<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="/source/rss.xsl.xml"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/">
<channel>
    <title>Changes in loop.h</title>
    <description></description>
    <language>en</language>
    <copyright>Copyright 2015</copyright>
    <generator>Java</generator><item>
        <title>033af2b3eb19c5ed96825572105bca3611635ada - io_uring: introduce callback driven main loop</title>
        <link>http://kernelsources.org:8080/source/history/linux/io_uring/loop.h#033af2b3eb19c5ed96825572105bca3611635ada</link>
        <description>io_uring: introduce callback driven main loopThe io_uring_enter() has a fixed order of execution: it submitsrequests, waits for completions, and returns to the user. Allow tooptionally replace it with a custom loop driven by a callback calledloop_step. The basic requirements to the callback is that it should beable to submit requests, wait for completions, parse them and repeat.Most of the communication including parameter passing can be implementedvia shared memory.The callback should return IOU_LOOP_CONTINUE to continue execution orIOU_LOOP_STOP to return to the user space. Note that the kernel maydecide to prematurely terminate it as well, e.g. in case the process wassignalled or killed.The hook takes a structure with parameters. It can be used to ask thekernel to wait for CQEs by setting cq_wait_idx to the CQE index it wantsto wait for. Spurious wake ups are possible and even likely, the callbackis expected to handle it. There will be more parameters in the futurelike timeout.It can be used with kernel callbacks, for example, as a slow pathdeprecation mechanism overwiting SQEs and emulating the wantedbehaviour, however it&apos;s more useful together with BPF programsimplemented in following patches.Note that keeping it separately from the normal io_uring wait loopmakes things much simpler and cleaner. It keeps it in one place insteadof spreading a bunch of checks in different places including disablingthe submission path. It holds the lock by default, which is a better fitfor BPF synchronisation and the loop execution model. It nicely avoidsexisting quirks like forced wake ups on timeout request completion. Andit should be easier to implement new features.Signed-off-by: Pavel Begunkov &lt;asml.silence@gmail.com&gt;Link: https://patch.msgid.link/a2d369aa1c9dd23ad7edac9220cffc563abcaed6.1772109579.git.asml.silence@gmail.comSigned-off-by: Jens Axboe &lt;axboe@kernel.dk&gt;

            List of files:
            /linux/io_uring/loop.h</description>
        <pubDate>Thu, 26 Feb 2026 13:48:38 +0100</pubDate>
        <dc:creator>Pavel Begunkov &lt;asml.silence@gmail.com&gt;</dc:creator>
    </item>
</channel>
</rss>
