Hello !
Few months ago some developers of linux-mips on R3K
(not all, but only _happy members_, having mips boxes with hard disks :-)
were discussing a swap functionality problem.
Discussion (see msg. below) was focused on random user program crashes
occurred in the system working with swapping turned on.
Unfortunately, we plug in a (SCSI) hard disk into our BAGET/R3081E just now,
and today we are looking to above problem. It looks that problem was fixed
in 2.3.xx, but our most 'production' version is still 2.2.1.
Looking a differences we found a lot of interesting fixes in
linux/include/asm-mips
since 2.2.1, but we have no time to investigate/migrate the code.
(customers are waiting a coolest dual-board baget computer, and only
this problem hold us before deadline :-)
Q: Maybe somebody remember a hint which was a key point for fixing this bug ?
Harald Koerfgen wrote:
>
> On 06-Jul-99 Ralf Baechle wrote:
> > I've received a report from some person who is working on his own R3081
> > port. He also observes data corruption and suspects reading of swapped
> > pages is causing that.
>
> That's definitely true for R3k DECstations, and no, flushing the icache in
> flush_tlb_page() does not help. I have added cacheflushing to all tlb
> routines,
> copy_page and even rw_swap_page_base() and swap_after_unlock_page() without
> success.
>
> Any ideas?
> ---
> Regards,
> Harald
>
Thanks in advance,
Vladimir.
|