On Thu, May 28, 2009 at 10:00, Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote: >> Product: SCSI Drivers >> Version: 2.5 >> Kernel Version: 2.6.27.21 > > Is it reproducible? If so, is there any change that it can be retested > under a 2.6.29-based kernel? This driver version: MPT_LINUX_VERSION_COMMON "4.00.43.00suse" seems not in the upstream kernel: MPT_LINUX_VERSION_COMMON "3.04.07" The failing code is only in the new driver. This bug should move to the Novell bugzilla, so the people who added the driver to the SUSE kernel can check with LSI directly. > OK, I assume that the scsi driver did something bad to sysfs and > that sysfs then fell on its face. > > Really, given the frequency and imaginativeness with which drivers > abuse sysfs, the driver-core should be more robust. > > Kay, would you have time to plunk through this and see if we can > strengthen the sysfs code a bit so it doesn't crash? Might be a missing lock/wrong refcounting issue in the caller, where two threads try to remove the same thing. With Eric's upcoming changes to the cleanup logic in sysfs, this should be handled better. Thanks, Kay -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
