Re: panic with CF drive + USB reader

Bernd,
> . 
> Seems that handling the stalled condition failed.
> Can you try the following patch:

I was *just* going to write you saying it was still running with this 
patch, which would have been the longest yet. But then I saw these messages:

umass0: BBB reset failed, STALLED
(da2:umass-sim0:0:0:0): AutoSense Failed
umass0: BBB reset failed, STALLED
(da2:umass-sim0:0:0:0): AutoSense Failed
initiate_write_filepage: already started
initiate_write_filepage: already started
initiate_write_filepage: already started
umass0: BBB reset failed, STALLED
(da2:umass-sim0:0:0:0): AutoSense Failed
umass0: BBB reset failed, STALLED
(da2:umass-sim0:0:0:0): AutoSense Failed
initiate_write_filepage: already started

And the box paniced again:

panic: initiate_write_inodeblock_ufs2: already started
cpuid = 0; lapic.id = 00000000
Stack backtrace:
backtrace(c033157d,0,c0341a7f,e96a1ab0,100) at backtrace+0x17
panic(c0341a7f,c01bc427,c7dabdf4,e96a1ad0,e96a1ad0) at panic+0x13d
initiate_write_inodeblock_ufs2(c6f4cb00,d29060b8,c01fd879,0,d29060b8) at 
initiate_write_inodeblock_ufs2+0x68f
softdep_disk_io_initiation(d29060b8,c17fbf30,e96a1bb4,c01fda2b,d29060b8) 
at softdep_disk_io_initiation+0x80
spec_xstrategy(c6d88490,d29060b8,e96a1bb4,c017dfcc,e96a1be0) at 
spec_xstrategy+0x104
spec_specstrategy(e96a1be0,e96a1bfc,c01f8550,e96a1be0,1) at 
spec_specstrategy+0x1b
spec_vnoperate(e96a1be0,1,c6bfc800,d2a5f680,e96a1bdc) at spec_vnoperate+0x18
bwrite(d29060b8,1000000,c63304c0,e96a1c3c,80012) at bwrite+0x403
vfs_bio_awrite(d29060b8,80012,0,c63304c0,c017dfcc) at vfs_bio_awrite+0x27b
vop_stdfsync(e96a1cdc,0,c6d88490,e96a1ca4,c017dfcc) at vop_stdfsync+0x1b1
spec_fsync(e96a1cdc,e96a1d18,c020b613,e96a1cdc,20002) at spec_fsync+0x31
spec_vnoperate(e96a1cdc,20002,c63304c0,c0335c1b,0) at spec_vnoperate+0x18
sched_sync(0,e96a1d48,0,0,0) at sched_sync+0x1e3
fork_exit(c020b430,0,e96a1d48) at fork_exit+0xb2
fork_trampoline() at fork_trampoline+0x8
--- trap 0x1, eip = 0, esp = 0xe96a1d7c, ebp = 0 ---
Debugger("panic")
Stopped at      Debugger+0x4f:  xchgl   %ebx,in_Debugger.0

It seems that this happens under light or no load only. I had been doing 
a heavy cycle of dump/restore/dd, etc, and all was well. Could it be 
that the MicroDrive does some kind of internal power management that 
delays its reponses some, and the kernel doesn't expect that?

Lars
-- 
Lars Eggert <larse_at_isi.edu>           USC Information Sciences Institute