linux-kernel - Re: cxacru usb_bulk_msg() firmware upload 36x slower with OHCI vs. UHCI

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]

Message-ID: <4B046E4C.7040205@simon.arlott.org.uk>
Date:	Wed, 18 Nov 2009 21:59:40 +0000
From:	Simon Arlott <simon@...e.lp0.eu>
To:	Alan Stern <stern@...land.harvard.edu>
CC:	USB list <linux-usb@...r.kernel.org>,
	Kernel development list <linux-kernel@...r.kernel.org>
Subject: Re: cxacru usb_bulk_msg() firmware upload 36x slower with OHCI vs.
 UHCI

On 18/11/09 21:28, Alan Stern wrote:
> If you don't have a good kernel to start from then there's nothing to
> search for.  I was assuming that a relatively recent kernel change
> might have caused the slow-down, but if it has been this way for a long
> time then a different approach is needed.

The earliest kernel I've had this happen with is 2.6.28-rc2, when I first
started using it with an OHCI controller.

> There's no particular significance to 256 ms that I know of, although
> if the hardware is malfunctioning it could easily pick such a rate.  
> The page size is significant because that's how the driver and the
> hardware divide up transfers; each Transfer Descriptor refers to at 
> most 4096 bytes of data.
> 
> On Wed, 18 Nov 2009, Simon Arlott wrote:
>> The OHCI code appears to split the data up into 4096 chunks, but even the
>> odd sized transfer of 25280 bytes at the end runs at the same speed:
>> 
>> [ 4774.830569] cxacru: sending fw 0x3 size 0x62c0 to #98668
>> [ 4776.410375] cxacru: sending fw 0x3 size 0x100 to #e0
> 
> Yep.  My intuition says "hardware problem", but there's no hard 
> evidence one way or another.
> 
> What happens with other sorts of devices, such as a USB flash drive?
 
On my desktop, I don't usually use the OHCI controller directly, but
through hubs on EHCI. I do have an USB Audio device which works ok
using OHCI.

The firmware load fails over EHCI (via a hub) despite running even faster:
[89633.219403] cxacru: sending fw 0x3 size 0x8000 to #7000
[89633.220147] cxacru: sending fw 0x3 size 0x8000 to #e000
[89633.220530] cxacru: sending fw 0x3 size 0x8000 to #15000
[89633.220774] cxacru: sending fw 0x3 size 0x8000 to #1c000
[89633.220898] cxacru: sending fw 0x3 size 0x8000 to #23000
[89633.221286] cxacru: sending fw 0x3 size 0x8000 to #2a000

338µs per 32KB... which is 775mbit/s+ and doesn't make sense.

I can write a 10MB file to an USB flash drive over OHCI, and umount+sync
takes around 13 seconds.

-- 
Simon Arlott
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/