netdev - Re: [PATCH 10/11][TG3]: Reduce spurious interrupts.

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, 04 May 2007 18:45:58 -0700
From:	"Michael Chan" <mchan@...adcom.com>
To:	"Jeff Garzik" <jeff@...zik.org>
cc:	"David Miller" <davem@...emloft.net>,
	"netdev" <netdev@...r.kernel.org>
Subject: Re: [PATCH 10/11][TG3]: Reduce spurious interrupts.

On Fri, 2007-05-04 at 20:42 -0400, Jeff Garzik wrote:
> Michael Chan wrote:
> > [TG3]: Reduce spurious interrupts.
> > 
> > Spurious interrupts are often encountered especially on systems
> > using the 8259 PIC mode.  This is because the I/O write to deassert
> > the interrupt is posted and won't get to the chip immediately.  As
> > a result, the IRQ may remain asserted after the IRQ handler exits,
> > causing spurious interrupts.
> > 
> > An unconditional read to flush the I/O write to force the IRQ to de-
> > assert immediately is not desirable because it impacts performance in
> > the fast path.  So we only do this after we have some indications of
> > spurious interrupts.
> > 
> > Signed-off-by: Michael Chan <mchan@...adcom.com>
> 
> hmmmm, this is a bit questionable that it needs to be here.
> 
> I think it's just a fact of life that it is important to flush certain 
> writes...
> 
> I'm not sure a driver needs to be adding code to avoid the obvious 
> solution.  It would be annoying if all drivers had code to do this.

We had a discussion about 2 years ago and David decided to remove the
I/O read to improve performance.  Since then a small number of users
have been complaining about spurious interrupts.  We can add back the
unconditional read or do this detection thing which I agree is somewhat
annoying.  David, what do you think?

-
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