Re: [PATCH v6 03/13] scripts/syscall.tbl: Use the common table

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] The perf tool's copy of `scripts/syscall.tbl` is left out of sync, and its syscall generator is unaware of the new common table.
- [Low] The use of `mktemp -p .` in build scripts is a GNU extension causing build failures on BSD-based hosts, and temporary files are leaked on error paths.
--

commit 81fbe53cf08247610dc3aa87d8f9c9e8106a9994
Author: André Almeida <andrealmeid@igalia.com>

scripts/syscall.tbl: Use the common table

Remove some of the duplicated code by using the common syscall number
table.

> diff --git a/scripts/Makefile.asm-headers b/scripts/Makefile.asm-headers
> index b38931314ad7c..e81d3ef8938ca 100644
> --- a/scripts/Makefile.asm-headers
> +++ b/scripts/Makefile.asm-headers
[ ... ]
> @@ -57,14 +58,14 @@ quiet_cmd_remove = REMOVE  $(unwanted)
>        cmd_remove = rm -f $(unwanted); find $(obj) -type d -empty -delete
>  
>  quiet_cmd_syshdr = SYSHDR  $@
> -      cmd_syshdr = $(CONFIG_SHELL) $(syshdr) \
> +      cmd_syshdr = $(CONFIG_SHELL) $(syshdr) --common-tbl $(syscalltbl_common) \

[Severity: Low]
Does this build script rely on a GNU-specific extension?

While not modified directly in this diff, scripts/syscallhdr.sh uses
mktemp -p ., which is a GNU extension:

scripts/syscallhdr.sh:gen_hdr() {
    ...
    tmpfile=$(mktemp -p .)
    ...
}

This will cause build failures on non-GNU hosts (such as macOS or FreeBSD)
where their mktemp implementations do not support the -p flag.

Additionally, if the script exits early via exit 1 in scripts/syscalltbl.sh
within gen_tbl(), the temporary file is permanently leaked in the build tree
because cleanup relies on a direct rm at the end instead of a shell trap:

scripts/syscalltbl.sh:gen_tbl() {
    ...
    tmpfile=$(mktemp -p .)
    ...
    exit 1
    ...
}

>  		   $(if $(syshdr-args-$*),$(syshdr-args-$*),$(syshdr-args)) \
>  		   $(if $(syscall_compat),--prefix "compat$*_") \
>  		   --abis $(subst $(space),$(comma),$(strip $(syscall_abis_$*))) \
>  		   $< $@
[ ... ]
> diff --git a/scripts/syscall.tbl b/scripts/syscall.tbl
> index 0ab5316051202..ee828d03094ed 100644
> --- a/scripts/syscall.tbl
> +++ b/scripts/syscall.tbl
> @@ -344,72 +344,3 @@
>  293	common	rseq				sys_rseq
>  294	common	kexec_file_load			sys_kexec_file_load
>  # 295 through 402 are unassigned to sync up with generic numbers don't use
> -403	32	clock_gettime64			sys_clock_gettime

[Severity: Medium]
Will this removal break perf's syscall resolution for architectures relying
on the generic syscall table?

By removing syscalls 403-472 from scripts/syscall.tbl, perf trace may fail
to resolve these syscalls by name (mapping them to unknown) because perf's
generator script is unaware of the new common table.

Looking at tools/perf/trace/beauty/syscalltbl.sh, the build_tables function
relies solely on scripts/syscall.tbl and has not been updated to include
scripts/syscall_common.tbl:

tools/perf/trace/beauty/syscalltbl.sh:build_tables() {
    ...
    build_tables "$tools_dir/scripts/syscall.tbl" "$outfile" common,32,riscv,memfd_secret EM_RISCV
    ...
}

This will also cause the perf build to emit warnings due to
tools/scripts/syscall.tbl being out of sync with scripts/syscall.tbl.

> -404	32	clock_settime64			sys_clock_settime
[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260828-tonyk-syscall_table-v6-0-fdd358456cf1@igalia.com?part=3