netdev - Re: r8169 mac reading/writing broken

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: <4BAE6C92.2060801@iki.fi>
Date:	Sat, 27 Mar 2010 22:37:38 +0200
From:	Timo Teräs <timo.teras@....fi>
To:	François Romieu <romieu@...zoreil.com>
CC:	Ivan Vecera <ivecera@...hat.com>, netdev@...r.kernel.org
Subject: Re: r8169 mac reading/writing broken

Timo Teräs wrote:
> François Romieu wrote:
>> Timo Teräs <timo.teras@....fi> :
>> [...]
>>> It seems that adding single printk between writing MAC0 and MAC4 
>>> fixes it.
>>> I guess it needs a bit of delay between the writes or something.
>>
>> Can you test with a single RTL_R32 after each MACx write ?
> 
> Adding reading back of the written value fixes it too. Though,
> disassembly says that it added an extra instructions also (needs to
> load the 'high' from stack before writing it) so the added delay is
> probably slightly more than just the io read.

I'm not too familiar with PCI details, but this smells a bit
like that write-combining is happening and the NIC does not like
that.

Any ideas how to check this?

The system experiencing this is a "VIA Eden 1.2Ghz" box.

Or is swapping MAC0/MAC4 writes, or adding the extra read an
acceptable fix/workaround?

- Timo
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html