Re: [PATCH] blk queue io tracing support

2005-09-02 Thread Jens Axboe
On Wed, Aug 31 2005, Nathan Scott wrote: > Hi Jens, > > On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: > > Ok, updated version. > > One thing I found a bit awkward was the way its putting all inodes > in the root of the relayfs namespace, with the cpuid tacked on the > end of the

Re: [PATCH] blk queue io tracing support

2005-09-02 Thread Jens Axboe
On Wed, Aug 31 2005, Nathan Scott wrote: Hi Jens, On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: Ok, updated version. One thing I found a bit awkward was the way its putting all inodes in the root of the relayfs namespace, with the cpuid tacked on the end of the bdevname -

Re: [PATCH] blk queue io tracing support

2005-08-31 Thread Jens Axboe
On Wed, Aug 31 2005, Nathan Scott wrote: > Hi Jens, > > On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: > > Ok, updated version. > > One thing I found a bit awkward was the way its putting all inodes > in the root of the relayfs namespace, with the cpuid tacked on the > end of the

Re: [PATCH] blk queue io tracing support

2005-08-31 Thread Jens Axboe
On Wed, Aug 31 2005, Nathan Scott wrote: > Hi Jens, > > On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: > > Patch attached is against 2.6.13-rc6-mm2. Still a good idea to apply the > > relayfs read update from the previous mail [*] as well. > > There's a small memory leak there on

Re: [PATCH] blk queue io tracing support

2005-08-31 Thread Jens Axboe
On Wed, Aug 31 2005, Nathan Scott wrote: Hi Jens, On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: Patch attached is against 2.6.13-rc6-mm2. Still a good idea to apply the relayfs read update from the previous mail [*] as well. There's a small memory leak there on one of the

Re: [PATCH] blk queue io tracing support

2005-08-31 Thread Jens Axboe
On Wed, Aug 31 2005, Nathan Scott wrote: Hi Jens, On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: Ok, updated version. One thing I found a bit awkward was the way its putting all inodes in the root of the relayfs namespace, with the cpuid tacked on the end of the bdevname -

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Tom Zanussi
Nathan Scott writes: > Hi Tom, > > On Tue, Aug 30, 2005 at 11:19:04PM -0500, Tom Zanussi wrote: > > You're right, it should be using simple_rmdir rather than > > simple_unlink for removing directories. Thanks for sending the patch, > > No problem. > > > which I've modified a bit to

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Nathan Scott
On Wed, Aug 31, 2005 at 02:33:10PM +1000, Nathan Scott wrote: > ... > On an unrelated note, are there any known issues with using epoll > on relayfs file descriptors? I'm having a few troubles, and just > wondering if its me doing something silly, or if its known to not > work...? Symptoms of

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Nathan Scott
Hi Tom, On Tue, Aug 30, 2005 at 11:19:04PM -0500, Tom Zanussi wrote: > You're right, it should be using simple_rmdir rather than > simple_unlink for removing directories. Thanks for sending the patch, No problem. > which I've modified a bit to avoid splitting the rmdir/unlink cases > into

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Tom Zanussi
Nathan Scott writes: > Hi there, > > On Wed, Aug 31, 2005 at 09:48:23AM +1000, Nathan Scott wrote: > > ... > > # find /relay > > /relay > > /relay/block > > /relay/block/sdd > > /relay/block/sdd/trace3 > > /relay/block/sdd/trace2 > > /relay/block/sdd/trace1 > > /relay/block/sdd/trace0

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Nathan Scott
Hi there, On Wed, Aug 31, 2005 at 09:48:23AM +1000, Nathan Scott wrote: > ... > # find /relay > /relay > /relay/block > /relay/block/sdd > /relay/block/sdd/trace3 > /relay/block/sdd/trace2 > /relay/block/sdd/trace1 > /relay/block/sdd/trace0 > /relay/block/sdb > /relay/block/sdb/trace3 >

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Nathan Scott
Hi Jens, On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: > Ok, updated version. One thing I found a bit awkward was the way its putting all inodes in the root of the relayfs namespace, with the cpuid tacked on the end of the bdevname - I was a bit confused at first when a trace of

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Nathan Scott
Hi Jens, On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: > Patch attached is against 2.6.13-rc6-mm2. Still a good idea to apply the > relayfs read update from the previous mail [*] as well. There's a small memory leak there on one of the start-tracing error paths (relay_open

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Nathan Scott
Hi Jens, On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: Patch attached is against 2.6.13-rc6-mm2. Still a good idea to apply the relayfs read update from the previous mail [*] as well. There's a small memory leak there on one of the start-tracing error paths (relay_open

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Nathan Scott
Hi Jens, On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: Ok, updated version. One thing I found a bit awkward was the way its putting all inodes in the root of the relayfs namespace, with the cpuid tacked on the end of the bdevname - I was a bit confused at first when a trace of sdd

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Nathan Scott
Hi there, On Wed, Aug 31, 2005 at 09:48:23AM +1000, Nathan Scott wrote: ... # find /relay /relay /relay/block /relay/block/sdd /relay/block/sdd/trace3 /relay/block/sdd/trace2 /relay/block/sdd/trace1 /relay/block/sdd/trace0 /relay/block/sdb /relay/block/sdb/trace3

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Tom Zanussi
Nathan Scott writes: Hi there, On Wed, Aug 31, 2005 at 09:48:23AM +1000, Nathan Scott wrote: ... # find /relay /relay /relay/block /relay/block/sdd /relay/block/sdd/trace3 /relay/block/sdd/trace2 /relay/block/sdd/trace1 /relay/block/sdd/trace0 /relay/block/sdb

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Nathan Scott
Hi Tom, On Tue, Aug 30, 2005 at 11:19:04PM -0500, Tom Zanussi wrote: You're right, it should be using simple_rmdir rather than simple_unlink for removing directories. Thanks for sending the patch, No problem. which I've modified a bit to avoid splitting the rmdir/unlink cases into separate

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Nathan Scott
On Wed, Aug 31, 2005 at 02:33:10PM +1000, Nathan Scott wrote: ... On an unrelated note, are there any known issues with using epoll on relayfs file descriptors? I'm having a few troubles, and just wondering if its me doing something silly, or if its known to not work...? Symptoms of the

Re: [PATCH] blk queue io tracing support

2005-08-30 Thread Tom Zanussi
Nathan Scott writes: Hi Tom, On Tue, Aug 30, 2005 at 11:19:04PM -0500, Tom Zanussi wrote: You're right, it should be using simple_rmdir rather than simple_unlink for removing directories. Thanks for sending the patch, No problem. which I've modified a bit to avoid

Re: [PATCH] blk queue io tracing support

2005-08-28 Thread Jens Axboe
On Mon, Aug 29 2005, Nathan Scott wrote: > On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: > > ... > > Patch attached is against 2.6.13-rc6-mm2. Still a good idea to apply the > > relayfs read update from the previous mail [*] as well. > > Hi Jens, > > There's a minor config botch in

Re: [PATCH] blk queue io tracing support

2005-08-28 Thread Nathan Scott
On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: > ... > Patch attached is against 2.6.13-rc6-mm2. Still a good idea to apply the > relayfs read update from the previous mail [*] as well. Hi Jens, There's a minor config botch in there, I get this: scripts/kconfig/conf -s

Re: [PATCH] blk queue io tracing support

2005-08-28 Thread Nathan Scott
On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: ... Patch attached is against 2.6.13-rc6-mm2. Still a good idea to apply the relayfs read update from the previous mail [*] as well. Hi Jens, There's a minor config botch in there, I get this: scripts/kconfig/conf -s

Re: [PATCH] blk queue io tracing support

2005-08-28 Thread Jens Axboe
On Mon, Aug 29 2005, Nathan Scott wrote: On Wed, Aug 24, 2005 at 11:28:39AM +0200, Jens Axboe wrote: ... Patch attached is against 2.6.13-rc6-mm2. Still a good idea to apply the relayfs read update from the previous mail [*] as well. Hi Jens, There's a minor config botch in there, I

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Jens Axboe
On Wed, Aug 24 2005, Jens Axboe wrote: > On Wed, Aug 24 2005, Nathan Scott wrote: > > On Wed, Aug 24, 2005 at 09:08:10AM +0200, Jens Axboe wrote: > > > ... > > > This isn't msec precision, it's usec. sched_clock() is in ns! I already > > > decided that msec is too coarse, but usec _should_ be

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Jens Axboe
On Wed, Aug 24 2005, Nathan Scott wrote: > On Wed, Aug 24, 2005 at 09:08:10AM +0200, Jens Axboe wrote: > > ... > > This isn't msec precision, it's usec. sched_clock() is in ns! I already > > decided that msec is too coarse, but usec _should_ be enough. > > Right you are (I was thinking

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Nathan Scott
On Wed, Aug 24, 2005 at 09:08:10AM +0200, Jens Axboe wrote: > ... > This isn't msec precision, it's usec. sched_clock() is in ns! I already > decided that msec is too coarse, but usec _should_ be enough. Right you are (I was thinking m-for-micro, not m-for-milli in my head ;) - but still, there

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Jens Axboe
On Wed, Aug 24 2005, Nathan Scott wrote: > On Tue, Aug 23, 2005 at 02:32:36PM +0200, Jens Axboe wrote: > > Hi, > > > > This is a little something I have played with. It allows you to see > > exactly what is going on in the block layer for a given queue. Currently > > it can logs request queueing

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Jens Axboe
On Wed, Aug 24 2005, Nathan Scott wrote: > Hi Jens, > > On Tue, Aug 23, 2005 at 02:32:36PM +0200, Jens Axboe wrote: > > ... > > + t.pid = current->pid; > > ... > > +/* > > + * The trace itself > > + */ > > +struct blk_io_trace { > > + u32 magic; /* MAGIC << 8 | version

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Nathan Scott
Hi Jens, On Tue, Aug 23, 2005 at 02:32:36PM +0200, Jens Axboe wrote: > ... > + t.pid = current->pid; > ... > +/* > + * The trace itself > + */ > +struct blk_io_trace { > + u32 magic; /* MAGIC << 8 | version */ > + u32 sequence; /* event number */ > +

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Nathan Scott
Hi Jens, On Tue, Aug 23, 2005 at 02:32:36PM +0200, Jens Axboe wrote: ... + t.pid = current-pid; ... +/* + * The trace itself + */ +struct blk_io_trace { + u32 magic; /* MAGIC 8 | version */ + u32 sequence; /* event number */ + u64 time;

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Jens Axboe
On Wed, Aug 24 2005, Nathan Scott wrote: Hi Jens, On Tue, Aug 23, 2005 at 02:32:36PM +0200, Jens Axboe wrote: ... + t.pid = current-pid; ... +/* + * The trace itself + */ +struct blk_io_trace { + u32 magic; /* MAGIC 8 | version */ + u32 sequence;

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Jens Axboe
On Wed, Aug 24 2005, Nathan Scott wrote: On Tue, Aug 23, 2005 at 02:32:36PM +0200, Jens Axboe wrote: Hi, This is a little something I have played with. It allows you to see exactly what is going on in the block layer for a given queue. Currently it can logs request queueing and

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Nathan Scott
On Wed, Aug 24, 2005 at 09:08:10AM +0200, Jens Axboe wrote: ... This isn't msec precision, it's usec. sched_clock() is in ns! I already decided that msec is too coarse, but usec _should_ be enough. Right you are (I was thinking m-for-micro, not m-for-milli in my head ;) - but still, there

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Jens Axboe
On Wed, Aug 24 2005, Nathan Scott wrote: On Wed, Aug 24, 2005 at 09:08:10AM +0200, Jens Axboe wrote: ... This isn't msec precision, it's usec. sched_clock() is in ns! I already decided that msec is too coarse, but usec _should_ be enough. Right you are (I was thinking m-for-micro, not

Re: [PATCH] blk queue io tracing support

2005-08-24 Thread Jens Axboe
On Wed, Aug 24 2005, Jens Axboe wrote: On Wed, Aug 24 2005, Nathan Scott wrote: On Wed, Aug 24, 2005 at 09:08:10AM +0200, Jens Axboe wrote: ... This isn't msec precision, it's usec. sched_clock() is in ns! I already decided that msec is too coarse, but usec _should_ be enough.

Re: [PATCH] blk queue io tracing support

2005-08-23 Thread Nathan Scott
On Tue, Aug 23, 2005 at 02:32:36PM +0200, Jens Axboe wrote: > Hi, > > This is a little something I have played with. It allows you to see > exactly what is going on in the block layer for a given queue. Currently > it can logs request queueing and building, dispatches, requeues, and >

Re: [PATCH] blk queue io tracing support

2005-08-23 Thread Nathan Scott
On Tue, Aug 23, 2005 at 02:32:36PM +0200, Jens Axboe wrote: Hi, This is a little something I have played with. It allows you to see exactly what is going on in the block layer for a given queue. Currently it can logs request queueing and building, dispatches, requeues, and completions. Ah,