From Bugzilla Helper: User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.5) Gecko/20041107 Firefox/1.0 Description of problem: I have a system with an older BIOS that has a 32G limit. Using the IDE size-limit jumpers and IDE "stroke" (which was a config option and is now a kernel parameter) is supposed to work around this issue. The problem is that the kernel is using the LBA48 SET_MAX_EXT command, which causes the attached Maxtor 4R120L4 drive to start spewing errors. The good news is that the LBA28 SET_MAX does work, and can be used to work around the issue by running the old "setmax.c" program at the beginning of /etc/rc.sysinit, with -d0, to reset the full disk capacity. It seems that this drive does support LBA48 I/O, but not the LBA48 SET_MAX_EXT. I suspect not many live BIOSes cannot handle LBA48, so it is likely a rare problem. (This system is an HP NetServer -- not all that old, but the BIOS assumes a SCSI-based system.) Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Install a Maxtor 4R120L4 with size-limit jumper in place 2. Boot with hdx=stroke Actual Results: Hundreds of drive error messages occur after the command is sent. Expected Results: Drive capacity goes from 32G to 120G Additional info: Probably not worth a lot of effort, but at least pass this info on to an ide-disk maintainer. Otherwise, any other people with this issue can use the setmax.c work-around.
Makes sense. We had a similar problem with some other Maxtors that report slightly odd but valid feature combinations. Will take a look.
It occured to me that any disk that can handle SET_MAX but not SET_MAX_EXT has to be under the LBA28 size limit, so one way to fix this is a kernel parameter to force LBA28. There is a no_lba48 flag, but it seems that right now only the low-level IDE drivers change this.
The change I've dropped into -ac for testing tries to use LBA48 disk size read, then if that fails LBA28 disk size read. That should be fine anyway. If the new size is within LBA28 range it issues the LBA28 command regardless of the drive claims.
Joe, what IDE controller is this disk connected to?
Should be fixed in the FC4 kernel, please retest (Bartlomiej - if this is fixed in FC4 its a bug in the -ac patches that isnt in base so don't worry about it)
An update has been released for Fedora Core 3 (kernel-2.6.12-1.1372_FC3) which may contain a fix for your problem. Please update to this new kernel, and report whether or not it fixes your problem. If you have updated to Fedora Core 4 since this bug was opened, and the problem still occurs with the latest updates for that release, please change the version field of this bug to 'fc4'. Thank you.