Skip to contentSkip to conversation

*** Fix reformat_objdump.awk ***

PATCH Superseded Samuel Zeter 3 participants 11 messages Jan 6, 2022 – Mar 4, 2022 x86

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

cover *** Fix reformat_objdump.awk ***

These are two small patches which originally dealt with the problem found at:

https://github.com/ClangBuiltLinux/linux/issues/1364

The original steps to reproduce were:

$ make -skj"$(nproc)" LLVM=1 LLVM_IAS=1 defconfig
$ scripts/config -e X86_DECODER_SELFTEST
$ make -skj"$(nproc)" LLVM=1 LLVM_IAS=1 olddefconfig bzImage

Which resulted in the error:
arch/x86/tools/insn_decoder_test: warning: objdump says 0 bytes, but insn_get_length() says 2

Upon inspection it turned out llvm-objdump was formatting its output differently, which caused objdump_reformat.awk to incorrectly output its values.

After fixing that bug, a second one was seen where the instruction "wait" was incorrectly matched with "fwait", which again caused insn_decoder_test to fail.

Samuel Zeter (2):
arch/x86/tools/objdump_reformat.awk: Ensure regex matches fwait arch/x86/tools/objdump_reformat.awk: Allow for spaces

 arch/x86/tools/objdump_reformat.awk | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
Signature · 1 line
2.32.0
#1 lore raw

PATCH 1/2 arch/x86/tools/objdump_reformat.awk: Ensure regex matches fwait

If there is "wait" mnemonic in the line being parsed, it is incorrectly handled by the script, and an extra line of "fwait" in objdump_reformat's output is inserted. As insn_decoder_test relies upon the formatted output, the test fails.

This is reproducible when disassembling with llvm-objdump:

Pre-processed lines:
ffffffff81033e72: 9b wait
ffffffff81033e73: 48 c7 c7 89 50 42 82 movq

After objdump_reformat.awk:

ffffffff81033e72:       9b      fwait
ffffffff81033e72:                               wait
ffffffff81033e73:       48 c7 c7 89 50 42 82    movq

This patch fixes the issue by requiring spaces, or tabs, along with the "fwait" instruction in the regex match.

---
 arch/x86/tools/objdump_reformat.awk | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/tools/objdump_reformat.awk +1 -1
index f418c91b71f0..276e572a6f60 100644
@@ -12,7 +12,7 @@ BEGIN {
 	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))"
-	fwait_expr = "^9b "
+	fwait_expr = "^9b[ \t]*fwait"
 	fwait_str="9b\tfwait"
 }
 
Signature · 1 line
2.32.0
#2 lore raw

PATCH 2/2 arch/x86/tools/objdump_reformat.awk: Allow for spaces

In reply to Samuel Zeter

objdump and llvm-objdump have differing output formats. Specifically, objump will format its output as: address:<tab>hex, whereas llvm-objdump displays its output as address:<space>hex.

objdump_reformat.awk incorrectly handles this discrepancy due to the unexpected space and as a result insn_decoder_test fails, as
its input is garbled.

This patch ensures that the instruction line being tokenized can support either a space and colon, or tab delimiter.

---
 arch/x86/tools/objdump_reformat.awk | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/tools/objdump_reformat.awk +1 -1
index 276e572a6f60..a4120d907277 100644
@@ -22,7 +22,7 @@ BEGIN {
 }
 
 /^ *[0-9a-f]+:/ {
-	if (split($0, field, "\t") < 3) {
+	if (split($0, field, /: |\t/) < 3) {
 		# This is a continuation of the same insn.
 		prev_hex = prev_hex field[2]
 	} else {
Signature · 1 line
2.32.0
Tested-by on the v1 series
#3 lore raw
In reply to Samuel Zeter
Quoted 14 lines · Samuel Zeter

These are two small patches which originally dealt with the problem found at:

https://github.com/ClangBuiltLinux/linux/issues/1364

The original steps to reproduce were:

$ make -skj"$(nproc)" LLVM=1 LLVM_IAS=1 defconfig
$ scripts/config -e X86_DECODER_SELFTEST
$ make -skj"$(nproc)" LLVM=1 LLVM_IAS=1 olddefconfig bzImage

Which resulted in the error:
arch/x86/tools/insn_decoder_test: warning: objdump says 0 bytes, but insn_get_length() says 2

Quoted 7 lines

Upon inspection it turned out llvm-objdump was formatting its output differently, which caused objdump_reformat.awk to incorrectly output its values.

After fixing that bug, a second one was seen where the instruction "wait" was incorrectly matched with "fwait", which again caused insn_decoder_test to fail.

Thanks a lot for sending these fixes!

I can confirm with this series and the removal of chkobjdump.awk [1] on top of v5.16-rc8, the insn_decoder_test now passes with LLVM 11 through 14.

For the future, I recommend putting the maintainers in the "To" field, rather than "Cc", to ensure they actually see it. Additionally, I see some small nits in the commit message that the tip maintainers might comment on, see

https://www.kernel.org/doc/html/latest/process/maintainer-tip.html#changelog

for some more info.

Masami Hiramatsu originally wrote this file and has a few fixes to it since, adding him now for review. The original thread is available at:

https://lore.kernel.org/r/[email protected]/

[1]: https://git.kernel.org/nathan/c/2f137c324b21f1c21b8830d8896cb9957009f969

Cheers,
Nathan

#4 lore raw

Thanks for the feedback, Nathan.

Quoted 7 lines

For the future, I recommend putting the maintainers in the "To" field, rather than "Cc", to ensure they actually see it. Additionally, I see some small nits in the commit message that the tip maintainers might comment on, see

https://www.kernel.org/doc/html/latest/process/maintainer-tip.html#changelog

Thanks for the link, I missed that one. What were the nits apparent to you in the commit message?

Sam.

#5 lore raw
Quoted 11 lines · Samuel Zeter

Thanks for the feedback, Nathan.

Quoted 7 lines

For the future, I recommend putting the maintainers in the "To" field, rather than "Cc", to ensure they actually see it. Additionally, I see some small nits in the commit message that the tip maintainers might comment on, see

https://www.kernel.org/doc/html/latest/process/maintainer-tip.html#changelog

Thanks for the link, I missed that one. What were the nits apparent to you in the commit message?

I primarily just saw a couple instances of "This patch", which is frowned upon in the main submitting patches document:

https://www.kernel.org/doc/html/latest/process/submitting-patches.html#submittingpatches

I thought I saw something else but upon further inspection, I didn't. It is minor enough that I would wait for further review comments from others to submit a v2.

Cheers,
Nathan

author
#6 lore raw

v2 cover *** Fix reformat_objdump.awk ***

v1 -> v2
- coding style commit message amendments.

Samuel Zeter (2):
arch/x86/tools/objdump_reformat.awk: Ensure regex matches fwait arch/x86/tools/objdump_reformat.awk: Allow for spaces

 arch/x86/tools/objdump_reformat.awk | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
Signature · 1 line
2.35.1
author
#7 lore raw

PATCH v2 1/2 arch/x86/tools/objdump_reformat.awk: Ensure regex matches fwait

From: Samuel Zeter <[email protected]>

If there is "wait" mnemonic in the line being parsed, it is incorrectly handled by the script, and an extra line of "fwait" in objdump_reformat's output is inserted. As insn_decoder_test relies upon the formatted output, the test fails.

This is reproducible when disassembling with llvm-objdump:

Pre-processed lines:
ffffffff81033e72: 9b wait
ffffffff81033e73: 48 c7 c7 89 50 42 82 movq

After objdump_reformat.awk:

ffffffff81033e72:       9b      fwait
ffffffff81033e72:                               wait
ffffffff81033e73:       48 c7 c7 89 50 42 82    movq

The regex match now accepts spaces or tabs, along with the "fwait" instruction.

---
 arch/x86/tools/objdump_reformat.awk | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/tools/objdump_reformat.awk +1 -1
index f418c91b7..276e572a6 100644
@@ -12,7 +12,7 @@ BEGIN {
 	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))"
-	fwait_expr = "^9b "
+	fwait_expr = "^9b[ \t]*fwait"
 	fwait_str="9b\tfwait"
 }
 
Signature · 1 line
2.35.1
author
#8 lore raw

PATCH v2 2/2 arch/x86/tools/objdump_reformat.awk: Allow for spaces

In reply to Sam Zeter

From: Samuel Zeter <[email protected]>

objdump and llvm-objdump have differing output formats. Specifically, objump will format its output as: address:<tab>hex, whereas llvm-objdump displays its output as address:<space>hex.

objdump_reformat.awk incorrectly handles this discrepancy due to the unexpected space and as a result insn_decoder_test fails, as its input is garbled.

The instruction line being tokenized now handles a space and colon, or tab delimiter.

---
 arch/x86/tools/objdump_reformat.awk | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/tools/objdump_reformat.awk +1 -1
index 276e572a6..a4120d907 100644
@@ -22,7 +22,7 @@ BEGIN {
 }
 
 /^ *[0-9a-f]+:/ {
-	if (split($0, field, "\t") < 3) {
+	if (split($0, field, /: |\t/) < 3) {
 		# This is a continuation of the same insn.
 		prev_hex = prev_hex field[2]
 	} else {
Signature · 1 line
2.35.1
Acked-by Masami Hiramatsu Tested-by on the v2 series
#9 lore raw
In reply to Sam Zeter

Hi Sam,

Quoted 9 lines · Sam Zeter

v1 -> v2
- coding style commit message amendments.

Samuel Zeter (2):
arch/x86/tools/objdump_reformat.awk: Ensure regex matches fwait arch/x86/tools/objdump_reformat.awk: Allow for spaces

 arch/x86/tools/objdump_reformat.awk | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Thank you for sending v2!

Since you only updated the commit messages (i.e., no functional changes), you should carry forward my tested-by and Masami's acked-by on both patches:

Additionally, for any future series/revisions, I would recommend leaving the rest of the cover letter's message intact, so that people who missed v1 are not in the dark on the purpose of the series (link provided for convenience):

https://lore.kernel.org/r/[email protected]/

Regardless, those are just small process suggestions. I think this should be ready to go, if the x86 maintainers have no objections. Great work!

Cheers,
Nathan

#10 lore raw
In reply to Nathan Chancellor

Hi Samuel and Nathan,

Thanks for working on this issue. I didn't noticed that the llvm has this difference. Anyway both patches look good to me.

for this series.

And if you resend it, please add "[email protected]" to Cc.

Thank you!

Quoted 49 lines · On Thu, 6 Jan 2022 12:02:25 -0700 Nathan Chancellor <[email protected]> wrote:
Quoted 14 lines · Samuel Zeter

These are two small patches which originally dealt with the problem found at:

https://github.com/ClangBuiltLinux/linux/issues/1364

The original steps to reproduce were:

$ make -skj"$(nproc)" LLVM=1 LLVM_IAS=1 defconfig
$ scripts/config -e X86_DECODER_SELFTEST
$ make -skj"$(nproc)" LLVM=1 LLVM_IAS=1 olddefconfig bzImage

Which resulted in the error:
arch/x86/tools/insn_decoder_test: warning: objdump says 0 bytes, but insn_get_length() says 2

Quoted 7 lines

Upon inspection it turned out llvm-objdump was formatting its output differently, which caused objdump_reformat.awk to incorrectly output its values.

After fixing that bug, a second one was seen where the instruction "wait" was incorrectly matched with "fwait", which again caused insn_decoder_test to fail.

Thanks a lot for sending these fixes!

I can confirm with this series and the removal of chkobjdump.awk [1] on top of v5.16-rc8, the insn_decoder_test now passes with LLVM 11 through 14.

For the future, I recommend putting the maintainers in the "To" field, rather than "Cc", to ensure they actually see it. Additionally, I see some small nits in the commit message that the tip maintainers might comment on, see

https://www.kernel.org/doc/html/latest/process/maintainer-tip.html#changelog

for some more info.

Masami Hiramatsu originally wrote this file and has a few fixes to it since, adding him now for review. The original thread is available at:

https://lore.kernel.org/r/[email protected]/

[1]: https://git.kernel.org/nathan/c/2f137c324b21f1c21b8830d8896cb9957009f969

Cheers,
Nathan

Signature · 1 line
Masami Hiramatsu <[email protected]>