[dm-devel] Re: [PATCH] dm-ioband-v1.12.2: I/O bandwidth controller

[Ryo Tsuruta <ryov valinux co jp>]

Hi Simon,

> Hi Tsuruta-san,
> 
> have you run sparse over this code? There seem to be a few interesting
> warnings generated (and some at the top noise too).

Thank you for your comment.
I'll do sparse every time from now on.
 
> $ make C=1 CF="-D__CHECK_ENDIAN__"
> [snip]
> include/trace/events/dm-ioband.h:9:1: warning: symbol 'ftrace_raw_output_ioband_hold_urgent_bio' was not declared. Should it be static?
[snip ftrace warnings]

> drivers/md/dm-ioband-ctl.c:178:2: warning: context problem in 'suspend_ioband_device': '_spin_unlock_irqrestore' expected different context
> drivers/md/dm-ioband-ctl.c:178:2:    context 'lock': wanted >= 1, got 0
> drivers/md/dm-ioband-ctl.c:600:3: warning: context imbalance in 'prevent_burst_bios': unexpected unlock
> drivers/md/dm-ioband-ctl.c:600:3:    context '<noident>': wanted 0, got -1
> drivers/md/dm-ioband-ctl.c:816:3: warning: context imbalance in 'ioband_map': __context__ statement expected different context
> drivers/md/dm-ioband-ctl.c:816:3:    context '<noident>': wanted >= 0, got -1
> drivers/md/dm-ioband-rangebw.c:287:4: warning: context imbalance in 'range_bw_queue_full': __context__ statement expected different context
> drivers/md/dm-ioband-rangebw.c:287:4:    context '<noident>': wanted >= 0, got -1

These imbalances are not warned by the latest sparse command retrieved
from the git repository, and actually they are not imbalanced.

> drivers/md/dm-ioband-rangebw.c:592:6: warning: symbol 'range_bw_timeover' was not declared. Should it be static?
> drivers/md/dm-ioband-rangebw.c:210: warning: 'io_mode' may be used uninitialized in this function

I'll fix these warnings and post the update soon.

Thanks,
Ryo Tsuruta