - it's unlikely that followup
   warnings are useful after the first warning is triggered.

 - Fix a number of objtool warnings in various drivers, core kernel
   code and architecture code. About half of them are potential
   problems related to out-of-bounds accesses or potential undefined
   behavior, the other half are additional objtool annotations.

 - Update objtool to latest (known) compiler quirks and
   objtool bugs triggered by compiler code generation

 Thanks,

	Ingo

------------------>
Josh Poimboeuf (25):
      objtool: Fix detection of consecutive jump tables on Clang 20
      objtool: Warn when disabling unreachable warnings
      objtool: Ignore entire functions rather than instructions
      objtool: Fix X86_FEATURE_SMAP alternative handling
      objtool: Fix CONFIG_OBJTOOL_WERROR for vmlinux.o
      objtool: Fix init_module() handling
      objtool: Silence more KCOV warnings
      objtool: Properly disable uaccess validation
      objtool: Improve error handling
      objtool: Reduce CONFIG_OBJTOOL_WERROR verbosity
      objtool: Fix up some outdated references to ENTRY/ENDPROC
      objtool: Remove --no-unreachable for noinstr-only vmlinux.o runs
      objtool: Remove redundant opts.noinstr dependency
      objtool, spi: amd: Fix out-of-bounds stack access in amd_set_spi_freq()
      objtool, nvmet: Fix out-of-bounds stack access in nvmet_ctrl_state_show()
      objtool, media: dib8000: Prevent divide-by-zero in dib8000_set_dds()
      objtool, panic: Disable SMAP in __stack_chk_fail()
      objtool, Input: cyapa - Remove undefined behavior in cyapa_update_fw_store()
      objtool, ASoC: codecs: wcd934x: Remove potential undefined behavior in wcd934x_slim_irq_handler()
      objtool, regulator: rk808: Remove potential undefined behavior in rk806_set_mode_dcdc()
      objtool, lkdtm: Obfuscate the do_nothing() pointer
      objtool: Fix NULL printf() '%s' argument in builtin-check.c:save_argv()
      objtool: Fix segfault in ignore_unreachable_insn()
      objtool: Fix STACK_FRAME_NON_STANDARD for cold subfunctions
      objtool, drm/vmwgfx: Don't ignore vmw_send_msg() for ORC


 arch/x86/include/asm/arch_hweight.h     |   6 +-
 arch/x86/include/asm/smap.h             |  23 +-
 arch/x86/include/asm/xen/hypercall.h    |   6 +-
 drivers/gpu/drm/vmwgfx/vmwgfx_msg.c     |   2 +-
 drivers/input/mouse/cyapa.c             |   4 +-
 drivers/media/dvb-frontends/dib8000.c   |   5 +-
 drivers/misc/lkdtm/perms.c              |  14 +-
 drivers/nvme/target/debugfs.c           |   2 +-
 drivers/regulator/rk808-regulator.c     |   4 +-
 drivers/spi/spi-amd.c                   |   2 +-
 include/linux/linkage.h                 |   4 -
 include/linux/objtool.h                 |   2 +-
 kernel/panic.c                          |   6 +
 scripts/Makefile.lib                    |   2 +-
 scripts/Makefile.vmlinux_o              |  15 +-
 sound/soc/codecs/wcd934x.c              |   2 +-
 tools/objtool/Documentation/objtool.txt |  10 +-
 tools/objtool/arch/x86/special.c        |  38 +--
 tools/objtool/builtin-check.c           | 146 +++++----
 tools/objtool/check.c                   | 527 ++++++++++++++++----------------
 tools/objtool/elf.c                     |  28 +-
 tools/objtool/include/objtool/builtin.h |   6 +-
 tools/objtool/include/objtool/check.h   |   3 +-
 tools/objtool/include/objtool/elf.h     |  28 +-
 tools/objtool/include/objtool/objtool.h |   2 +-
 tools/objtool/include/objtool/special.h |   4 +-
 tools/objtool/include/objtool/warn.h    |  13 +-
 tools/objtool/objtool.c                 |  11 +-
 tools/objtool/special.c                 |  12 +-
 29 files changed, 472 insertions(+), 455 deletions(-)

Re: [GIT PULL] objtool fixes and updates

[Borislav Petkov]

On Fri, Mar 28, 2025 at 10:18:57PM +0100, Ingo Molnar wrote:
> Linus,
> 
> Please pull the latest objtool/urgent Git tree from:
> 
>    git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git objtool-urgent-2025-03-28
> 
>    # HEAD: ae958b12940bcd4ffa32c44684e4f2878bc5e140 objtool, drm/vmwgfx: Don't ignore vmw_send_msg() for ORC
> 
> [ Merge note: not all driver fixes below have maintainer acks. ]

Btw, test bot complains:

https://lore.kernel.org/r/202503292202.Sge7ZEUc-lkp@intel.com

Date: Sat, 29 Mar 2025 22:29:12 +0800
From: kernel test robot <lkp@intel.com>
To: x86-ml <x86@kernel.org>
Cc: linux-kernel@vger.kernel.org
Subject: [tip:objtool/urgent] BUILD REGRESSION ae958b12940bcd4ffa32c44684e4f2878bc5e140

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: [GIT PULL] objtool fixes and updates

[Linus Torvalds]

On Sat, 29 Mar 2025 at 08:33, Borislav Petkov <bp@alien8.de> wrote:
>
> Btw, test bot complains:
>
> https://lore.kernel.org/r/202503292202.Sge7ZEUc-lkp@intel.com

That's not a very helpful error message, but I've dropped this pull
request for now. Ingo?

             Linus

Re: [GIT PULL] objtool fixes and updates

[Borislav Petkov]

On Sun, Mar 30, 2025 at 03:19:40PM -0700, Linus Torvalds wrote:
> On Sat, 29 Mar 2025 at 08:33, Borislav Petkov <bp@alien8.de> wrote:
> >
> > Btw, test bot complains:
> >
> > https://lore.kernel.org/r/202503292202.Sge7ZEUc-lkp@intel.com
> 
> That's not a very helpful error message

I found this:

https://lore.kernel.org/r/202503280703.OARM8SrY-lkp@intel.com

which looks like the original report.

Looks unsolved yet...

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: [GIT PULL] objtool fixes and updates

[Josh Poimboeuf]

On Mon, Mar 31, 2025 at 01:13:55AM +0200, Borislav Petkov wrote:
> On Sun, Mar 30, 2025 at 03:19:40PM -0700, Linus Torvalds wrote:
> > On Sat, 29 Mar 2025 at 08:33, Borislav Petkov <bp@alien8.de> wrote:
> > >
> > > Btw, test bot complains:
> > >
> > > https://lore.kernel.org/r/202503292202.Sge7ZEUc-lkp@intel.com
> > 
> > That's not a very helpful error message
> 
> I found this:
> 
> https://lore.kernel.org/r/202503280703.OARM8SrY-lkp@intel.com
> 
> which looks like the original report.
> 
> Looks unsolved yet...

The "new" warning is just the "skipping duplicate warning", which was
already merged with commit 0a7fb6f07e3a ("objtool: Increase per-function
WARN_FUNC() rate limit").  So none of the warnings are specific to this
pull request.

Tiezhu, can you please look at this warning?
  
   arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
   arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[23]=-1+0 reg2[23]=-2-152

-- 
Josh

Re: [GIT PULL] objtool fixes and updates

[Ingo Molnar]

* Josh Poimboeuf <jpoimboe@kernel.org> wrote:

> On Mon, Mar 31, 2025 at 01:13:55AM +0200, Borislav Petkov wrote:
> > On Sun, Mar 30, 2025 at 03:19:40PM -0700, Linus Torvalds wrote:
> > > On Sat, 29 Mar 2025 at 08:33, Borislav Petkov <bp@alien8.de> wrote:
> > > >
> > > > Btw, test bot complains:
> > > >
> > > > https://lore.kernel.org/r/202503292202.Sge7ZEUc-lkp@intel.com
> > > 
> > > That's not a very helpful error message
> > 
> > I found this:
> > 
> > https://lore.kernel.org/r/202503280703.OARM8SrY-lkp@intel.com
> > 
> > which looks like the original report.
> > 
> > Looks unsolved yet...
> 
> The "new" warning is just the "skipping duplicate warning", which was
> already merged with commit 0a7fb6f07e3a ("objtool: Increase per-function
> WARN_FUNC() rate limit").  So none of the warnings are specific to this
> pull request.
> 
> Tiezhu, can you please look at this warning?
>   
>    arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
>    arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[23]=-1+0 reg2[23]=-2-152

I'm wondering why LKP didn't pick up the primary warning, only the 
final warning that skips duplicate warnings, which is kinda totally 
useless in isolation:

    `-- arch-loongarch-kernel-traps.o:warning:objtool:show_stack:skipping-duplicate-warning(s)

Maybe we should just do what is below - output a single warning, a bit 
like lockdep.

Or delay the reporting to until the entire pass is finished, and 
output something like:

  warning: objtool: show_stack+0xe0: stack state mismatch: reg1[22]=-1+0 reg2[22]=-2-160 [2+ more warnings]

Note: we shouldn't output the real warning count - just that it's 
multiple warnings that might or might not be resolved by fixing the 
primary warning.

Or something like that.

Thanks,

	Ingo

============>
 tools/objtool/include/objtool/warn.h | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/tools/objtool/include/objtool/warn.h b/tools/objtool/include/objtool/warn.h
index b29ac144e4f5..69dd29a3c267 100644
--- a/tools/objtool/include/objtool/warn.h
+++ b/tools/objtool/include/objtool/warn.h
@@ -57,7 +57,7 @@ static inline char *offstr(struct section *sec, unsigned long offset)
 	free(_str);					\
 })
 
-#define WARN_LIMIT 2
+#define WARN_LIMIT 1
 
 #define WARN_INSN(insn, format, ...)					\
 ({									\
@@ -69,8 +69,6 @@ static inline char *offstr(struct section *sec, unsigned long offset)
 		if (_insn->sym)						\
 			_insn->sym->warnings++;				\
 	} else if (_insn->sym && _insn->sym->warnings == WARN_LIMIT) {	\
-		WARN_FUNC("skipping duplicate warning(s)",		\
-			  _insn->sec, _insn->offset);			\
 		_insn->sym->warnings++;					\
 	}								\
 })

Re: [GIT PULL] objtool fixes and updates

[Ingo Molnar]

* Borislav Petkov <bp@alien8.de> wrote:

> On Fri, Mar 28, 2025 at 10:18:57PM +0100, Ingo Molnar wrote:
> > Linus,
> > 
> > Please pull the latest objtool/urgent Git tree from:
> > 
> >    git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git objtool-urgent-2025-03-28
> > 
> >    # HEAD: ae958b12940bcd4ffa32c44684e4f2878bc5e140 objtool, drm/vmwgfx: Don't ignore vmw_send_msg() for ORC
> > 
> > [ Merge note: not all driver fixes below have maintainer acks. ]
> 
> Btw, test bot complains:
> 
> https://lore.kernel.org/r/202503292202.Sge7ZEUc-lkp@intel.com

AFAICS that particular report is not a new regression, but essentially 
a test-bot false positive due to a reporting change, because the 
objtool warning output changed.

This can be seen in the full report:

  https://lore.kernel.org/oe-kbuild-all/202503280703.OARM8SrY-lkp@intel.com/

  All warnings (new ones prefixed by >>):

     arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
     arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[23]=-1+0 reg2[23]=-2-152
  >> arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: skipping duplicate warning(s)

Note how '>>' is the new warning - the summary line added recently that 
suggests that there's more warnings. It appears to me the test-bot 
considers the other warnings old regressions, but I couldn't find any 
trace of them being reported before. Maybe they weren't Cc:-ed to lkml.

Or maybe these *are* all new warnings. I've Cc:-ed the LKP folks.

To add confusion to confusion, the test bot also incorrectly attributed 
ae958b1294 for that new warning line, and incorrectly categorized it as 
a 'recent_error':

  recent_errors
  `-- loongarch-randconfig-001-20250328
      `-- arch-loongarch-kernel-traps.o:warning:objtool:show_stack:skipping-duplicate-warning(s)

While the commit that added this new reporting line is:

  0a7fb6f07e3a ("objtool: Increase per-function WARN_FUNC() rate limit")

... which is a few weeks old and already upstream. We didn't get a 
loongson related test-bot regression report for that commit or any 
other commits, other than the weeks-late & self-contradictory report 
against ae958b1294.

So something's wonky with the test-bot reporting in this particular 
case for loongson, or my parsing thereof.

Thanks,

	Ingo

Re: [GIT PULL] objtool fixes and updates

[Philip Li]

On Mon, Mar 31, 2025 at 01:55:25PM +0200, Ingo Molnar wrote:
> 
> * Borislav Petkov <bp@alien8.de> wrote:
> 
> > On Fri, Mar 28, 2025 at 10:18:57PM +0100, Ingo Molnar wrote:
> > > Linus,
> > > 
> > > Please pull the latest objtool/urgent Git tree from:
> > > 
> > >    git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git objtool-urgent-2025-03-28
> > > 
> > >    # HEAD: ae958b12940bcd4ffa32c44684e4f2878bc5e140 objtool, drm/vmwgfx: Don't ignore vmw_send_msg() for ORC
> > > 
> > > [ Merge note: not all driver fixes below have maintainer acks. ]
> > 
> > Btw, test bot complains:
> > 
> > https://lore.kernel.org/r/202503292202.Sge7ZEUc-lkp@intel.com
> 
> AFAICS that particular report is not a new regression, but essentially 
> a test-bot false positive due to a reporting change, because the 
> objtool warning output changed.
> 
> This can be seen in the full report:
> 
>   https://lore.kernel.org/oe-kbuild-all/202503280703.OARM8SrY-lkp@intel.com/
> 
>   All warnings (new ones prefixed by >>):
> 
>      arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
>      arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[23]=-1+0 reg2[23]=-2-152
>   >> arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: skipping duplicate warning(s)
> 
> Note how '>>' is the new warning - the summary line added recently that 
> suggests that there's more warnings. It appears to me the test-bot 
> considers the other warnings old regressions, but I couldn't find any 
> trace of them being reported before. Maybe they weren't Cc:-ed to lkml.
> 
> Or maybe these *are* all new warnings. I've Cc:-ed the LKP folks.

Hi Ingo and all, sorry for confusion, here only the line with >> is considered
as new warning, due to the commit 0a7fb6f07e3a you mentioned.

From the bot logic, the "stack state mismatch" are not regression (not new warnings),
which already exists in earlier commit.

> 
> To add confusion to confusion, the test bot also incorrectly attributed 
> ae958b1294 for that new warning line, and incorrectly categorized it as 
> a 'recent_error':
> 
>   recent_errors
>   `-- loongarch-randconfig-001-20250328
>       `-- arch-loongarch-kernel-traps.o:warning:objtool:show_stack:skipping-duplicate-warning(s)
> 
> While the commit that added this new reporting line is:
> 
>   0a7fb6f07e3a ("objtool: Increase per-function WARN_FUNC() rate limit")
> 
> ... which is a few weeks old and already upstream. We didn't get a 
> loongson related test-bot regression report for that commit or any 
> other commits, other than the weeks-late & self-contradictory report 
> against ae958b1294.

For 0a7fb6f07e3a, the bot only reported 2 times on x86 [1][2]. For this loongarch report,
the bisection is wrong and is a false positive, I will further check. Meanwhile, the bot
will ignore the bisection of this new objtool message as it is not really a new kernel issue.

[1] https://lore.kernel.org/oe-kbuild-all/202503191453.uFfxQy5R-lkp@intel.com/
[2] https://lore.kernel.org/oe-kbuild-all/202503181736.zkZUBv4N-lkp@intel.com/

> 
> So something's wonky with the test-bot reporting in this particular 
> case for loongson, or my parsing thereof.
> 
> Thanks,
> 
> 	Ingo
> 

Re: [GIT PULL] objtool fixes and updates

[Borislav Petkov]

On Mon, Mar 31, 2025 at 08:31:08PM +0800, Philip Li wrote:
> For 0a7fb6f07e3a, the bot only reported 2 times on x86 [1][2]. For this
> loongarch report, the bisection is wrong and is a false positive, I will
> further check. Meanwhile, the bot will ignore the bisection of this new
> objtool message as it is not really a new kernel issue.

Can you guys get a human being to double-check and vet those reports?! Please!

Because we kinda trust them but

1. they're not really helpful and hard to understand what you're reporting

2. that summary thing is especially useless:

"Error/Warning ids grouped by kconfigs:

recent_errors
`-- loongarch-randconfig-001-20250328
    `-- arch-loongarch-kernel-traps.o:warning:objtool:show_stack:skipping-duplicate-warning(s)"

When I see this, I need to go look for the original reports and somehow
scratch them together. And you have all that info, why don't you simply dump
a URL with the bug materials so that one can inspect them?

3. Last but not least, if this doesn't change I will start ignoring them.
   Because they're not really helping - on the contrary - they're actively
   interfering.

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: [GIT PULL] objtool fixes and updates

[Philip Li]

On Mon, Mar 31, 2025 at 02:36:49PM +0200, Borislav Petkov wrote:
> On Mon, Mar 31, 2025 at 08:31:08PM +0800, Philip Li wrote:
> > For 0a7fb6f07e3a, the bot only reported 2 times on x86 [1][2]. For this
> > loongarch report, the bisection is wrong and is a false positive, I will
> > further check. Meanwhile, the bot will ignore the bisection of this new
> > objtool message as it is not really a new kernel issue.
> 
> Can you guys get a human being to double-check and vet those reports?! Please!

Really sorry, we will add more strict check of objtool warnings. We have
logic to detect the low confidence one and send to https://lore.kernel.org/oe-kbuild/
for manual check before sending out. But there's still case that we fail to
detect.

> 
> Because we kinda trust them but
> 
> 1. they're not really helpful and hard to understand what you're reporting
> 
> 2. that summary thing is especially useless:
> 
> "Error/Warning ids grouped by kconfigs:
> 
> recent_errors
> `-- loongarch-randconfig-001-20250328
>     `-- arch-loongarch-kernel-traps.o:warning:objtool:show_stack:skipping-duplicate-warning(s)"
> 
> When I see this, I need to go look for the original reports and somehow
> scratch them together. And you have all that info, why don't you simply dump
> a URL with the bug materials so that one can inspect them?

I will also check this as it is designed to have link together with reported
error in the summary. It is another bug that the bot should be fixed.

> 
> 3. Last but not least, if this doesn't change I will start ignoring them.
>    Because they're not really helping - on the contrary - they're actively
>    interfering.

Apologize again, we will continue improving the bot to make it really useful.

> 
> Thx.
> 
> -- 
> Regards/Gruss,
>     Boris.
> 
> https://people.kernel.org/tglx/notes-about-netiquette

Re: [GIT PULL] objtool fixes and updates

[Borislav Petkov]

On Mon, Mar 31, 2025 at 08:44:25PM +0800, Philip Li wrote:
> I will also check this as it is designed to have link together with reported
> error in the summary. It is another bug that the bot should be fixed.

Please.

> Apologize again, we will continue improving the bot to make it really useful.

Thanks, I appreciate the effort.

If you need guinea pigs to look at reports and give you feedback, holler!

:-)

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: [GIT PULL] objtool fixes and updates

[Philip Li]

On Mon, Mar 31, 2025 at 02:49:17PM +0200, Borislav Petkov wrote:
> On Mon, Mar 31, 2025 at 08:44:25PM +0800, Philip Li wrote:
> > I will also check this as it is designed to have link together with reported
> > error in the summary. It is another bug that the bot should be fixed.
> 
> Please.
> 
> > Apologize again, we will continue improving the bot to make it really useful.
> 
> Thanks, I appreciate the effort.
> 
> If you need guinea pigs to look at reports and give you feedback, holler!

Got it, thanks a lot!

> 
> :-)
> 
> Thx.
> 
> -- 
> Regards/Gruss,
>     Boris.
> 
> https://people.kernel.org/tglx/notes-about-netiquette

Re: [GIT PULL] objtool fixes and updates

[Josh Poimboeuf]

On Mon, Mar 31, 2025 at 11:39:13AM +0200, Ingo Molnar wrote:
> I'm wondering why LKP didn't pick up the primary warning, only the 
> final warning that skips duplicate warnings, which is kinda totally 
> useless in isolation:

At least it alerted us to the other loongarch warnings we hadn't seen ;-)

>     `-- arch-loongarch-kernel-traps.o:warning:objtool:show_stack:skipping-duplicate-warning(s)
>
> Maybe we should just do what is below - output a single warning, a bit 
> like lockdep.

Yeah, "skipping duplicate warnings" isn't technically a warning so it
probably shouldn't be presented as one.

I'll just revert 0a7fb6f07e3a and go back to the simple "one warning per
function" which worked fine before.

-- 
Josh

Re: [GIT PULL] objtool fixes and updates

[Ingo Molnar]

* Josh Poimboeuf <jpoimboe@kernel.org> wrote:

> On Mon, Mar 31, 2025 at 11:39:13AM +0200, Ingo Molnar wrote:
> > I'm wondering why LKP didn't pick up the primary warning, only the 
> > final warning that skips duplicate warnings, which is kinda totally 
> > useless in isolation:
> 
> At least it alerted us to the other loongarch warnings we hadn't seen ;-)

There's that :-)

> >     `-- arch-loongarch-kernel-traps.o:warning:objtool:show_stack:skipping-duplicate-warning(s)
> >
> > Maybe we should just do what is below - output a single warning, a bit 
> > like lockdep.
> 
> Yeah, "skipping duplicate warnings" isn't technically a warning so it 
> probably shouldn't be presented as one.
> 
> I'll just revert 0a7fb6f07e3a and go back to the simple "one warning 
> per function" which worked fine before.

Thanks! Our experience with lockdep is that the fewer warnings per 
build/bootup, the better: the value of secondary warnings is 
diminishing, sometimes even negative.

It might even be OK to generate just one objtool warning per .o file.

	Ingo

Re: [GIT PULL] objtool fixes and updates

[Ingo Molnar]

* Philip Li <philip.li@intel.com> wrote:

> On Mon, Mar 31, 2025 at 01:55:25PM +0200, Ingo Molnar wrote:
> > 
> > * Borislav Petkov <bp@alien8.de> wrote:
> > 
> > > On Fri, Mar 28, 2025 at 10:18:57PM +0100, Ingo Molnar wrote:
> > > > Linus,
> > > > 
> > > > Please pull the latest objtool/urgent Git tree from:
> > > > 
> > > >    git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git objtool-urgent-2025-03-28
> > > > 
> > > >    # HEAD: ae958b12940bcd4ffa32c44684e4f2878bc5e140 objtool, drm/vmwgfx: Don't ignore vmw_send_msg() for ORC
> > > > 
> > > > [ Merge note: not all driver fixes below have maintainer acks. ]
> > > 
> > > Btw, test bot complains:
> > > 
> > > https://lore.kernel.org/r/202503292202.Sge7ZEUc-lkp@intel.com
> > 
> > AFAICS that particular report is not a new regression, but essentially 
> > a test-bot false positive due to a reporting change, because the 
> > objtool warning output changed.
> > 
> > This can be seen in the full report:
> > 
> >   https://lore.kernel.org/oe-kbuild-all/202503280703.OARM8SrY-lkp@intel.com/
> > 
> >   All warnings (new ones prefixed by >>):
> > 
> >      arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
> >      arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[23]=-1+0 reg2[23]=-2-152
> >   >> arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: skipping duplicate warning(s)
> > 
> > Note how '>>' is the new warning - the summary line added recently that 
> > suggests that there's more warnings. It appears to me the test-bot 
> > considers the other warnings old regressions, but I couldn't find any 
> > trace of them being reported before. Maybe they weren't Cc:-ed to lkml.
> > 
> > Or maybe these *are* all new warnings. I've Cc:-ed the LKP folks.
> 
> Hi Ingo and all, sorry for confusion, here only the line with >> is considered
> as new warning, due to the commit 0a7fb6f07e3a you mentioned.

Oh, so the loongson 'stack state mismatch' warnings started sometime in 
the past, but were never reported by the bot? Or were they reported 
somewhere? I'd like to discover the timeline of those warnings, if 
that's possible.

Thanks,

	Ingo

Re: [GIT PULL] objtool fixes and updates

[Josh Poimboeuf]

On Mon, Mar 31, 2025 at 06:10:45PM +0200, Ingo Molnar wrote:
> > > This can be seen in the full report:
> > > 
> > >   https://lore.kernel.org/oe-kbuild-all/202503280703.OARM8SrY-lkp@intel.com/
> > > 
> > >   All warnings (new ones prefixed by >>):
> > > 
> > >      arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
> > >      arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[23]=-1+0 reg2[23]=-2-152
> > >   >> arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: skipping duplicate warning(s)
> > > 
> > > Note how '>>' is the new warning - the summary line added recently that 
> > > suggests that there's more warnings. It appears to me the test-bot 
> > > considers the other warnings old regressions, but I couldn't find any 
> > > trace of them being reported before. Maybe they weren't Cc:-ed to lkml.
> > > 
> > > Or maybe these *are* all new warnings. I've Cc:-ed the LKP folks.
> > 
> > Hi Ingo and all, sorry for confusion, here only the line with >> is considered
> > as new warning, due to the commit 0a7fb6f07e3a you mentioned.
> 
> Oh, so the loongson 'stack state mismatch' warnings started sometime in 
> the past, but were never reported by the bot? Or were they reported 
> somewhere? I'd like to discover the timeline of those warnings, if 
> that's possible.

I believe those warnings were introduced a year ago with:

  cb8a2ef0848c ("LoongArch: Add ORC stack unwinder support")

but I wasn't able to find the original report.

-- 
Josh

Re: [GIT PULL] objtool fixes and updates

[Josh Poimboeuf]

On Sun, Mar 30, 2025 at 06:39:51PM -0700, Josh Poimboeuf wrote:
> On Mon, Mar 31, 2025 at 01:13:55AM +0200, Borislav Petkov wrote:
> > On Sun, Mar 30, 2025 at 03:19:40PM -0700, Linus Torvalds wrote:
> > > On Sat, 29 Mar 2025 at 08:33, Borislav Petkov <bp@alien8.de> wrote:
> > > >
> > > > Btw, test bot complains:
> > > >
> > > > https://lore.kernel.org/r/202503292202.Sge7ZEUc-lkp@intel.com
> > > 
> > > That's not a very helpful error message
> > 
> > I found this:
> > 
> > https://lore.kernel.org/r/202503280703.OARM8SrY-lkp@intel.com
> > 
> > which looks like the original report.
> > 
> > Looks unsolved yet...
> 
> The "new" warning is just the "skipping duplicate warning", which was
> already merged with commit 0a7fb6f07e3a ("objtool: Increase per-function
> WARN_FUNC() rate limit").  So none of the warnings are specific to this
> pull request.
> 
> Tiezhu, can you please look at this warning?
>   
>    arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
>    arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[23]=-1+0 reg2[23]=-2-152

Here's a fix.  Will post a real fix soon, along with another pile of
fixes.

diff --git a/arch/loongarch/include/asm/stacktrace.h b/arch/loongarch/include/asm/stacktrace.h
index f23adb15f418..fc8b64773794 100644
--- a/arch/loongarch/include/asm/stacktrace.h
+++ b/arch/loongarch/include/asm/stacktrace.h
@@ -8,6 +8,7 @@
 #include <asm/asm.h>
 #include <asm/ptrace.h>
 #include <asm/loongarch.h>
+#include <asm/unwind_hints.h>
 #include <linux/stringify.h>
 
 enum stack_type {
@@ -43,6 +44,7 @@ int get_stack_info(unsigned long stack, struct task_struct *task, struct stack_i
 static __always_inline void prepare_frametrace(struct pt_regs *regs)
 {
 	__asm__ __volatile__(
+		UNWIND_HINT_SAVE
 		/* Save $ra */
 		STORE_ONE_REG(1)
 		/* Use $ra to save PC */
@@ -80,6 +82,7 @@ static __always_inline void prepare_frametrace(struct pt_regs *regs)
 		STORE_ONE_REG(29)
 		STORE_ONE_REG(30)
 		STORE_ONE_REG(31)
+		UNWIND_HINT_RESTORE
 		: "=m" (regs->csr_era)
 		: "r" (regs->regs)
 		: "memory");
diff --git a/arch/loongarch/include/asm/unwind_hints.h b/arch/loongarch/include/asm/unwind_hints.h
index a01086ad9dde..2c68bc72736c 100644
--- a/arch/loongarch/include/asm/unwind_hints.h
+++ b/arch/loongarch/include/asm/unwind_hints.h
@@ -23,6 +23,14 @@
 	UNWIND_HINT sp_reg=ORC_REG_SP type=UNWIND_HINT_TYPE_CALL
 .endm
 
-#endif /* __ASSEMBLY__ */
+#else /* !__ASSEMBLY__ */
+
+#define UNWIND_HINT_SAVE \
+	UNWIND_HINT(UNWIND_HINT_TYPE_SAVE, 0, 0, 0)
+
+#define UNWIND_HINT_RESTORE \
+	UNWIND_HINT(UNWIND_HINT_TYPE_RESTORE, 0, 0, 0)
+
+#endif /* !__ASSEMBLY__ */
 
 #endif /* _ASM_LOONGARCH_UNWIND_HINTS_H */

Re: [GIT PULL] objtool fixes and updates

[Philip Li]

On Mon, Mar 31, 2025 at 03:17:51PM -0700, Josh Poimboeuf wrote:
> On Mon, Mar 31, 2025 at 06:10:45PM +0200, Ingo Molnar wrote:
> > > > This can be seen in the full report:
> > > > 
> > > >   https://lore.kernel.org/oe-kbuild-all/202503280703.OARM8SrY-lkp@intel.com/
> > > > 
> > > >   All warnings (new ones prefixed by >>):
> > > > 
> > > >      arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
> > > >      arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[23]=-1+0 reg2[23]=-2-152
> > > >   >> arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: skipping duplicate warning(s)
> > > > 
> > > > Note how '>>' is the new warning - the summary line added recently that 
> > > > suggests that there's more warnings. It appears to me the test-bot 
> > > > considers the other warnings old regressions, but I couldn't find any 
> > > > trace of them being reported before. Maybe they weren't Cc:-ed to lkml.
> > > > 
> > > > Or maybe these *are* all new warnings. I've Cc:-ed the LKP folks.
> > > 
> > > Hi Ingo and all, sorry for confusion, here only the line with >> is considered
> > > as new warning, due to the commit 0a7fb6f07e3a you mentioned.
> > 
> > Oh, so the loongson 'stack state mismatch' warnings started sometime in 
> > the past, but were never reported by the bot? Or were they reported 
> > somewhere? I'd like to discover the timeline of those warnings, if 
> > that's possible.
> 
> I believe those warnings were introduced a year ago with:
> 
>   cb8a2ef0848c ("LoongArch: Add ORC stack unwinder support")
> 
> but I wasn't able to find the original report.

Hi Ingo and Josh, for this cb8a2ef0848c commit, the bot has reported at [1]
but with a different randconfig and a different issue pattern. And the bot
didn't report new ones as it considered the bad commit was already reported.

	tree:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
	commit: cb8a2ef0848ca80d67d6d56e2df757cfdf6b3355 LoongArch: Add ORC stack unwinder support
	config: loongarch-randconfig-002-20240812 
(https://download.01.org/0day-ci/archive/20240812/202408120955.qls5oNQY-lkp@intel.com/config)

	WARNING: modpost: missing MODULE_DESCRIPTION() in net/ieee802154/6lowpan/ieee802154_6lowpan.o
	>> ERROR: modpost: "kvm_restore_lsx" [arch/loongarch/kvm/kvm.ko] undefined!
	>> ERROR: modpost: "kvm_restore_lasx" [arch/loongarch/kvm/kvm.ko] undefined!

Later, there's fix at [2], but seems it doesn't solve all possible issues since
the bot didn't report out explicitly.

[1] https://lore.kernel.org/oe-kbuild-all/202408120955.qls5oNQY-lkp@intel.com/
[2] https://lore.kernel.org/all/20240604150741.30252-1-xry111@xry111.site/

Thanks

> 
> -- 
> Josh

Re: [GIT PULL] objtool fixes and updates

[Tiezhu Yang]

Hi all,

Sorry for the late reply.

(1) The following objtool warnings:

modpost: "kvm_restore_lsx" [arch/loongarch/kvm/kvm.ko] undefined!
modpost: "kvm_restore_lasx" [arch/loongarch/kvm/kvm.ko] undefined!

have been fixed and merged (2024-08-26):

https://web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=80376323e2b6

There are "Reported-by" and "Closes" tags in the commit message.

Reported-by: kernel test robot <lkp@intel.com>
Closes: 
https://lore.kernel.org/oe-kbuild-all/202408120955.qls5oNQY-lkp@intel.com/

(2) I did not meet the following objtool warnings before this merge
window:

arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack 
state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack 
state mismatch: reg1[23]=-1+0 reg2[23]=-2-152

and I did not notice the robot report due to there are many unread mails
in my inbox, I will try to reproduce it and take a deeper look at the
problem, maybe it is related with some special configs.

Thanks,
Tiezhu

Re: [GIT PULL] objtool fixes and updates

[Josh Poimboeuf]

On Tue, Apr 01, 2025 at 10:05:41AM +0800, Tiezhu Yang wrote:
> (2) I did not meet the following objtool warnings before this merge
> window:
> 
> arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack
> state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
> arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack
> state mismatch: reg1[23]=-1+0 reg2[23]=-2-152
> 
> and I did not notice the robot report due to there are many unread mails
> in my inbox, I will try to reproduce it and take a deeper look at the
> problem, maybe it is related with some special configs.

I actually already have a patch for it, will post soon.

-- 
Josh

Re: [GIT PULL] objtool fixes and updates

[Tiezhu Yang]

On 04/01/2025 12:15 PM, Josh Poimboeuf wrote:
> On Tue, Apr 01, 2025 at 10:05:41AM +0800, Tiezhu Yang wrote:
>> (2) I did not meet the following objtool warnings before this merge
>> window:
>>
>> arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack
>> state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
>> arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack
>> state mismatch: reg1[23]=-1+0 reg2[23]=-2-152
>>
>> and I did not notice the robot report due to there are many unread mails
>> in my inbox, I will try to reproduce it and take a deeper look at the
>> problem, maybe it is related with some special configs.
>
> I actually already have a patch for it, will post soon.

Thanks for the fix, I will test it once you post it.

I can confirm that the objtool warning is not related with
the patches in the branch objtool/urgent of tip.git, it is
related with CONFIG_STACKPROTECTOR which is default y, but
CONFIG_STACKPROTECTOR is not set in the robot randconfig
loongarch-randconfig-001-20250328 [1], there is no objtool
warning if CONFIG_STACKPROTECTOR is set.

I can also reproduce this issue when compiling with v6.14
if CONFIG_STACKPROTECTOR is not set.

[1] 
https://download.01.org/0day-ci/archive/20250328/202503280703.OARM8SrY-lkp@intel.com/config

Thanks,
Tiezhu

Re: [GIT PULL] objtool fixes and updates

[Ingo Molnar]

* Josh Poimboeuf <jpoimboe@kernel.org> wrote:

> On Sun, Mar 30, 2025 at 06:39:51PM -0700, Josh Poimboeuf wrote:
> > On Mon, Mar 31, 2025 at 01:13:55AM +0200, Borislav Petkov wrote:
> > > On Sun, Mar 30, 2025 at 03:19:40PM -0700, Linus Torvalds wrote:
> > > > On Sat, 29 Mar 2025 at 08:33, Borislav Petkov <bp@alien8.de> wrote:
> > > > >
> > > > > Btw, test bot complains:
> > > > >
> > > > > https://lore.kernel.org/r/202503292202.Sge7ZEUc-lkp@intel.com
> > > > 
> > > > That's not a very helpful error message
> > > 
> > > I found this:
> > > 
> > > https://lore.kernel.org/r/202503280703.OARM8SrY-lkp@intel.com
> > > 
> > > which looks like the original report.
> > > 
> > > Looks unsolved yet...
> > 
> > The "new" warning is just the "skipping duplicate warning", which was
> > already merged with commit 0a7fb6f07e3a ("objtool: Increase per-function
> > WARN_FUNC() rate limit").  So none of the warnings are specific to this
> > pull request.
> > 
> > Tiezhu, can you please look at this warning?
> >   
> >    arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[22]=-1+0 reg2[22]=-2-160
> >    arch/loongarch/kernel/traps.o: warning: objtool: show_stack+0xe0: stack state mismatch: reg1[23]=-1+0 reg2[23]=-2-152
> 
> Here's a fix.  Will post a real fix soon, along with another pile of 
> fixes.

And just to make it clear, these objtool warnings were not a new 
regression, they were introduced more than a year ago, via:

  cb8a2ef0848c ("LoongArch: Add ORC stack unwinder support")

So, to bring this thread to a conclusion, I think by getting rid of the 
summary warning line:

  c5610071a69d ("Revert "objtool: Increase per-function WARN_FUNC() rate limit"")

... the CI test-bots ought to be back to the v6.14 baseline even taking 
such false positives into account.

I'll send the updated objtool/urgent tree to Linus later today, unless 
some last-minute problem pops up.

Thanks,

	Ingo

Re: [GIT PULL] objtool fixes and updates

[pr-tracker-bot]

The pull request you sent on Fri, 28 Mar 2025 22:18:57 +0100:

> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git objtool-urgent-2025-03-28

has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/55c78035a1a8dfb05f1472018ce2a651701adb7d

Thank you!

-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/prtracker.html