[PATCH 0/14] JANITOR: Remove include/linux/sys.h

[PATCH 0/14] JANITOR: Remove include/linux/sys.h

[Mika Kukkonen]

Hello,

File include/linux/sys.h has a comment stating that the "file is no longer used or needed",
but it is still included in multiple places under several architectures. It does contain
bunch of #define's but those have been #ifdef'd out, and grepping did not show any uses
of those #define's in the source tree outside this file.

The uses of the file seem partially be accidental remains of some old use, but almost all
are in architecture specific code with only one outside in file net/core/pktgen.c. So this
patchset has been divided to 14 parts, addressing 12 architectures (c6x, cris, frv, h8300,
microblaze, mips, mn10300, parisc, powerpc, sh, unicore32 and x86) and the outlier in net
and then the final patch removes the file itself. All patches can be applied invidually
with the obvious exception of the last one which actually removes the file and should be
last.

Diffstat:
 arch/c6x/kernel/entry.S                    |  1 -
 arch/cris/arch-v10/kernel/entry.S          |  1 -
 arch/cris/arch-v32/kernel/entry.S          |  1 -
 arch/frv/kernel/cmode.S                    |  1 -
 arch/frv/kernel/entry-table.S              |  1 -
 arch/frv/kernel/sleep.S                    |  1 -
 arch/frv/mm/tlb-flush.S                    |  1 -
 arch/frv/mm/tlb-miss.S                     |  1 -
 arch/h8300/kernel/entry.S                  |  1 -
 arch/h8300/kernel/syscalls.S               |  1 -
 arch/microblaze/kernel/entry.S             |  1 -
 arch/microblaze/kernel/misc.S              |  1 -
 arch/microblaze/kernel/sys_microblaze.c    |  1 -
 arch/mips/kernel/process.c                 |  1 -
 arch/mn10300/kernel/entry.S                |  1 -
 arch/mn10300/kernel/gdb-io-serial-low.S    |  1 -
 arch/mn10300/kernel/gdb-io-ttysm-low.S     |  1 -
 arch/mn10300/kernel/gdb-low.S              |  1 -
 arch/mn10300/kernel/mn10300-serial-low.S   |  1 -
 arch/mn10300/kernel/mn10300-watchdog-low.S |  1 -
 arch/mn10300/kernel/profile-low.S          |  1 -
 arch/mn10300/kernel/smp-low.S              |  1 -
 arch/mn10300/kernel/switch_to.S            |  1 -
 arch/mn10300/mm/cache-dbg-flush-by-reg.S   |  1 -
 arch/mn10300/mm/cache-dbg-flush-by-tag.S   |  1 -
 arch/mn10300/mm/cache-dbg-inv-by-reg.S     |  1 -
 arch/mn10300/mm/cache-dbg-inv-by-tag.S     |  1 -
 arch/mn10300/mm/cache-dbg-inv.S            |  1 -
 arch/mn10300/mm/cache-flush-by-reg.S       |  1 -
 arch/mn10300/mm/cache-flush-by-tag.S       |  1 -
 arch/mn10300/mm/cache-inv-by-reg.S         |  1 -
 arch/mn10300/mm/cache-inv-by-tag.S         |  1 -
 arch/mn10300/mm/tlb-mn10300.S              |  1 -
 arch/parisc/hpux/entry_hpux.S              |  1 -
 arch/powerpc/kernel/entry_32.S             |  1 -
 arch/powerpc/kernel/misc_32.S              |  1 -
 arch/powerpc/kernel/misc_64.S              |  1 -
 arch/powerpc/kernel/syscalls.c             |  1 -
 arch/powerpc/lib/checksum_32.S             |  1 -
 arch/powerpc/lib/checksum_64.S             |  1 -
 arch/sh/boards/mach-ap325rxa/sdram.S       |  1 -
 arch/sh/boards/mach-ecovec24/sdram.S       |  1 -
 arch/sh/boards/mach-kfr2r09/sdram.S        |  1 -
 arch/sh/boards/mach-migor/sdram.S          |  1 -
 arch/sh/boards/mach-se/7724/sdram.S        |  1 -
 arch/sh/kernel/cpu/sh3/entry.S             |  1 -
 arch/sh/kernel/cpu/sh3/swsusp.S            |  1 -
 arch/sh/kernel/cpu/sh5/entry.S             |  1 -
 arch/sh/kernel/cpu/shmobile/sleep.S        |  1 -
 arch/sh/kernel/debugtraps.S                |  1 -
 arch/sh/kernel/syscalls_32.S               |  1 -
 arch/sh/kernel/syscalls_64.S               |  2 --
 arch/unicore32/kernel/hibernate_asm.S      |  1 -
 arch/x86/ia32/syscall_ia32.c               |  1 -
 arch/x86/kernel/syscall_32.c               |  1 -
 arch/x86/kernel/syscall_64.c               |  1 -
 arch/x86/um/sys_call_table_32.c            |  1 -
 arch/x86/um/sys_call_table_64.c            |  1 -
 include/linux/sys.h                        | 29 -----------------------------
 net/core/pktgen.c                          |  1 -
 60 files changed, 89 deletions(-)

Compile tested only on x86, but should be obviously correct since the symbols are
not defined or used anywhere.

The patchset is also available in my personal github tree as a branch janitor_retire_sys_h:
https://github.com/mikukkon/linux/tree/janitor_retire_sys_h.

Signed-off-by: Mika Kukkonen <m.kukkonen@samsung.com>

--MiKu

Re: [PATCH 0/14] JANITOR: Remove include/linux/sys.h

[Dan Carpenter]

On Tue, Oct 29, 2013 at 11:53:48AM +0200, Mika Kukkonen wrote:
> Hello,
> 
> File include/linux/sys.h has a comment stating that the "file is no longer used or needed",
> but it is still included in multiple places under several architectures. It does contain
> bunch of #define's but those have been #ifdef'd out, and grepping did not show any uses
> of those #define's in the source tree outside this file.
> 
> The uses of the file seem partially be accidental remains of some old use, but almost all
> are in architecture specific code with only one outside in file net/core/pktgen.c. So this
> patchset has been divided to 14 parts, addressing 12 architectures (c6x, cris, frv, h8300,
> microblaze, mips, mn10300, parisc, powerpc, sh, unicore32 and x86) and the outlier in net
> and then the final patch removes the file itself. All patches can be applied invidually
> with the obvious exception of the last one which actually removes the file and should be
> last.
> 

Kernel Janitors doesn't have an a patch tree.  In theory these could all
go through trivial@kernel.org or they could go to the subsystem
maintainers.  You pretty much have broken it up by subsystem already.

The subjects need to be different.  Each should reflect the subsystem
conventions.  Find the correct subject by doing:
git log --oneline arch/c6x/
git log --oneline arch/cris/

The complication with doing it by the subsystem is that the last patch
as to go in last.  I would suggest sending the first 13 patches right
away and holding off until v3.13-rc1 has been released in 4 weeks time.
By then the earlier patches should have been merged.  Otherwise wait
until v3.14 to apply the last patch.  It's not a race, and it will make
it easiest if the 13 have gone to Linus before we take the final step.

regards,
dan carpenter

Re: [PATCH 0/14] JANITOR: Remove include/linux/sys.h

[Mika Kukkonen]

On Tue, Oct 29, 2013 at 01:36:09PM +0000, Dan Carpenter wrote:
> 
> Kernel Janitors doesn't have an a patch tree.  In theory these could all
> go through trivial@kernel.org or they could go to the subsystem
> maintainers.  You pretty much have broken it up by subsystem already.

Yes, but I chickened out and did not want to CC 14 subsystem maintainers. ;-)

> The complication with doing it by the subsystem is that the last patch
> as to go in last.  I would suggest sending the first 13 patches right
> away and holding off until v3.13-rc1 has been released in 4 weeks time.
> By then the earlier patches should have been merged.  Otherwise wait
> until v3.14 to apply the last patch.  It's not a race, and it will make
> it easiest if the 13 have gone to Linus before we take the final step.

Well, this is in no way urgent issue, so I'll revisit it in the next release
cycle if nothing happends before that. Thanks for your feedback!

--MiKu

Re: [PATCH 0/14] JANITOR: Remove include/linux/sys.h

[Julia Lawall]

On Wed, 30 Oct 2013, Mika Kukkonen wrote:

> On Tue, Oct 29, 2013 at 01:36:09PM +0000, Dan Carpenter wrote:
> >
> > Kernel Janitors doesn't have an a patch tree.  In theory these could all
> > go through trivial@kernel.org or they could go to the subsystem
> > maintainers.  You pretty much have broken it up by subsystem already.
>
> Yes, but I chickened out and did not want to CC 14 subsystem maintainers. ;-)

I've deleted the mail with the patches, but if it would be useful for the
maintainers to see the patches, you should send them to them.  That is,
send each patch to the people indicated by get_maintainers.

julia

>
> > The complication with doing it by the subsystem is that the last patch
> > as to go in last.  I would suggest sending the first 13 patches right
> > away and holding off until v3.13-rc1 has been released in 4 weeks time.
> > By then the earlier patches should have been merged.  Otherwise wait
> > until v3.14 to apply the last patch.  It's not a race, and it will make
> > it easiest if the 13 have gone to Linus before we take the final step.
>
> Well, this is in no way urgent issue, so I'll revisit it in the next release
> cycle if nothing happends before that. Thanks for your feedback!
>
> --MiKu
> --
> To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: [PATCH 0/14] JANITOR: Remove include/linux/sys.h

[Mika Kukkonen]

On Wed, Oct 30, 2013 at 11:03:57AM +0000, Julia Lawall wrote:
> On Wed, 30 Oct 2013, Mika Kukkonen wrote:
> 
> > On Tue, Oct 29, 2013 at 01:36:09PM +0000, Dan Carpenter wrote:
> > >
> > > Kernel Janitors doesn't have an a patch tree.  In theory these could all
> > > go through trivial@kernel.org or they could go to the subsystem
> > > maintainers.  You pretty much have broken it up by subsystem already.
> >
> > Yes, but I chickened out and did not want to CC 14 subsystem maintainers. ;-)
> 
> I've deleted the mail with the patches, but if it would be useful for the
> maintainers to see the patches, you should send them to them.  That is,
> send each patch to the people indicated by get_maintainers.

Yeah, I'll probably do that, but I got bunch other janitorial stuff(*) in pipeline,
so I will probably delay this to next merge window as this is in no way urgent.

--MiKu

(*) Basically I am trying to make "make W=1" usefull, I just happened to stumble
    upon this <linux/sys.h> while working on that.