Re: [PATCH -mm] ide_end_drive_cmd(): avoid instruction pipelinestall

[Andrew Morton]

On Fri, 30 Jun 2006 18:26:56 +0100
Alan Cox <alan@xxxxxxxxxxxxxxxxxxx> wrote:

> Ar Gwe, 2006-06-30 am 18:13 +0200, ysgrifennodd Andreas Mohr:
> > Use an independently-formatted "unsigned int" for data instead of a
> > restrictive "u16" to avoid instruction fetch pipeline stalls
> > probably caused by the byte calculations later.
> 
> drivers/ide is on its way out.

Like sound/oss ;)

> I'm also curious that this shows up given
> that the inw() is going to cause a PCI sequence and stall the CPU
> entirely for ages anyway.

I guess because he was profiling for IFU_MEM_STALL, not for wall-time.

> NAK because
> 1. This is a gcc problem
> 2. Not everyone is using an intel x86-32 box which has such problems
> 3. IDE is in life-support mode and the relatives are already planning
> the flowers.

Well.  If the patch breaks anything we can dine on hats for a month.  Seems
pretty inoffensive to me.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/