[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180628150218.GA12970@localhost.localdomain>
Date: Thu, 28 Jun 2018 09:02:18 -0600
From: Keith Busch <keith.busch@...ux.intel.com>
To: Sagi Grimberg <sagi@...mberg.me>
Cc: Keith Busch <keith.busch@...el.com>,
Christoph Hellwig <hch@....de>,
Linux Kernel Mailinglist <linux-kernel@...r.kernel.org>,
Linux NVMe Mailinglist <linux-nvme@...ts.infradead.org>,
Johannes Thumshirn <jthumshirn@...e.de>
Subject: Re: [PATCH] nvme: trace: add disk name to tracepoints
On Thu, Jun 28, 2018 at 05:34:07PM +0300, Sagi Grimberg wrote:
>
> > Thinking more on this, not using the hw qid really limits the utility
> > out of using these trace events: We may not be able to match a completion
> > to the submission without it since cmdid alone isn't enough to match up
> > the two events.
> >
> > Here's an updated proposal and actually tested. I was also able to
> > combine admin and io submissions.
>
> This looks good!
>
> One thing that is missing is the controller id when we have multiple
> controllers/subsystems in the host like:
>
> TP_printk("nvme%d: qid=%d, nsid=%u, cmdid=%u, flags=0x%x, meta=0x%llx,
> cmd=(%s %s)",
>
> But Johannes can add it I guess.
Yep, I'm sort of going off on a tangent from what Johannes was adding.
We definitely need both. I have some time today, I'll send out the a
series with the combined results.
Powered by blists - more mailing lists