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]

Message-ID: <CA+55aFxZ7H=x2-rgQgRE9b=ky-h0WM1x-48Zs8a0dVt0NvkB_g@mail.gmail.com>
Date:	Thu, 12 Jan 2012 16:13:30 -0800
From:	Linus Torvalds <torvalds@...ux-foundation.org>
To:	Rafał Miłecki <zajec5@...il.com>
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..

2012/1/12 Rafał Miłecki <zajec5@...il.com>:
>
> Have you tried booting with bcma & brcmsmac blacklisted? Does
> suspend&resume work then?
>
> Have you tried blacklisting just brcmsmac (letting bcma load)? Does
> s&r work then?

If I unload brcmsmac, I can suspend/resume. Once. It can't suspend a
second time.

I did see some message flash about "does not have a release()
function", but don't know if that was bcma or something else.

I do notice that both the bcma and suspend/resume seems quite broken.
It's using the legacy suspend/resume stuff and does the PCI resume on
its own (with no matching suspend!). That *really* isn't a good idea
these days.

The way to do it these days is to have a struct dev_pm_ops embedded in
the struct pci_driver (".driver.pm"), and let the PCI layer handle all
the generic PCI suspend/resume  details - you only handle the
device-specific ones (ie in this case suspending/resuming the bcma bus
itself).

The generic PCI layer will do all the PCI stuff correctly, including
all the nasty races with shared interrupts etc. In a way that no
driver ever got it right. And it simplifies the driver too.

And the brcms driver does suspend/resume *completely* wrong, and seems
to actually re-suspend and re-resume the PCI device.

I'm surprised it has ever worked for anybody. It certainly doesn't work for me.

                      Linus
--
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