linux-kernel - Re: [REGRESSION] asix: Lots of asix_rx_fixup() errors and slow transmissions

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, 6 May 2016 10:40:47 -0700
From:	John Stultz <john.stultz@...aro.org>
To:	Dean Jenkins <Dean_Jenkins@...tor.com>
Cc:	Guodong Xu <guodong.xu@...aro.org>,
	lkml <linux-kernel@...r.kernel.org>,
	Mark Craske <Mark_Craske@...tor.com>,
	"David S. Miller" <davem@...emloft.net>,
	YongQin Liu <yongqin.liu@...aro.org>,
	linux-usb@...r.kernel.org, netdev@...r.kernel.org,
	Ivan Vecera <ivecera@...hat.com>,
	"David B. Robins" <linux@...idrobins.net>
Subject: Re: [REGRESSION] asix: Lots of asix_rx_fixup() errors and slow transmissions

On Fri, May 6, 2016 at 8:00 AM, Dean Jenkins <Dean_Jenkins@...tor.com> wrote:
> My conclusion is that your USB to Ethernet Adaptor is not running at high
> speed (480Mbps) mode which is causing a partial loss (corruption) of
> Ethernet frames across the USB link. A USB Protocol Analyser or software
> tool usbmon could be used to confirm this scenario.
>
> Therefore please retest with a working high-speed USB hub or remove the
> full-speed USB hub from the test environment and directly connect the USB to
> Ethernet Adaptor to the root hub of the USB port. Then repeat the tests to
> see whether anything improved.
>
> In other words, you need to eliminate the dmesg messages saying "not running
> at top speed; connect to a high speed hub".

The aarch64 system has a quirk that at the moment limits it to the
slower full-speed mode, which also exacerbates the issue (basically
taking a fairly slow 1.1.Mb/s network connection without your patch,
to an almost unusable 30Kb/s with it).

But that isn't the case on the x86_64 system, which is seeing a very
similar problem (though the performance effect isn't nearly as bad, as
the error rate in time seems relatively similar on both, and I think
my scp transmissions are cpu bound on this atom board :).

thanks
-john