From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.8) Gecko/20050512 Red Hat/1.0.4-1.4.1 Firefox/1.0.4 Description of problem: Emulex has recently found several serious bugs in the the lpfc driver version we have in U2. The most critical is a fix for a race between eh_abort handler and the scsi_done up-call: "This allowed premature scsi completions, which in our tests have lead to data corruption. As such, we'd like very much to have this updated driver in RHEL4u2." The fix is to hold the host_lock in driver's done routine. There are several other serious bug fixes since late May when they sent us the driver we currently have in U2. These include various error handling and stability issues, including the FAN issue that I just posted for Taroon. The complete changelog is attached. Unfortunately, there are also a bunch of less important changes in their current revision, including a pile of annoying copyright changes, etc. Version-Release number of selected component (if applicable): kernel-2.6.9-11 How reproducible: Always Steps to Reproduce: 1.boot the kernel with emulax driver 2.run some disk related tests 3. Additional info: We need to make a difficult decision about whether to take their latest revision as-is, so we stay in synch with the testing that Emulex and our shared OEMs are doing, or whether we should minimize change in the U2 stream, by forcing Emulex to give is a custom driver that only has very critical bug fixes. My opinion is that we are better off staying in sync with the rest of the world, because their FC testing is much more thorough than ours and all of our beta testers combined. I have reviewed this patch in detail, and I don't see anything that adds risk w.r.t. interaction with the Linux SCSI subsystem. I have also built it and done some regression testing. I have been told that EMC, IBM and others are already testing this driver revision as well.
*** This bug has been marked as a duplicate of 158789 ***