RE: [linux-vr-cvsupdate] Update to module linux

[Mike K. <mf...@po...>]

Actually, Ralf pointed me at that one.  Apparently it's broken for all R4K
in the Linux-MIPS tree, and he hasn't been able to fix it due to hardware
problems.

The icache line sizing thing I just happened to stumble across due to the
flush_icache_page bug, since I noticed it was deadlocking in a page fault
inside r4k_flush_icache_page_i32, which is the 32 byte line size version of
that function.

Mike Klar

> -----Original Message-----
> From: lin...@li...
> [mailto:lin...@li...]On Behalf Of
> Hiroshi Kawashima
> Sent: Sunday, September 17, 2000 7:37 PM
> To: Michael Klar
> Cc: lin...@li...
> Subject: Re: [linux-vr-cvsupdate] Update to module linux
>
>
> Hi, Mike.
>
> Thank you following hack.
> Just my curious, how you can find out this bug ?
> Do you have any test code to reproduce this problem ?
> (I suspect if this effect (fix) X11 on hard-float probelm...hopefully...)
>
> > Modified Files:
> > 	r4xx0.c
> > Log Message:
> > Hack around a deadlock in flush_icache_page, and fix an icache
> sizing bug
> ----
> Kawashima
> _______________________________________________
> linux-vr-cvsupdate mailing list
> lin...@li...
> http://lists.sourceforge.net/mailman/listinfo/linux-vr-cvsupdate
>