Prev: mm: avoid overflowing preempt_count() in mmu_take_all_locks()
Next: Aw: [PATCH] power/ds2782: fix clientdata on removal
From: Kei Tokunaga on 1 Apr 2010 07:40
Hi,

This is v4 patch series of SCSI trace. Only a cosmetic change
from v3.

Because it's been a while since the first posting, I'll put brief
descriptions of the patchset.

Basically, this is a ftrace based SCSI tracer. It prints the
following information for scsi-3 commands.

o scsi bus number ("host")
o scsi channel number ("channel")
o scsi id ("id")
o LUN ("lun")
o command info
- op name (ex. "WRITE_10")
- LBA ("lba") (for certain commands only)
- transfer length ("txlen") (for certain commands only)
- CDB (ex. "raw=28 00 03 60 a2 cb 00 00 08 00")
o result of the command
(ex. "result=(driver=DRIVER_OK host=DID_OK
message=COMMAND_COMPLETE status=SAM_STAT_GOOD")

Those information are printed either at a time:

o when a command is to be dispatched,
o when a command dispatched returns in error,
o when a command is done,
o when an error handler is to waken up, or
o when timeout of a command is deteced.

To use it quickly, try these steps:

1) Build a kernel with the patchset applied
(ftrace related kernel configs need to be enabled.)
2) Boot the kernel
3) Mount debugfs
ex. mount -t debugfs none /sys/kernel/debug/
4) Indicate events to trace
ex. echo "scsi:*" > /sys/kernel/debug/tracing/set_event
5) Make some I/Os on your scsi devices
ex. ls
6) Retrieve trace log
ex. cat /sys/kernel/debug/tracing/trace

Here is an outputs example:

# tracer: nop
#
# TASK-PID CPU# TIMESTAMP FUNCTION
# | | | | |
flush-8:0-3746 [001] 1374.517152:
scsi_dispatch_cmd_start: host_no=0 channel=0 id=0 lun=0
cmnd=(WRITE_10 lba=32267539 txlen=8 raw=2a 00 01 ec 5d 13 00 00
08 00)
<idle>-0 [000] 1374.524964:
scsi_dispatch_cmd_done: host_no=0 channel=0 id=0 lun=0
cmnd=(WRITE_10 lba=32267539 txlen=8 raw=2a 00 01 ec 5d 13 00 00
08 00) result=(driver=DRIVER_OK host=DID_OK
message=COMMAND_COMPLETE status=SAM_STAT_GOOD)

This patchset applies to 2.6.34-rc3.

Thanks,
Kei

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
 | 
Pages: 1
Prev: mm: avoid overflowing preempt_count() in mmu_take_all_locks()
Next: Aw: [PATCH] power/ds2782: fix clientdata on removal