Re: [PATCH spice-common v2] codegen: Make the compiler work out better way to write unaligned memory

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, Sep 18, 2017 at 10:27:01AM -0400, Frediano Ziglio wrote:
> > 
> > Has this been tested on actual problematic hardware, and improved
> > performance? I remember last time this was discussed, the outcome was
> > not always what was expected.
> > 
> > Christophe
> > 
> 
> I don't remember old discussions, there were actually 2 different ones
> which were confusing as one was referring an old embedded ARM version
> and another a more recent.
> I remember a memcpy version was proposed.
> I don't own an ARM to test.
> Was tested with x64 with no regressions introduced.
> Knowing the machine instruction, which one fault is causing the issue
> and the different platforms I can try and verify the assembly output
> cross compiling.

Given there were past attempts which should have been working, but did
not when testing on real hardware, I'd prefer that we postpone this
patch until someone could run tests on real hardware. This is in an area
of the code which does not change much, so it should not need painful
rebases.

Christophe
_______________________________________________
Spice-devel mailing list
Spice-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/spice-devel




[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Security]     [Bugtraq]     [Linux]     [Linux OMAP]     [Linux MIPS]     [ECOS]     [Asterisk Internet PBX]     [Linux API]     [Monitors]