scripts: update namespace.pl
cover scripts: update namespace.pl
This patch series updates scripts/namespace.pl:
1/4: handle rename of lib-ksyms.o to lib.a by scripts/Makefile.build 2/4: add a few x86 object files to the namespace exceptions list 3/4: include trailing ".o" in the namespace exceptions list of
file names and change/fix "acpia" to "acpica"
4/4: add a blank line to separate a section of the output
scripts/namespace.pl | 48 +++++++++++++++++++++++------------------ 1 file changed, 27 insertions(+), 21 deletions(-)
PATCH 1/4 scripts: namespace.pl: handle lib-ksyms.o as lib.a
Fix these warnings:
No source file found for drivers/firmware/efi/libstub/lib-ksyms.o No source file found for arch/x86/lib/lib-ksyms.o
No source file found for lib/lib-ksyms.o
In scripts/Makefile.build, any resulting lib-ksyms.o file is renamed to lib.a, so handle that rename here also.
---
patch 1/4:
scripts/namespace.pl | 3 +++ 1 file changed, 3 insertions(+)
scripts/namespace.pl +3 -0
@@ -231,6 +231,9 @@ sub do_nm return; } ($source = $basename) =~ s/\.o$//; + if ($basename eq "lib-ksyms.o") { # modified in scripts/Makefile.build + $basename = "lib.a"; # from lib.a to lib-ksyms.o + } if (-e "$source.c" || -e "$source.S") { $source = File::Spec->catfile($objtree, $File::Find::dir, $source) } else {
PATCH 2/4 scripts: namespace.pl: add some x86 object files to nmexception
In reply to [email protected]These x86 object files do not have any exports or public symbols so they can be excluded from the namespace checking.
---
patch 2/4:
scripts/namespace.pl | 2 ++ 1 file changed, 2 insertions(+)
scripts/namespace.pl +2 -0
@@ -91,6 +91,8 @@ my %nmexception = ( 'arch/x86/lib/thunk_32' => 1, 'arch/x86/lib/cmpxchg' => 1, 'arch/x86/vdso/vdso32/note' => 1, + 'arch/x86/entry/vdso/vdso-note.o' => 1, + 'arch/x86/entry/vdso/vdso32/note.o' => 1, 'lib/irq_regs' => 1, 'usr/initramfs_data' => 1, 'drivers/scsi/aic94xx/aic94xx_dump' => 1,
PATCH 3/4 scripts: namespace.pl: add .o to nmexception file names and fix "acpica"
In reply to [email protected]'nm' exceptions are checked against $fullname, which includes the trailing '.o', so the list of file names should include the trailing '.o'.
Also, the sub-directory for drivers/acpi/acpica/ has a typo, so s/acpia/acpica/g.
The acpica list of exceptions almost surely needs to be updated.
---
patch 3/4:
scripts/namespace.pl | 42 ++++++++++++++++++++--------------------- 1 file changed, 21 insertions(+), 21 deletions(-)
scripts/namespace.pl +21 -21
@@ -86,29 +86,29 @@ my %ref = (); # $ref{$name} exists if t my %export = (); # $export{$name} exists if there is an EXPORT_... of $name my %nmexception = ( - 'fs/ext3/bitmap' => 1, - 'fs/ext4/bitmap' => 1, - 'arch/x86/lib/thunk_32' => 1, - 'arch/x86/lib/cmpxchg' => 1, - 'arch/x86/vdso/vdso32/note' => 1, + 'fs/ext3/bitmap.o' => 1, + 'fs/ext4/bitmap.o' => 1, + 'arch/x86/lib/thunk_32.o' => 1, + 'arch/x86/lib/cmpxchg.o' => 1, + 'arch/x86/vdso/vdso32/note.o' => 1, 'arch/x86/entry/vdso/vdso-note.o' => 1, 'arch/x86/entry/vdso/vdso32/note.o' => 1, - 'lib/irq_regs' => 1, - 'usr/initramfs_data' => 1, - 'drivers/scsi/aic94xx/aic94xx_dump' => 1, - 'drivers/scsi/libsas/sas_dump' => 1, - 'lib/dec_and_lock' => 1, - 'drivers/ide/ide-probe-mini' => 1, - 'usr/initramfs_data' => 1, - 'drivers/acpi/acpia/exdump' => 1, - 'drivers/acpi/acpia/rsdump' => 1, - 'drivers/acpi/acpia/nsdumpdv' => 1, - 'drivers/acpi/acpia/nsdump' => 1, - 'arch/ia64/sn/kernel/sn2/io' => 1, - 'arch/ia64/kernel/gate-data' => 1, - 'security/capability' => 1, - 'fs/ntfs/sysctl' => 1, - 'fs/jfs/jfs_debug' => 1, + 'lib/irq_regs.o' => 1, + 'usr/initramfs_data.o' => 1, + 'drivers/scsi/aic94xx/aic94xx_dump.o' => 1, + 'drivers/scsi/libsas/sas_dump.o' => 1, + 'lib/dec_and_lock.o' => 1, + 'drivers/ide/ide-probe-mini.o' => 1, + 'usr/initramfs_data.o' => 1, + 'drivers/acpi/acpica/exdump.o' => 1, + 'drivers/acpi/acpica/rsdump.o' => 1, + 'drivers/acpi/acpica/nsdumpdv.o' => 1, + 'drivers/acpi/acpica/nsdump.o' => 1, + 'arch/ia64/sn/kernel/sn2/io.o' => 1, + 'arch/ia64/kernel/gate-data.o' => 1, + 'security/capability.o' => 1, + 'fs/ntfs/sysctl.o' => 1, + 'fs/jfs/jfs_debug.o' => 1, ); my %nameexception = (
PATCH 4/4 scripts: namespace.pl: add blank section separator line
In reply to [email protected]Most namespace report sections are separated by blank lines, but a blank line is needed before the "multiply defined" output as a separator, to make it easier to find the beginning of this section of output.
---
patch 4/4:
scripts/namespace.pl | 1 + 1 file changed, 1 insertion(+)
scripts/namespace.pl +1 -0
@@ -368,6 +368,7 @@ sub drop_def sub list_multiply_defined { + printf "\n"; foreach my $name (keys(%def)) { if ($#{$def{$name}} > 0) { # Special case for cond_syscall
Re: [PATCH 1/4] scripts: namespace.pl: handle lib-ksyms.o as lib.a
In reply to [email protected]Randy,
Quoted 2 lines · On Thu, Oct 3, 2019 at 12:18 PM <[email protected]> wrote:
In
scripts/Makefile.build, any resulting lib-ksyms.o file is renamed to lib.a, so handle that rename here also.
This description is incorrect.
lib-ksyms.o is included in built-in.a
The code attempts to suppress the warnings
in a strange way.
I think the following is a more straightforward way.
scripts/namespace.pl +1 -0
@@ -166,6 +166,7 @@ sub linux_objects if (/.*\.o$/ && ! ( m:/built-in.a$: + || m:/lib-ksyms.o$: || m:arch/x86/vdso/: || m:arch/x86/boot/: || m:arch/ia64/ia32/ia32.o$:
--
Best Regards
Masahiro Yamada
Re: [PATCH 2/4] scripts: namespace.pl: add some x86 object files to nmexception
In reply to [email protected]On Thu, Oct 3, 2019 at 12:18 PM <[email protected]> wrote:
You can grow %nmexception until
you no longer see "No nm data for ..." messages,
but is this maintainable?
I think it is more realistic to
delete the printf "No nm data for $fullname\n" line.
Signature · 2 lines
Best Regards Masahiro Yamada
Re: [PATCH 3/4] scripts: namespace.pl: add .o to nmexception file names and fix "acpica"
In reply to [email protected]On Thu, Oct 3, 2019 at 12:18 PM <[email protected]> wrote:
The same comment as 2/4.
I doubt if %nmexception is maintainable.
For example, you are fixing fs/ext3/bitmap to fs/ext3/bitmap.o but I do not see the directory fs/ext3/.
How many entries in the %nmexception are still valid?
--
Best Regards
Masahiro Yamada