Messages in this thread | | | Date | Fri, 7 Nov 1997 23:43:12 -0500 (EST) | From | Jason Duerstock <> | Subject | Re: pre-2.0.32... |
| |
On Fri, 7 Nov 1997, Linus Torvalds wrote:
> as I haven't seen much noise about 31 brokenness and fixing that up, I > put together my personal pre-32 patch and put it on ftp.kernel.org (in the > usual place: pub/linux/kernel/testing). I hoped that others (Doug?) would > have picked up the 2.0.x series, but that failing I hope to at least do a > minimal band-aid. > > How does my pre-32 look to people?
It looks broken. :)
kernel: Unable to handle kernel paging request at virtual address 0000000000c15433 kswapd(3): Oops 0 pc = [<fffffc000032cea0>] ps = 0000 rp = [<fffffc00003336b4>] sp = fffffc0000005de0 r0=0 r1=c1540b r2=2ca1 r3=c1540b r8=40000000000 r16=0 r17=0 r18=1 r19=155558d4000 r20=0 r21=0 r22=0 r23=300000 r24=2f800000000 r25=6f800000000 r26=fffffc00003336b4 r27=fffffc000032cda8 r28=1 r29=fffffc0000454a10 r30=fffffc0000005de0 Code: 47ff041f 47ff041f 47ff041f <a0230028> 48209781 40651402 e020000c a8c30028 44c21001
I have to admit not having a lot of experience reading alpha dumps, but if pc == program counter, then it bombed in mmap_shrink in mm/filemap.c.
3d4: 1f 04 ff 47 nop 3d8: 1f 04 ff 47 nop 3dc: 1f 04 ff 47 nop /usr/src/linux-2.0.32/include/asm/bitops.h:85 3e0: 28 00 23 a0 ldl t0,40(t2) ^^^^- here is where it bombed 3e4: 81 97 20 48 sra t0,0x4,t0 /usr/src/linux-2.0.32/include/asm/bitops.h:84 3e8: 02 14 65 40 addq t2,0x28,t1 /usr/src/linux-2.0.32/mm/filemap.c:144 3ec: 0c 00 20 e0 blbc t0,420 <shrink_mmap..ng+130> /usr/src/linux-2.0.32/include/asm/bitops.h:45 3f0: 28 00 c3 a8 ldl_l t5,40(t2) 3f4: 01 10 c2 44 and t5,0x10,t0
bh = page->buffers; if (bh) { tmp = bh; do { if (buffer_touched(tmp)) { ^^^^---here clear_bit(BH_Touched, &tmp->b_state); set_bit(PG_referenced, &page->flags); } tmp = tmp->b_this_page; } while (tmp != bh); }
Apparently it was called from try_to_free_page().
If you need any more info, let me know.
Thanks!
Jason Duerstock
| |