netdev - Re: [PATCH v3 00/10] split out emac cpdma and mdio for reuse

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:	Thu, 09 Sep 2010 15:51:58 -0400
From:	Cyril Chemparathy <cyril@...com>
To:	Michael Williamson <michael.williamson@...ticallink.com>
CC:	Kevin Hilman <khilman@...prootsystems.com>,
	"tony@...mide.com" <tony@...mide.com>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	"davinci-linux-open-source@...ux.davincidsp.com" 
	<davinci-linux-open-source@...ux.davincidsp.com>,
	"linux-omap@...r.kernel.org" <linux-omap@...r.kernel.org>,
	"davem@...emloft.net" <davem@...emloft.net>
Subject: Re: [PATCH v3 00/10] split out emac cpdma and mdio for reuse

Hi Mike,

[...]
> The hang is in wait_for_user_access() in the davinci_mdio_read() call.  Looks like
> the state machine got put back into IDLE somewhere between the MDIO probe and the 
> EMAC probe. Seems like there should be some sort of time-out and error message 
> in the wait_for_user_access() method.... (maybe even a check for IDLE??)
> 
> If I add a patch to check the state machine for IDLE and then re-enable it in the
> davinci_mdio_read() call, it is able to press on and come up.  I don't see any calls
> to the davinci_mdio_suspend() call, so I am wondering if the EMAC probe routine, 
> particularly the application of the SOFTRESET, is causing the MDIO to drop back to 
> IDLE / disabled.
> 
> I can post the patch if you like, but it is a bit of a hack...

An EMAC soft-reset clobbering the MDIO controller state is a
possibility.  I will poll TI designers to see if this could be the case.

In any case, a couple of unanswered questions remain:

  1. Why don't other davinci devices display similar behavior?

  2. If the answer to #1 above is that the timing window is pretty slim
     (i.e., only if an MDIO read/write is in progress during EMAC
     soft-reset), why do we hit this situation consistently on
     mityomap?

I have put together a quick patch (tested dm365).  See attached.

Regards
Cyril.

View attachment "mdio-check-idle.patch" of type "text/x-patch" (7776 bytes)