netdev
[Top] [All Lists]

Re: [patch 4/10] s390: network driver.

To: jamal <hadi@xxxxxxxxxx>
Subject: Re: [patch 4/10] s390: network driver.
From: Paul Jakma <paul@xxxxxxxx>
Date: Sun, 19 Dec 2004 23:54:49 +0000 (UTC)
Cc: Thomas Spatzier <thomas.spatzier@xxxxxxxxxx>, "David S. Miller" <davem@xxxxxxxxxxxxx>, Hasso Tepper <hasso@xxxxxxxxx>, Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>, jgarzik@xxxxxxxxx, netdev@xxxxxxxxxxx
In-reply-to: <1103484552.1046.155.camel@xxxxxxxxxxxxxxxx>
Mail-followup-to: paul@xxxxxxxxxxxxxxxxxx
References: <OF28701C56.81E1D26E-ONC1256F6B.00513EDD-C1256F6B.0052AF84@xxxxxxxxxx> <1103484552.1046.155.camel@xxxxxxxxxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
On Sun, 19 Dec 2004, jamal wrote:

This is the strange part. Anyone still willing to provide a sample
program that hangs?

I can provide instructions, or you can wait a wee bit - I didnt have any e1000 hardware to test with (e1000 being one of the drivers which has this behavious, AFAIK/TTBOMK) - but a computer with an e1000 arrived Friday. So, give me a bit and I'll try come up with a test case. A simple UDP send should be enough.

Still, I think we need to resolve the original problem.
And I have a feeling we wont be seeing any program which
reproduces it ;->

Just send a UDP packet - AIUI when link goes down the application can (or at least a window exists) where the application can send a packet down the file descriptor without error, but it will be queued rather than sent and the fd eventually blocks/returns EAGAIN.

That window is the problem.

I'll test tomorrow.

cheers,
jamal

regards,
--
Paul Jakma      paul@xxxxxxxx   paul@xxxxxxxxx  Key ID: 64A2FF6A
Fortune:
You want to know why I kept getting promoted?  Because my mouth knows more
than my brain.
        -- W.G.

<Prev in Thread] Current Thread [Next in Thread>