netdev - Re: Regression since 2.6.30-rc1 - Page allocation failures

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>] [day] [month] [year] [list]

Message-Id: <1241048343.25477.7.camel@johannes.local>
Date:	Thu, 30 Apr 2009 01:39:02 +0200
From:	Johannes Berg <johannes@...solutions.net>
To:	Larry Finger <Larry.Finger@...inger.net>
Cc:	wireless <linux-wireless@...r.kernel.org>,
	netdev <netdev@...r.kernel.org>
Subject: Re: Regression since 2.6.30-rc1 - Page allocation failures

(dropped lkml, added netdev)

> I get the "failed to reallocate TX buffer" message. Not all are related to the
> page allocation failure. I am also getting "b43-phy0 debug: DMA RX:
> setup_rx_descbuffer() failed" messages.
> 
> So far, the numbers printed by the revised printk are:
> 
> kernel: phy0: failed to reallocate TX buffer 0, 404, need 0, 0
> kernel: protocol 0008 is buggy, dev eth1
> 
> kernel: phy0: failed to reallocate TX buffer 0, 1728, need 0, 18
> 
> Device eth1 is the b43 wireless device.

Thanks.

This is a little odd -- I don't see why
	pskb_expand_head(skb, 0, 0)
or
	pskb_expand_head(skb, 0, 18)

would try to do an order 1 allocation when skb->len is 404 or 1728. It
looks like something is forcing larger alignments than should be
necessary, but I have no idea what.

(for netdev: http://mid.gmane.org/49F8C03D.6040909@lwfinger.net has more
information)

johannes

Download attachment "signature.asc" of type "application/pgp-signature" (837 bytes)