[PATCH] fix comment in Light-weight-syscall code

[PATCH] fix comment in Light-weight-syscall code

[Helge Deller]

The comment for the lws-atomic system call userspace helper 
function mentions r22 and r23 as input registers, but I couldn't 
find any usage of those registers in the code.
This patch fixes this and corrects another small typo as well.

Signed-off-by: Helge Deller <deller@gmx.de>

diff --git a/arch/parisc/kernel/syscall.S b/arch/parisc/kernel/syscall.S
index 69b6eeb..ae45262 100644
--- a/arch/parisc/kernel/syscall.S
+++ b/arch/parisc/kernel/syscall.S
@@ -368,11 +368,11 @@ tracesys_sigexit:
 		Light-weight-syscall code
 
 		r20 - lws number
-		r26,r25,r24,r23,r22 - Input registers
+		r26,r25,r24 - Input registers
 		r28 - Function return register
 		r21 - Error code.
 
-		Scracth: Any of the above that aren't being
+		Scratch: Any of the above that aren't being
 		currently used, including r1. 
 
 		Return pointer: r31 (Not usable)

Re: [PATCH] fix comment in Light-weight-syscall code

[Kyle McMartin]

On Thu, Jul 10, 2008 at 10:04:15PM +0200, Helge Deller wrote:
> The comment for the lws-atomic system call userspace helper 
> function mentions r22 and r23 as input registers, but I couldn't 
> find any usage of those registers in the code.
> This patch fixes this and corrects another small typo as well.
> 
> Signed-off-by: Helge Deller <deller@gmx.de>
> 

If Carlos thinks this is OK, it's OK by me.

Re: [PATCH] fix comment in Light-weight-syscall code

[Carlos O'Donell]

On Thu, Jul 10, 2008 at 4:04 PM, Helge Deller <deller@gmx.de> wrote:
> The comment for the lws-atomic system call userspace helper
> function mentions r22 and r23 as input registers, but I couldn't
> find any usage of those registers in the code.
> This patch fixes this and corrects another small typo as well.

That's the LWS syscall ABI. It has 5 argument registers, 1 return
register, 1 error code register, and 1 lws number register.

Cheers,
Carlos.

Re: [PATCH] fix comment in Light-weight-syscall code

[Kyle McMartin]

On Fri, Jul 11, 2008 at 01:02:10PM -0400, Carlos O'Donell wrote:
> On Thu, Jul 10, 2008 at 4:04 PM, Helge Deller <deller@gmx.de> wrote:
> > The comment for the lws-atomic system call userspace helper
> > function mentions r22 and r23 as input registers, but I couldn't
> > find any usage of those registers in the code.
> > This patch fixes this and corrects another small typo as well.
> 
> That's the LWS syscall ABI. It has 5 argument registers, 1 return
> register, 1 error code register, and 1 lws number register.
> 

Yeah, but I think what Helge is asking is, if there's no plans to ever
add more, or more arguments, can he change it to be 3 args?

r, Kyle

Re: [PATCH] fix comment in Light-weight-syscall code

[Carlos O'Donell]

On Fri, Jul 11, 2008 at 2:33 PM, Kyle McMartin <kyle@mcmartin.ca> wrote:
> On Fri, Jul 11, 2008 at 01:02:10PM -0400, Carlos O'Donell wrote:
>> On Thu, Jul 10, 2008 at 4:04 PM, Helge Deller <deller@gmx.de> wrote:
>> > The comment for the lws-atomic system call userspace helper
>> > function mentions r22 and r23 as input registers, but I couldn't
>> > find any usage of those registers in the code.
>> > This patch fixes this and corrects another small typo as well.
>>
>> That's the LWS syscall ABI. It has 5 argument registers, 1 return
>> register, 1 error code register, and 1 lws number register.
>>
>
> Yeah, but I think what Helge is asking is, if there's no plans to ever
> add more, or more arguments, can he change it to be 3 args?

No, I'm not OK with it, I think we should keep the extra two argument
registers for future use.

I've been talking to Helge about this. Once we lose those registers
they are gone from the ABI forever, and if we need some fixup, or an
additional arg in the future, it will be hard to handle.

Cheers,
Carlos.