Re: More envelope info on "Socket destroy delayed"

Scott Jennings (smj@smudge.oro.net)
3 Jun 1996 07:46:34 GMT

Messages sorted by: [ date ][ thread ][ subject ][ author ]
Next message: Ingo Molnar: "Re: Ext2 speedup (No atime update), concerns"
Previous message: Steffen Grunewald: "Re: Going from 32M to 48M, Linux crawls!"

In every case that I've been able to watch closely enough, the
onslaught of "Socket destroy delayed" messages are preceeded by a
single "double lock on device queue" message.

This has been seen in 1.99.9.

Have just started a few 1.99.10 kernels... have not seen it yet,
but am *still* getting massive "UDP: bad checksum" errors, which
I suspect to be related. (kernels which do not produce the
"delayed" messages do not produce the "checksum" messages either)

-smj

Next message: Ingo Molnar: "Re: Ext2 speedup (No atime update), concerns"
Previous message: Steffen Grunewald: "Re: Going from 32M to 48M, Linux crawls!"