svn commit: r274914 - in head/sys: kern sys ufs/ffs vm

[Ruslan Bukin]

Hi,

There is a theory that a problem is related to CPU caches, and it can be BERI-cpu
specific only.

And sorry I didn't realized virtio was still in the kernel when I was trying
MDROOT (we do dcache operations in virtio).

The workaround I found is to disable all the instruction cache invalidation calls
(there are a few calls on cpu startup only).

Ruslan

On Thu, Nov 27, 2014 at 08:51:04AM -0800, Adrian Chadd wrote:
> Hi,
> 
> Has there been any further progress with this?
> 
> If there hasn't, then I'd like to see it backed out for the time
> being. I'm worried that this has introduced some subtle bugs and this
> isn't yet fully understood.
> 
> Thanks,
> 
> 
> 
> -adrian
> 
> 
> On 25 November 2014 at 09:38, Ruslan Bukin <br at freebsd.org> wrote:
> > On Tue, Nov 25, 2014 at 06:29:22PM +0300, Gleb Smirnoff wrote:
> >> R> Un-unioning helps partially: I receive much less panics and able to boot until login prompt in ~ 1/3 attempts with patch attached.
> >>
> >> What happens then? Same panic as w/o patch?
> >>
> >
> > Same random panics:
> >
> > [..]
> > Starting file system checks:
> > /dev/vtbd0: 344 files, 22588 used, 3419 free (3 frags, 427 blocks, 0.0% fragmentation)
> > MAP: No valid partition found at vtbd0
> > MAP: No valid partition found at md0
> > MAP: No valid partition found at md0
> > panic: vm_page_insert_after: msucc doesn't succeed pindex
> > KDB: enter: panic
> > [ thread pid 53 tid 100030 ]
> > Stopped at      kdb_enter+0x8c: lui     at,0x0
> > db> bt
> > Tracing pid 53 tid 100030 td 0x9800000000b8f000
> > db_trace_thread+40 (?,?,?,?) ra ffffffff80117654 sp c0000000056d90f0 sz 16
> > ffffffff801174c8+18c (0,?,ffffffff,?) ra ffffffff80116730 sp c0000000056d9100 sz 48
> > ffffffff801162a8+488 (?,?,?,?) ra ffffffff801169fc sp c0000000056d9130 sz 192
> > db_command_loop+f4 (?,?,?,?) ra ffffffff8011a570 sp c0000000056d91f0 sz 16
> > ffffffff8011a3c8+1a8 (?,?,?,?) ra ffffffff80282338 sp c0000000056d9200 sz 816
> > kdb_trap+188 (?,?,?,?) ra ffffffff804b10a4 sp c0000000056d9530 sz 64
> > trap+1114 (?,?,?,?) ra ffffffff8049c63c sp c0000000056d9570 sz 288
> > MipsKernGenException+154 (0,4,804f60d0,12f) ra ffffffff80281ed4 sp c0000000056d9690 sz 368
> > kdb_enter+8c (?,?,?,?) ra ffffffff80234398 sp c0000000056d9800 sz 16
> > ffffffff802341e8+1b0 (?,?,?,?) ra ffffffff802344c8 sp c0000000056d9810 sz 32
> > kassert_panic+c8 (?,4,b82c00,3d7) ra ffffffff8047c580 sp c0000000056d9830 sz 96
> > ffffffff8047c460+120 (?,?,?,?) ra ffffffff8047ff28 sp c0000000056d9890 sz 64
> > vm_page_alloc+7f8 (?,?,?,?) ra ffffffff80464524 sp c0000000056d98d0 sz 80
> > ffffffff80463a00+b24 (?,?,?,?) ra 0 sp c0000000056d9920 sz 272
> > pid 53
> >
> > Starting file system checks:
> > /dev/vtbd0: 344 files, 22588 used, 3419 free (3 frags, 427 blocks, 0.0% fragmentation)
> > MAP: No valid partition found at vtbd0
> > MAP: No valid partition found at md0
> > MAP: No valid partition found at md0
> > panic: Bad link elm 0x9800000000bb3000 prev->next != elm
> > KDB: enter: panic
> > [ thread pid 44 tid 100030 ]
> > Stopped at      kdb_enter+0x8c: lui     at,0x0
> > db> bt
> > Tracing pid 44 tid 100030 td 0x9800000000b8f000
> > db_trace_thread+40 (?,?,?,?) ra ffffffff80117654 sp c0000000056d8ea0 sz 16
> > ffffffff801174c8+18c (0,?,ffffffff,?) ra ffffffff80116730 sp c0000000056d8eb0 sz 48
> > ffffffff801162a8+488 (?,?,?,?) ra ffffffff801169fc sp c0000000056d8ee0 sz 192
> > db_command_loop+f4 (?,?,?,?) ra ffffffff8011a570 sp c0000000056d8fa0 sz 16
> > ffffffff8011a3c8+1a8 (?,?,?,?) ra ffffffff80282338 sp c0000000056d8fb0 sz 816
> > kdb_trap+188 (?,?,?,?) ra ffffffff804b10a4 sp c0000000056d92e0 sz 64
> > trap+1114 (?,?,?,?) ra ffffffff8049c63c sp c0000000056d9320 sz 288
> > MipsKernGenException+154 (0,4,804f60d0,12f) ra ffffffff80281ed4 sp c0000000056d9440 sz 368
> > kdb_enter+8c (?,?,?,?) ra ffffffff80234398 sp c0000000056d95b0 sz 16
> > ffffffff802341e8+1b0 (?,?,?,?) ra ffffffff80234400 sp c0000000056d95c0 sz 32
> > panic+30 (?,bb3000,8051f738,8051f738) ra ffffffff804790c8 sp c0000000056d95e0 sz 96
> > vm_object_deallocate+960 (?,?,?,?) ra ffffffff8046b184 sp c0000000056d9640 sz 96
> > ffffffff8046b158+2c (?,?,?,?) ra ffffffff8046b2a4 sp c0000000056d96a0 sz 32
> > ffffffff8046b1e8+bc (?,?,?,?) ra ffffffff8046b308 sp c0000000056d96c0 sz 32
> > _vm_map_unlock+40 (?,?,?,?) ra ffffffff8046d990 sp c0000000056d96e0 sz 16
> > vm_map_remove+98 (?,?,?,?) ra ffffffff801eb2e4 sp c0000000056d96f0 sz 32
> > exec_new_vmspace+274 (?,?,?,?) ra ffffffff801c4ea0 sp c0000000056d9710 sz 80
> > ffffffff801c4828+678 (?,?,?,?) ra ffffffff801eb81c sp c0000000056d9760 sz 304
> > kern_execve+404 (?,?,?,?) ra ffffffff801ec6b4 sp c0000000056d9890 sz 784
> > sys_execve+44 (?,?,?,?) ra ffffffff804b0a88 sp c0000000056d9ba0 sz 80
> > trap+af8 (?,?,?,?) ra ffffffff8049c880 sp c0000000056d9bf0 sz 288
> > MipsUserGenException+13c (?,?,?,600e30b4) ra 0 sp c0000000056d9d10 sz 0
> > pid 44
> > db>
> >
> > Ruslan
> >
>