[PATCH] x86/tools: objdump_reformat.awk: Skip bad instructions from llvm-objdump

[PATCH] x86/tools: objdump_reformat.awk: Skip bad instructions from llvm-objdump

[Nathan Chancellor]

When running the instruction decoder selftest with LLVM=1 +
CONFIG_PVH=y, there is a series of warnings:

  arch/x86/tools/insn_decoder_test: warning: Found an x86 instruction decoder bug, please report this.
  arch/x86/tools/insn_decoder_test: warning: ffffffff81000050     ea                      <unknown>
  arch/x86/tools/insn_decoder_test: warning: objdump says 1 bytes, but insn_get_length() says 7
  arch/x86/tools/insn_decoder_test: warning: Decoded and checked 7214721 instructions with 1 failures

GNU objdump outputs "(bad)" instead of "<unknown>", which is already
handled in the bad_expr regex, so there is no warning.

  $ objdump -d arch/x86/platform/pvh/head.o | grep -E '50:\s+ea'
  50:   ea                      (bad)

  $ llvm-objdump -d arch/x86/platform/pvh/head.o | grep -E '50:\s+ea'
        50: ea                            <unknown>

Add "<unknown>" to the bad_expr regex to clear up the warning, allowing
the instruction decoder selftest to fully pass with llvm-objdump.

Signed-off-by: Nathan Chancellor <nathan@kernel.org>
---
 arch/x86/tools/objdump_reformat.awk | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/tools/objdump_reformat.awk b/arch/x86/tools/objdump_reformat.awk
index a4120d907277..20b08a6c4d33 100644
--- a/arch/x86/tools/objdump_reformat.awk
+++ b/arch/x86/tools/objdump_reformat.awk
@@ -11,7 +11,7 @@ BEGIN {
 	prev_addr = ""
 	prev_hex = ""
 	prev_mnemonic = ""
-	bad_expr = "(\\(bad\\)|^rex|^.byte|^rep(z|nz)$|^lock$|^es$|^cs$|^ss$|^ds$|^fs$|^gs$|^data(16|32)$|^addr(16|32|64))"
+	bad_expr = "(\\(bad\\)|<unknown>|^rex|^.byte|^rep(z|nz)$|^lock$|^es$|^cs$|^ss$|^ds$|^fs$|^gs$|^data(16|32)$|^addr(16|32|64))"
 	fwait_expr = "^9b[ \t]*fwait"
 	fwait_str="9b\tfwait"
 }

---
base-commit: 5225952d74d43e4c054731c74b8afd700b23a94a
change-id: 20231205-objdump_reformat-awk-handle-llvm-objdump-bad_expr-9e74cd2a08b5

Best regards,
-- 
Nathan Chancellor <nathan@kernel.org>

Re: [PATCH] x86/tools: objdump_reformat.awk: Skip bad instructions from llvm-objdump

[Nathan Chancellor]

Ping? I am still seeing this issue.

On Tue, Dec 05, 2023 at 12:53:08PM -0700, Nathan Chancellor wrote:
> When running the instruction decoder selftest with LLVM=1 +
> CONFIG_PVH=y, there is a series of warnings:
> 
>   arch/x86/tools/insn_decoder_test: warning: Found an x86 instruction decoder bug, please report this.
>   arch/x86/tools/insn_decoder_test: warning: ffffffff81000050     ea                      <unknown>
>   arch/x86/tools/insn_decoder_test: warning: objdump says 1 bytes, but insn_get_length() says 7
>   arch/x86/tools/insn_decoder_test: warning: Decoded and checked 7214721 instructions with 1 failures
> 
> GNU objdump outputs "(bad)" instead of "<unknown>", which is already
> handled in the bad_expr regex, so there is no warning.
> 
>   $ objdump -d arch/x86/platform/pvh/head.o | grep -E '50:\s+ea'
>   50:   ea                      (bad)
> 
>   $ llvm-objdump -d arch/x86/platform/pvh/head.o | grep -E '50:\s+ea'
>         50: ea                            <unknown>
> 
> Add "<unknown>" to the bad_expr regex to clear up the warning, allowing
> the instruction decoder selftest to fully pass with llvm-objdump.
> 
> Signed-off-by: Nathan Chancellor <nathan@kernel.org>
> ---
>  arch/x86/tools/objdump_reformat.awk | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/x86/tools/objdump_reformat.awk b/arch/x86/tools/objdump_reformat.awk
> index a4120d907277..20b08a6c4d33 100644
> --- a/arch/x86/tools/objdump_reformat.awk
> +++ b/arch/x86/tools/objdump_reformat.awk
> @@ -11,7 +11,7 @@ BEGIN {
>  	prev_addr = ""
>  	prev_hex = ""
>  	prev_mnemonic = ""
> -	bad_expr = "(\\(bad\\)|^rex|^.byte|^rep(z|nz)$|^lock$|^es$|^cs$|^ss$|^ds$|^fs$|^gs$|^data(16|32)$|^addr(16|32|64))"
> +	bad_expr = "(\\(bad\\)|<unknown>|^rex|^.byte|^rep(z|nz)$|^lock$|^es$|^cs$|^ss$|^ds$|^fs$|^gs$|^data(16|32)$|^addr(16|32|64))"
>  	fwait_expr = "^9b[ \t]*fwait"
>  	fwait_str="9b\tfwait"
>  }
> 
> ---
> base-commit: 5225952d74d43e4c054731c74b8afd700b23a94a
> change-id: 20231205-objdump_reformat-awk-handle-llvm-objdump-bad_expr-9e74cd2a08b5
> 
> Best regards,
> -- 
> Nathan Chancellor <nathan@kernel.org>
> 
> 

Re: [PATCH] x86/tools: objdump_reformat.awk: Skip bad instructions from llvm-objdump

[Borislav Petkov]

On Wed, Jan 03, 2024 at 11:15:42AM -0700, Nathan Chancellor wrote:
> Ping? I am still seeing this issue.

Does this need a Fixes tag and needs to go to Linus now or are you fine
with 6.8-rc0?

I.e., the answer probably depends on what kernels you're running the
llvm tests...

-- 
Regards/Gruss,
    Boris.

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

Re: [PATCH] x86/tools: objdump_reformat.awk: Skip bad instructions from llvm-objdump

[Nathan Chancellor]

On Wed, Jan 03, 2024 at 07:18:52PM +0100, Borislav Petkov wrote:
> On Wed, Jan 03, 2024 at 11:15:42AM -0700, Nathan Chancellor wrote:
> > Ping? I am still seeing this issue.
> 
> Does this need a Fixes tag and needs to go to Linus now or are you fine
> with 6.8-rc0?

This is only needed due to the recent changes from Sam and myself in
x86/build, so no need to rush it to Linus. I just wanted to make sure it
was not lost before the chaos of the merge window.

Cheers,
Nathan

Re: [PATCH] x86/tools: objdump_reformat.awk: Skip bad instructions from llvm-objdump

[Borislav Petkov]

On Wed, Jan 03, 2024 at 01:55:06PM -0700, Nathan Chancellor wrote:
> On Wed, Jan 03, 2024 at 07:18:52PM +0100, Borislav Petkov wrote:
> > On Wed, Jan 03, 2024 at 11:15:42AM -0700, Nathan Chancellor wrote:
> > > Ping? I am still seeing this issue.
> > 
> > Does this need a Fixes tag and needs to go to Linus now or are you fine
> > with 6.8-rc0?
> 
> This is only needed due to the recent changes from Sam and myself in
> x86/build

I don't understand: those three commits seem unrelated to LLVM objdump
outputting "<unknown>".

Or are you saying that you've *started* running the insn decoder selftest
with llvm's objdump and those three commits are addressing differences
in the output and this outstanding commit is needed too for the bad
opcode case?

Thx.

-- 
Regards/Gruss,
    Boris.

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

Re: [PATCH] x86/tools: objdump_reformat.awk: Skip bad instructions from llvm-objdump

[Nathan Chancellor]

On Wed, Jan 03, 2024 at 10:26:16PM +0100, Borislav Petkov wrote:
> On Wed, Jan 03, 2024 at 01:55:06PM -0700, Nathan Chancellor wrote:
> > On Wed, Jan 03, 2024 at 07:18:52PM +0100, Borislav Petkov wrote:
> > > On Wed, Jan 03, 2024 at 11:15:42AM -0700, Nathan Chancellor wrote:
> > > > Ping? I am still seeing this issue.
> > > 
> > > Does this need a Fixes tag and needs to go to Linus now or are you fine
> > > with 6.8-rc0?
> > 
> > This is only needed due to the recent changes from Sam and myself in
> > x86/build
> 
> I don't understand: those three commits seem unrelated to LLVM objdump
> outputting "<unknown>".
> 
> Or are you saying that you've *started* running the insn decoder selftest
> with llvm's objdump and those three commits are addressing differences
> in the output and this outstanding commit is needed too for the bad
> opcode case?

Prior to commit 5225952d74d4 ("x86/tools: Remove chkobjdump.awk"), the
insn decoder selftest would not actually run with llvm-objdump
altogether because chkobjdump.awk would fail (because it was only
written for GNU objdump). The two commits prior to 5225952d74d4 were
cleaning up differences between the output of each objdump
implementations and this change should have been a part of that work as
well, I just did not build enough configurations to see it. Hopefully
that clears things up.

Cheers,
Nathan

Re: [PATCH] x86/tools: objdump_reformat.awk: Skip bad instructions from llvm-objdump

[Nathan Chancellor]

On Wed, Jan 03, 2024 at 02:48:09PM -0700, Nathan Chancellor wrote:
> On Wed, Jan 03, 2024 at 10:26:16PM +0100, Borislav Petkov wrote:
> > On Wed, Jan 03, 2024 at 01:55:06PM -0700, Nathan Chancellor wrote:
> > > On Wed, Jan 03, 2024 at 07:18:52PM +0100, Borislav Petkov wrote:
> > > > On Wed, Jan 03, 2024 at 11:15:42AM -0700, Nathan Chancellor wrote:
> > > > > Ping? I am still seeing this issue.
> > > > 
> > > > Does this need a Fixes tag and needs to go to Linus now or are you fine
> > > > with 6.8-rc0?
> > > 
> > > This is only needed due to the recent changes from Sam and myself in
> > > x86/build
> > 
> > I don't understand: those three commits seem unrelated to LLVM objdump
> > outputting "<unknown>".
> > 
> > Or are you saying that you've *started* running the insn decoder selftest
> > with llvm's objdump and those three commits are addressing differences
> > in the output and this outstanding commit is needed too for the bad
> > opcode case?
> 
> Prior to commit 5225952d74d4 ("x86/tools: Remove chkobjdump.awk"), the
> insn decoder selftest would not actually run with llvm-objdump
> altogether because chkobjdump.awk would fail (because it was only
> written for GNU objdump). The two commits prior to 5225952d74d4 were
> cleaning up differences between the output of each objdump
> implementations and this change should have been a part of that work as
> well, I just did not build enough configurations to see it. Hopefully
> that clears things up.

For the record, this explanation really should have been in the commit
message of 5225952d74d4 but I guess I was not thinking at the time.

Cheers,
Nathan

Re: [PATCH] x86/tools: objdump_reformat.awk: Skip bad instructions from llvm-objdump

[Borislav Petkov]

On Wed, Jan 03, 2024 at 02:51:19PM -0700, Nathan Chancellor wrote:
> For the record, this explanation really should have been in the commit
> message of 5225952d74d4 but I guess I was not thinking at the time.

That's fine - we have the Link tag. Lemme refer to that thread and we're
good.

Thx.

-- 
Regards/Gruss,
    Boris.

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