On Wed, Jul 05, 2006 at 08:49:27AM -0700, Auke Kok wrote: > Zhang, Yanmin wrote: > >On Fri, 2006-06-30 at 00:26, Linas Vepstas wrote: > >>Adds PCI Error recovery callbacks to the Intel 10-gigabit ethernet > >>ixgb device driver. Lightly tested, works. > > > >Both pci_disable_device and ixgb_down would access the device. It doesn't > >follow Documentation/pci-error-recovery.txt that error_detected shouldn't > >do > >any access to the device. > > Moreover, it was Linas who wrote this documentation in the first place :)
On the pSeries, its harmless to try to do i/o; the i/o will e blocked. > Linas, have you tried moving the e1000_down() call into the _reset part? I > suspect that the e1000_reset() in there however may already be sufficient. I wanted to perform all of the "down" type functions BEFORE the reset. The idea is to get the device driver and the various parts of the Linux kernel into a state that would be consisten with a reset. I don't want to do these functions after the reset, since, at this point, the card is a "clean slate"; it has the PCI bars set, but nothing else. Doing random i/o to it at this point could confuse the card; instead, one wants to bring the card up using the usual bringup sequence. For example, I tipped over one rather confusing bug: new code in the -mm tree blocks a pci_enable_device() if it thinks the card is already enabled (even if its not). Doing I/O to a card that is not enabled will cause either a target abort or a master abort. Thus, I found I had to call pci_disable_device(); and it seemed that the best time to do this would be before the reset, not afterwords. However, I did not play at length with other possibilities. I recently lost access to my ixgb cards, and so can't do more testing just right now. --linas - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html