netdev - Re: brcm80211 breakage..

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, 13 Jan 2012 07:50:41 +0100
From:	Rafał Miłecki <zajec5@...il.com>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	Arend van Spriel <arend@...adcom.com>,
	Larry Finger <Larry.Finger@...inger.net>,
	Alwin Beukers <alwin@...adcom.com>,
	Roland Vossen <rvossen@...adcom.com>,
	"John W. Linville" <linville@...driver.com>,
	Network Development <netdev@...r.kernel.org>,
	"Franky (Zhenhui) Lin" <frankyl@...adcom.com>
Subject: Re: brcm80211 breakage..

W dniu 13 stycznia 2012 06:34 użytkownik Linus Torvalds
<torvalds@...ux-foundation.org> napisał:
> 2012/1/12 Linus Torvalds <torvalds@...ux-foundation.org>:
>> 2012/1/12 Linus Torvalds <torvalds@...ux-foundation.org>:
>>>
>>> I'm surprised it has ever worked for anybody. It certainly doesn't work for me.
>>
>> So I can suspend the bcma driver on its own until the cows come home.
>>
>> But after I have suspended the bcma driver even once, just doing a
>> "modprobe brcmsmac" will hang the machine hard. Dunno where, but this
>> is probably the same thing as "hangs on resume".
>
> Guys, has suspend/resume with the bcma interface been tested AT ALL?
>
> The suspend/resume fields of "struct bcma_driver" are COMPLETELY
> UNUSED. The only place in the kernel that uses them is the brcmsmac
> driver that does this write-only assignment:
>
>        .suspend  = brcms_suspend,
>        .resume   = brcms_resume,
>
> nothing else uses them. NOTHING. I tested by just removing the fields
> and compiling the bcma subsystem, just in case there was something
> really subtle that I was missing and was hidden through some magic
> hidden approach. But no.
>
> Seriously - how was something that isn't even connected ever supposed
> to work at all? And why was the BCMA conversion of that driver sent
> up-stream if something as fundamental as suspend/resume had never been
> done, and didn't actually work?
>
> What am I missing now? How the hell can this ever have worked for
> ANYBODY? What kind of f*&*ing sick joke is this all?

The suspend&resume wasn't implemented for some time because my PC
doesn't s&r. And I don't have access to notebook with mini PCIe slot.

I've implemented support for s&r in bcma when I got to open my Sony
VAIO to replace A/C power slot. It was one time I was able to change
WiFi card in my notebook which has really-ugly-hidden mini PCIe slot.

S&r was working fine for me with bcma&b43 after writing that patch!
That includes suspending and resuming multiple times. And tests were
done with the same card you're using.

The lock up on (resume|loading brcmsmac) means bus wasn't initialized
correctly after resume. It does not have to be brcmsmac bug. We're
accessing some registers before they're ready.

Linus: can you do one trivial test for me? Please simply try unloading
bcma before suspending. Then resume and load bcma and brcmsmac. Does
it still lockup your machine?

-- 
Rafał
--
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