From: Al Viro <viro@ZenIV.linux.org.uk>
To: Christoph Hellwig <hch@lst.de>
Cc: Avi Kivity <avi@scylladb.com>,
linux-aio@kvack.org, linux-fsdevel@vger.kernel.org,
netdev@vger.kernel.org, linux-api@vger.kernel.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH 10/32] aio: implement IOCB_CMD_POLL
Date: Sun, 20 May 2018 08:33:39 +0100 [thread overview]
Message-ID: <20180520073332.GA30522@ZenIV.linux.org.uk> (raw)
In-Reply-To: <20180520053219.GZ30522@ZenIV.linux.org.uk>
On Sun, May 20, 2018 at 06:32:25AM +0100, Al Viro wrote:
> > + spin_lock_irqsave(&ctx->ctx_lock, flags);
> > + list_add_tail(&aiocb->ki_list, &ctx->delayed_cancel_reqs);
> > + spin_unlock(&ctx->ctx_lock);
>
> ... and io_cancel(2) comes, finds it and inhume^Wcompletes it, leaving us to...
>
> > + spin_lock(&req->head->lock);
>
> ... get buggered on attempt to dereference a pointer fetched from freed and
> reused object.
FWIW, how painful would it be to pull the following trick:
* insert into wait queue under ->ctx_lock
* have wakeup do schedule_work() with aio_complete() done from that
* have ->ki_cancel() grab queue lock, remove from queue and use
the same schedule_work()
That way you'd get ->ki_cancel() with the same semantics as originally for
everything - "ask politely to finish ASAP", and called in the same locking
environment for everyone - under ->ctx_lock, that is. queue lock nests
inside ->ctx_lock; no magical flags, etc.
The cost is schedule_work() for each async poll-related completion as you
have for fsync. I don't know whether that's too costly or not; it certainly
simplifies the things, but whether it's OK performance-wise...
Comments?
next prev parent reply other threads:[~2018-05-20 7:34 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-15 19:48 aio poll and a new in-kernel poll API V11 Christoph Hellwig
2018-05-15 19:48 ` [PATCH 01/32] fs: unexport poll_schedule_timeout Christoph Hellwig
2018-05-15 19:48 ` [PATCH 02/32] fs: cleanup do_pollfd Christoph Hellwig
2018-05-15 19:48 ` [PATCH 03/32] fs: update documentation to mention __poll_t and match the code Christoph Hellwig
2018-05-15 19:48 ` [PATCH 04/32] fs: add new vfs_poll and file_can_poll helpers Christoph Hellwig
2018-05-15 19:48 ` [PATCH 05/32] fs: introduce new ->get_poll_head and ->poll_mask methods Christoph Hellwig
2018-05-15 19:48 ` [PATCH 06/32] aio: simplify KIOCB_KEY handling Christoph Hellwig
2018-05-15 19:48 ` [PATCH 07/32] aio: simplify cancellation Christoph Hellwig
2018-05-15 19:48 ` [PATCH 08/32] aio: replace kiocb_set_cancel_fn with a cancel_kiocb file operation Christoph Hellwig
2018-05-20 5:27 ` Al Viro
2018-05-15 19:48 ` [PATCH 09/32] aio: add delayed cancel support Christoph Hellwig
2018-05-15 19:48 ` [PATCH 10/32] aio: implement IOCB_CMD_POLL Christoph Hellwig
2018-05-20 5:32 ` Al Viro
2018-05-20 7:33 ` Al Viro [this message]
2018-05-20 17:32 ` Christoph Hellwig
2018-05-15 19:48 ` [PATCH 11/32] net: refactor socket_poll Christoph Hellwig
2018-05-15 19:48 ` [PATCH 12/32] net: add support for ->poll_mask in proto_ops Christoph Hellwig
2018-05-15 19:48 ` [PATCH 13/32] net: remove sock_no_poll Christoph Hellwig
2018-05-15 19:48 ` [PATCH 14/32] net/tcp: convert to ->poll_mask Christoph Hellwig
2018-05-15 19:48 ` [PATCH 15/32] net/unix: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 16/32] net: convert datagram_poll users tp ->poll_mask Christoph Hellwig
2018-05-15 19:48 ` [PATCH 17/32] net/dccp: convert to ->poll_mask Christoph Hellwig
2018-05-15 19:48 ` [PATCH 18/32] net/atm: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 19/32] net/vmw_vsock: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 20/32] net/tipc: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 21/32] net/sctp: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 22/32] net/bluetooth: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 23/32] net/caif: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 24/32] net/nfc: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 25/32] net/phonet: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 26/32] net/iucv: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 27/32] net/rxrpc: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 28/32] crypto: af_alg: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 29/32] pipe: " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 30/32] eventfd: switch " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 31/32] timerfd: convert " Christoph Hellwig
2018-05-15 19:48 ` [PATCH 32/32] random: " Christoph Hellwig
-- strict thread matches above, loose matches on Subject: below --
2018-05-11 11:07 aio poll and a new in-kernel poll API V10 Christoph Hellwig
2018-05-11 11:07 ` [PATCH 10/32] aio: implement IOCB_CMD_POLL Christoph Hellwig
2018-03-30 15:07 aio poll and a new in-kernel poll API V9 Christoph Hellwig
2018-03-30 15:07 ` [PATCH 10/32] aio: implement IOCB_CMD_POLL Christoph Hellwig
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180520073332.GA30522@ZenIV.linux.org.uk \
--to=viro@zeniv.linux.org.uk \
--cc=avi@scylladb.com \
--cc=hch@lst.de \
--cc=linux-aio@kvack.org \
--cc=linux-api@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).