From: David Miller <davem@davemloft.net>
To: mirq-linux@rere.qmqm.pl
Cc: netdev@vger.kernel.org
Subject: Re: [PATCH v2 00/46] Clean up RX copybreak and DMA handling
Date: Mon, 11 Jul 2011 02:24:03 -0700 (PDT) [thread overview]
Message-ID: <20110711.022403.532062888492669176.davem@davemloft.net> (raw)
In-Reply-To: <20110711091649.GA6380@rere.qmqm.pl>
From: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Date: Mon, 11 Jul 2011 11:16:49 +0200
> Packet is indicated in queue N, theres no memory for new skb, so its
> dropped, and the buffer goes back to free list. In parallel, queue M
> (!= N) indicates new packet. Still, there's no memory for new skb so
> its also dropped and its buffer is reused. The effect is that all
> packets are dropped, whatever queue they appear on.
Why would queue M (!= N) fail just because N did? They may be
allocating out of different NUMA nodes, and thus succeed.
> The HOL blocking does not matter here, because there's only one head
> --- the system memory. If I misunderstood this point, please explain
> it further.
Multiqueue drivers are moving towards placing the queues on different
NUMA nodes, and in that scenerio one queue might succeed even if the
other fails.
Back to the hardware hanging issue, it's real. Getting into a
situation where the RX ring lacks any buffers at all is the least
tested path for these chips.
Testing fate is a really bad idea, and this is why I always propose to
keep the hardware with RX buffers to use in all circumstances.
next prev parent reply other threads:[~2011-07-11 9:24 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-07-11 0:52 [PATCH v2 00/46] Clean up RX copybreak and DMA handling Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 06/46] net/tokenring: 3c359: fix DMA API usage Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 02/46] net: wrap common patterns of rx handler code Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 01/46] net: introduce __netdev_alloc_skb_aligned() Michał Mirosław
2011-07-11 5:46 ` [PATCH net-next-2.6] net: introduce build_skb() Eric Dumazet
2011-07-11 10:53 ` Michał Mirosław
2011-07-12 15:40 ` Eric Dumazet
2011-07-12 15:54 ` Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 05/46] net: bnx2x: fix DMA sync direction Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 03/46] net drivers: remove unnecessary dma_sync_to_device(DMA_FROM_DEVICE) Michał Mirosław
2011-07-11 8:30 ` Vlad Zolotarov
2011-07-11 9:29 ` Michał Mirosław
2011-07-11 9:46 ` Vlad Zolotarov
2011-07-11 0:52 ` [PATCH v2 04/46] net/wireless: p54: remove useless dma_sync_single_for_device(DMA_FROM_DEVICE) Michał Mirosław
[not found] ` <86c8bde08b005ca7eb4806ea77aec1f3212d63fc.1310339688.git.mirq-linux-CoA6ZxLDdyEEUmgCuDUIdw@public.gmane.org>
2011-07-11 15:15 ` Pavel Roskin
2011-07-12 4:50 ` Felix Fietkau
2011-07-11 0:52 ` [PATCH v2 08/46] net/wireless: b43: fix DMA direction for RX buffers Michał Mirosław
[not found] ` <cover.1310339688.git.mirq-linux-CoA6ZxLDdyEEUmgCuDUIdw@public.gmane.org>
2011-07-11 0:52 ` [PATCH v2 07/46] net/wireless: ath9k: fix DMA API usage Michał Mirosław
2011-07-12 4:36 ` Felix Fietkau
[not found] ` <4E1BCF36.2010506-p3rKhJxN3npAfugRpC6u6w@public.gmane.org>
2011-07-12 5:30 ` [ath9k-devel] " Ben Greear
2011-07-12 9:55 ` Michał Mirosław
2011-07-12 12:54 ` Felix Fietkau
[not found] ` <B4765EFC-B5C9-4E2D-BE00-ED5519D13A4E-Vt+b4OUoWG0@public.gmane.org>
2011-07-12 13:03 ` [ath9k-devel] " Michał Mirosław
[not found] ` <20110712130316.GA8621-CoA6ZxLDdyEEUmgCuDUIdw@public.gmane.org>
2011-07-12 14:21 ` Felix Fietkau
2011-07-12 15:58 ` Michał Mirosław
[not found] ` <20110712155849.GB10651-CoA6ZxLDdyEEUmgCuDUIdw@public.gmane.org>
2011-07-12 16:04 ` Felix Fietkau
2011-07-12 19:13 ` Michał Mirosław
2011-07-12 19:32 ` Ralf Baechle
[not found] ` <20110712193204.GB13413-6z/3iImG2C8G8FEW9MqTrA@public.gmane.org>
2011-07-12 20:53 ` Michał Mirosław
[not found] ` <20110712205316.GA13503-CoA6ZxLDdyEEUmgCuDUIdw@public.gmane.org>
2011-07-12 20:59 ` Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 09/46] net: octeon_mgmt: fix DMA unmap size Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 10/46] net: jme: convert to generic DMA API Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 11/46] net: sungem: cleanup RX skb allocation Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 12/46] net: sunhme: " Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 13/46] net: sunbmac: " Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 46/46] net: mark drivers that drop packets from rx queue head under memory pressure Michał Mirosław
2011-07-11 5:40 ` Francois Romieu
2011-07-11 6:47 ` Eilon Greenstein
2011-07-11 10:04 ` Michał Mirosław
2011-07-11 10:16 ` Eilon Greenstein
2011-07-11 15:24 ` Stephen Hemminger
2011-07-11 0:52 ` [PATCH v2 16/46] net: cxgb3: don't drop packets on memory pressure in driver Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 15/46] net/wireless: b43: use kfree_skb() for untouched skbs Michał Mirosław
2011-07-11 0:52 ` [PATCH v2 14/46] net: sunbmac: cleanup magic '34' Michał Mirosław
2011-07-11 6:54 ` [PATCH v2 00/46] Clean up RX copybreak and DMA handling David Miller
2011-07-11 9:16 ` Michał Mirosław
2011-07-11 9:24 ` David Miller [this message]
2011-07-11 9:47 ` Michał Mirosław
2011-07-11 10:11 ` David Miller
2011-07-11 11:17 ` Michał Mirosław
2011-07-11 12:36 ` Ben Hutchings
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=20110711.022403.532062888492669176.davem@davemloft.net \
--to=davem@davemloft.net \
--cc=mirq-linux@rere.qmqm.pl \
--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).