Most recent kernel where this bug did *NOT* occur: 2.6.20 Distribution: fedora core 6, new kernel Hardware Environment: HP dv9000z laptop (AMD 64x2 SMP, nforce chipset, 2 GB ram) Software Environment: kernel 2.6.21-rc1 built using FC6 .config options. Problem Description: after boot, eth0 stops working Steps to reproduce: build kernel, install, boot. driver loads, but after about 7 packets, stops working - no packets getting through. back out just the forcedeth.c diffs, kernel works as well as 2.6.20 (better, actually, since sata_nv seems to be less buggy), in particular eth0 works.
Did you try the following patch set? http://marc.theaimsgroup.com/?l=linux-netdev&m=117199770027017&w=2
Ayaz - the patch you mention here does fix this problem for me. Thanks!
The patch that fixes this bug is now in Linus' tree.