From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 Description of problem: During installation on a system with a 3Ware 6410 or 6800 IDE Raid card, a crash occurs prior to formatting the file systems: <4> 3w-xxxx:scsi 0:unit #0: Command (df36f400) timed out Unable to handle kernel NULL pointer dereference at virtual address.... ... scsi_eh_0: ... The ...'s represent a stak trace Redhat 8 installs just fine, bu 9 has this crash. Version-Release number of selected component (if applicable): 3w-xxxx driver / scsi_eh_0 How reproducible: Always Steps to Reproduce: 1.Install Redhat 9 on a system with a 3ware Escalade 6410 or 6800 2.When ready to format... kaboom! 3. Actual Results: <4> 3w-xxxx:scsi 0:unit #0: Command (df36f400) timed out Unable to handle kernel NULL pointer dereference at virtual address.... ... scsi_eh_0: ... Expected Results: The installation should have proceeded with bad block checks and formatting of partitions Additional info:
Is there any progress with resolving this issue? I'm willing to test this and gather more info for you. Please update me as to the status of your investigation. Thanks, Ken
*** Bug 88993 has been marked as a duplicate of this bug. ***
What is going on to resolve this issue? I have not heard a status report on this issue for some time. Is this being worked on? Or is my 3ware card now garbage as far as Linux goes...
I also have this problem too. What Red Hat needs to do is provide a drive disk that overrides the default 3ware driver shipped with RH 9. I spoke with 3Ware on this subject, and their response is that Red Hat shipped a bad driver with RH9. If that is the case, there should be either an anaconda boot disk errata to download, or a driver disk so we can start the installer with "linux dd". A lot of people use 3Ware cards, its not something that can be ignored and have happy customers.
Also present in kernel-2.4.20-xxx for RH 7.3 I have come across this bug not only in RH 9 upgrade/install, but also in RH 7.3, trying to upgrade from kernel-2.4.18-27.7.x to kernel-2.4.20-xxxx (any of the kernel-2.4.20 upgrades) Version of 3w-xxxx.c driver that causes this problem is 1.02.00.032. This is the version of the driver that is present in RH 9 kernels as well as in RH 7.3 kernel-2.4.20 upgrades. Last version of 3w-xxxx.c driver that I can confirm working is 1.02.00.025. This is the version in RH 7.3 kernel-2.4.18-27.7.x . There is a comment in 3ware web site knowledge base. They don't seem to allow deep linking but here is the Question ID: Q10364 . Essentially they say to use a different driver.
RH 9 install workaround Hello, I haven't tried this on the machine with the 6800 card yet, but I have tried it on a 7500 card machine and at least the install/modify procedure works. I took the 3w-xxxx.c and 3w-xxxx.h from kernel-2.4.18-27.7.x . I replaced the files in kernel-2.4.20-8 with these and re-compiled the source, creating all the different flavors of kernel binary RPMS. I also took the compiled 3w-xxxx.o from kernel-BOOT that I compiled from modified source. I then took drvblock.img, opened it up, unpacked modules.cgz, replaced 3w-xxxx.o with the new 3w-xxxx.o and re-packed everything. I created an FTP install tree for RH 9, and replaced the default kernel RPMs with the ones that I compiled. I created floppies of bootdisk.img, drvnet.img with the stock images and drvblock.img with the modified image file. I then did "dd" during install and used the drvblock.img diskette.
Thanks for the bug report. However, Red Hat no longer maintains this version of the product. Please upgrade to the latest version and open a new bug if the problem persists. The Fedora Legacy project (http://fedoralegacy.org/) maintains some older releases, and if you believe this bug is interesting to them, please report the problem in the bug tracker at: http://bugzilla.fedora.us/