netdev - Re: ppoll() stuck on POLLIN while TCP peer is sending

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Fri, 04 Jan 2013 09:15:03 -0800
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Mel Gorman <mgorman@...e.de>
Cc:	Eric Wong <normalperson@...t.net>, linux-mm@...ck.org,
	netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
	Rik van Riel <riel@...hat.com>,
	Minchan Kim <minchan@...nel.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Linus Torvalds <torvalds@...ux-foundation.org>
Subject: Re: ppoll() stuck on POLLIN while TCP peer is sending

On Fri, 2013-01-04 at 16:01 +0000, Mel Gorman wrote:

> Implying that it's stuck in compaction somewhere. It could be the case
> that compaction alters timing enough to trigger another bug. You say it
> tests differently depending on whether TCP or unix sockets are used
> which might indicate multiple problems. However, lets try and see if
> compaction is the primary problem or not.

One difference between TCP or unix socket is that :

Unix sockets try hard to limit the order of allocations.

For a 16KB (+ skb overhead) send(), we will probably use one order-2
page and one order-0 page as a frag (data_len being not 0) :

vi +1484 net/unix/af_unix.c

       if (len > SKB_MAX_ALLOC)
                data_len = min_t(size_t,
                                 len - SKB_MAX_ALLOC,
                                 MAX_SKB_FRAGS * PAGE_SIZE);

        skb = sock_alloc_send_pskb(sk, len - data_len, data_len,
                                   msg->msg_flags & MSG_DONTWAIT, &err);

While TCP could use order-3 pages if available

Eric, you could try to change SKB_FRAG_PAGE_ORDER in net/core/sock.c to
lower values (16384, 8192, 4096) and check if the hang can disappear or
not.

Alternatively (no kernel patching needed), you could try to hang AF_UNIX
using buffers of 90KB, to force order-3 allocations as well (one 32KB
allocation plus 16 * 4KB frags)

Thanks


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html