[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH 00/12] nbd improvements
From: |
Kevin Wolf |
Subject: |
Re: [Qemu-devel] [PATCH 00/12] nbd improvements |
Date: |
Fri, 09 Sep 2011 12:42:57 +0200 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:6.0) Gecko/20110816 Thunderbird/6.0 |
Am 09.09.2011 12:29, schrieb Paolo Bonzini:
> On 09/09/2011 11:00 AM, Kevin Wolf wrote:
>> There is anonther patch enabling AIO for NBD on the list [1], by
>> Nicholas Thomas (CCed), that lacked review so far. Can you guys please
>> review each others approach and then converge on a solution? I guess
>> Paolo's patches 1-7 can be applied in any case, probably causing minor
>> conflicts, but for the rest we need to decide which one to pick.
>
> Stefan also pointed me to Nicholas's patches yesterday. I would go with
> mine, if only because his work predates coroutines (at least the older
> versions) and are much more complex.
>
> On the other hand, Nicholas's work includes timeout and reconnect. I'm
> not sure how complicated it is to include it in my series, but probably
> not much. With coroutines, preserving the list of outstanding I/O
> requests is done implicitly by the CoMutex, so you basically have to
> check errno for ECONNRESET and similar errors, reconnect, and retry
> issuing the current request only.
>
> Timeout can be done with a QEMUTimer that shuts down the socket
> (shutdown(2) I mean). This triggers an EPIPE when writing, or a
> zero-sized read when reading. The timeout can be set every time the
> coroutine is (re)entered, and reset before exiting nbd_co_readv/writev.
>
> What do you think?
I haven't really had a look at your patches yet (I hope to do so later
today), but from the patch descriptions I would think that indeed your
patches could be the better base for a converged series.
What I would like you and Nick to do is to see what is missing from your
series compared to his one (you describe a couple of things, but let's
see if Nick knows anything else) and to agree on the next steps. I think
that a possible way is to merge your series and do the other
improvements on top, but as I said I haven't really looked into the
details yet.
Kevin
- Re: [Qemu-devel] [PATCH v2 09/12] sheepdog: move coroutine send/recv function to generic code, (continued)
- [Qemu-devel] [PATCH 10/12] block: add bdrv_co_flush support, Paolo Bonzini, 2011/09/08
- [Qemu-devel] [PATCH 04/12] nbd: add support for NBD_CMD_FLUSH, Paolo Bonzini, 2011/09/08
- [Qemu-devel] [PATCH 07/12] sheepdog: add coroutine_fn markers, Paolo Bonzini, 2011/09/08
- [Qemu-devel] [PATCH 05/12] nbd: add support for NBD_CMD_FLAG_FUA, Paolo Bonzini, 2011/09/08
- Re: [Qemu-devel] [PATCH 00/12] nbd improvements, Kevin Wolf, 2011/09/09
Re: [Qemu-devel] [PATCH 00/12] nbd improvements, Kevin Wolf, 2011/09/14