From: Greg Ungerer <gerg@uclinux.org>
To: Philippe De Muyter <phdm@macq.eu>
Cc: linux-m68k@vger.kernel.org
Subject: Re: [PATCH v2] m68k: use conventional function parameters for do_sigreturn
Date: Mon, 15 Feb 2016 23:06:15 +1000 [thread overview]
Message-ID: <56C1CD47.8000109@uclinux.org> (raw)
In-Reply-To: <20160215100301.GA30557@frolo.macqel>
Hi Philippe,
On 15/02/16 20:03, Philippe De Muyter wrote:
> On Mon, Feb 15, 2016 at 04:36:29PM +1000, Greg Ungerer wrote:
>> Create conventional stack parameters for the calls to do_sigreturn and
>> do_rt_sigreturn. The current C code for do_sigreturn and do_rt_sigreturn
>> dig into the stack to create local pointers to the saved switch stack
>> and the pt_regs structs.
>>
>> The motivation for this change is a problem with non-MMU targets that
>> have broken signal return paths on newer versions of gcc. It appears as
>> though gcc has determined that the pointers into the saved stack structs,
>> and the saved structs themselves, are function parameters and updates to
>> them will be lost on function return, so they are optimized away. This
>> results in large parts of restore_sigcontext() and mangle_kernel_stack()
>> functions being removed. Of course this results in non-functional code
>> causing kernel oops. This problem has been observed with gcc version
>> 5.2 and 5.3, and probably exists in earlier versions as well.
>>
>> Using conventional stack parameter pointers passed to these functions has
>> the advantage of the code here not needing to know the exact details of
>> how the underlying entry handler layed these structs out on the stack.
>> So the rather ugly pointer setup casting and arg referencing can be
>> removed.
>>
>> The resulting code after this change is a few bytes larger (due to the
>> overhead of creating the stack args and their tear down). Not being hot
>> paths I don't think this is too much of a problem here.
>
> As an intermediate solution, I think that you can avoid part of that overhead ...
>
>>
>> An alternative solution is to put a barrier() in the do_sigreturn() code,
>> but this doesn't feel quite as clean as this solution.
>>
>> This change has been compile tested on all defconfigs, and run tested on
>> Atari (through aranym), ColdFire with MMU (M5407EVB) and ColdFire with
>> no-MMU (QEMU and M5208EVB).
>>
>> Signed-off-by: Greg Ungerer <gerg@uclinux.org>
>> ---
>> v2: reworded the commit log message with better description of problem
>>
>> arch/m68k/kernel/entry.S | 6 ++++++
>> arch/m68k/kernel/signal.c | 8 ++------
>> 2 files changed, 8 insertions(+), 6 deletions(-)
>>
>> diff --git a/arch/m68k/kernel/entry.S b/arch/m68k/kernel/entry.S
>> index b54ac7a..97cd3ea 100644
>> --- a/arch/m68k/kernel/entry.S
>> +++ b/arch/m68k/kernel/entry.S
>> @@ -71,13 +71,19 @@ ENTRY(__sys_vfork)
>>
>> ENTRY(sys_sigreturn)
>> SAVE_SWITCH_STACK
>> + movel %sp,%sp@- | switch_stack pointer
>> + pea %sp@(SWITCH_STACK_SIZE+4) | pt_regs pointer
>
> if you remove the above asm instruction ...
>
>> jbsr do_sigreturn
>> + addql #8,%sp
>
> change the 8 by 4 here ...
>
>> RESTORE_SWITCH_STACK
>> rts
>>
>> ENTRY(sys_rt_sigreturn)
>> SAVE_SWITCH_STACK
>> + movel %sp,%sp@- | switch_stack pointer
>> + pea %sp@(SWITCH_STACK_SIZE+4) | pt_regs pointer
>
> ditto here ...
>
>> jbsr do_rt_sigreturn
>> + addql #8,%sp
>
> and here ...
>
>> RESTORE_SWITCH_STACK
>> rts
>>
>> diff --git a/arch/m68k/kernel/signal.c b/arch/m68k/kernel/signal.c
>> index af1c4f3..2dcee3a 100644
>> --- a/arch/m68k/kernel/signal.c
>> +++ b/arch/m68k/kernel/signal.c
>> @@ -737,10 +737,8 @@ badframe:
>> return 1;
>> }
>>
>> -asmlinkage int do_sigreturn(unsigned long __unused)
>> +asmlinkage int do_sigreturn(struct pt_regs *regs, struct switch_stack *sw)
>
> and if you use the following prototype :
> asmlinkage int do_sigreturn(struct switch_stack *sw) ...
>
>> {
>> - struct switch_stack *sw = (struct switch_stack *) &__unused;
>> - struct pt_regs *regs = (struct pt_regs *) (sw + 1);
>
> and keep the 'C' instruction above
>
>> unsigned long usp = rdusp();
>> struct sigframe __user *frame = (struct sigframe __user *)(usp - 4);
>> sigset_t set;
>> @@ -764,10 +762,8 @@ badframe:
>> return 0;
>> }
>>
>> -asmlinkage int do_rt_sigreturn(unsigned long __unused)
>> +asmlinkage int do_rt_sigreturn(struct pt_regs *regs, struct switch_stack *sw)
> same here
>> {
>> - struct switch_stack *sw = (struct switch_stack *) &__unused;
>> - struct pt_regs *regs = (struct pt_regs *) (sw + 1);
> and here
>> unsigned long usp = rdusp();
>> struct rt_sigframe __user *frame = (struct rt_sigframe __user *)(usp - 4);
>> sigset_t set;
Those changes ultimately make no difference to code size.
Although the entry point code is smaller the resulting
generated code for do_sigreturn() and do_rt_sigreturn()
is slightly larger. (At least that is the case with the gcc-5.3
compiler I am currently using).
Personally I don't like having to keep that ugly stack referencing
definition of regs based on sw either. I would like to remove that.
"sw" isn't actually used in do_sigreturn() at all, so that could
be removed from the its parameters. I chose to keep it there in this
patch for consistency with do_rt_sigreturn().
Regards
Greg
next prev parent reply other threads:[~2016-02-15 13:06 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-02-15 6:36 [PATCH v2] m68k: use conventional function parameters for do_sigreturn Greg Ungerer
2016-02-15 10:03 ` Philippe De Muyter
2016-02-15 13:06 ` Greg Ungerer [this message]
2016-02-15 14:57 ` Philippe De Muyter
2016-02-24 1:10 ` Greg Ungerer
2016-02-25 10:42 ` Geert Uytterhoeven
2016-02-25 10:55 ` Andreas Schwab
2016-02-29 9:36 ` Geert Uytterhoeven
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=56C1CD47.8000109@uclinux.org \
--to=gerg@uclinux.org \
--cc=linux-m68k@vger.kernel.org \
--cc=phdm@macq.eu \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox