From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040115 Description of problem: I have a Soyo SY-P4I875P motherboard that contains an Intel 875P chipset. The controllers on it are the ICH5 (IDE/SATA), HPT372 (IDE RAID) and Silicon Image 3112A (SATA RAID). The HPT372 isn't used, since I have no drives connected to it. Connected are: ICH5 SATA: 2 x IDE hard drives with SATA adapters ICH5 IDE: 2 x DVD-ROM (actually, 1 DVD-ROM, 1 DVD-R) - connected on secondary channel Si3112A: 2 x IDE hard drives with SATA adapters The BIOS is setup in Combined Mode, so that the SATA drives on the ICH5 become primary master/slave, and the DVD-ROM/DVD-R become secondary master/slave. When the Si3112A SATA chipset is disabled on the motherboard (through the BIOS), everything is fine, and the output from dmesg is in the attachment okay.txt. When the Si3112A chipset is enabled on the motherboard, there are Assertion failures on the DVD-ROM and DVD-R drives (crash.txt): ata2: PATA max UDMA/33 cmd 0x170 ctl 0x376 bmdma 0xF008 irq 15 ata2: dev 0 cfg 49:0f00 82:0218 83:4000 84:4000 85:0218 86:0000 87:4000 88:101f Assertion failed! dev->class == ATA_DEV_ATA,drivers/scsi/libata-core.c,ata_dev_p arse_strings,line=691 ata2: dev 0 ATAPI, max UDMA/66 ata2: dev 1 cfg 49:0f00 82:0218 83:4000 84:4000 85:0218 86:0000 87:4000 88:0407 Assertion failed! dev->class == ATA_DEV_ATA,drivers/scsi/libata-core.c,ata_dev_p arse_strings,line=691 ata2: dev 1 ATAPI, max UDMA/33 ata2: dev 0 configured for UDMA/33 ata2: dev 1 configured for UDMA/33 Note that crash.txt is a bit incomplete, because of the multiple Si3112A error messages. I believe that this is unrelated since I have seen it on the kernel mailing list archives with real SATA drives (just in case it is the SATA adapters causing the problem). This setup also fails under Fedora Core 1 (I believe with the same issue, I didn't do too much testing), but works fine under Knoppix 3.3 (using the 2.4.22 kernel). The error messages from the Si3112A driver also appear in Knoppix, but I do not have any problems with the DVD-ROM/DVD-R drives. Version-Release number of selected component (if applicable): kernel-2.6.2-1.85 and kernel-2.6.1-1.65 How reproducible: Always Steps to Reproduce: 1. In BIOS of motherboard, enable SATA on the ICH5, and set it to "combined" mode, when the SATA drives being primary and the IDE drives being secondary 2. Enable Si3112A SATA controller in motherboard BIOS 3. Boot machine. Fails every time, DVD-ROM is not available, but drives on Si3112A are. 4. Go into motherboard BIOS, disable Si3112A SATA controller. 5. Boot machine. Works fine (although I can't use the hard drives connected to that controller) Expected Results: Knoppix 3.3 works with this configuration, so I would hope that Fedora can as well (I prefer Fedora to Knoppix). Additional info: Attached are two dumps from dmesg, "okay.txt" and "crash.txt". Okay.txt is with the Si3112A disabled, crash.txt is when it is enabled, causing libata to have problems. Stock Fedora Core 2 Test 1 is affected, as is an upgrade to kernel-2.6.2-1.85. From crash.txt: ata2: PATA max UDMA/33 cmd 0x170 ctl 0x376 bmdma 0xF008 irq 15 ata2: dev 0 cfg 49:0f00 82:0218 83:4000 84:4000 85:0218 86:0000 87:4000 88:101f Assertion failed! dev->class == ATA_DEV_ATA,drivers/scsi/libata-core.c,ata_dev_p arse_strings,line=691 ata2: dev 0 ATAPI, max UDMA/66 ata2: dev 1 cfg 49:0f00 82:0218 83:4000 84:4000 85:0218 86:0000 87:4000 88:0407 Assertion failed! dev->class == ATA_DEV_ATA,drivers/scsi/libata-core.c,ata_dev_p arse_strings,line=691 ata2: dev 1 ATAPI, max UDMA/33 ata2: dev 0 configured for UDMA/33 ata2: dev 1 configured for UDMA/33
Created attachment 97781 [details] Output of dmesg when failed assertion occurs (crash.txt)
Created attachment 97782 [details] Output of dmesg when Si3112A is disabled (okay.txt)
Does this happen with newer kernels?
Yes. Although I'm not completely sure about the crash under Fedora Core 2 Test 3, my CD-ROMs are inaccessible during the installation, so there's still a problem. Using a vanilla 2.6.5 (from kernel.org) on my Slackware partition reveals that there's a problem with the vanilla kernel as well - I'm seeing the libata exception. For now, this is not much of an issue, since I'm using a Promise SATA card, which seems to work much better than the Sil3112A. -- Joe
Fedora Core 2 has now reached end of life, and no further updates will be provided by Red Hat. The Fedora legacy project will be producing further kernel updates for security problems only. If this bug has not been fixed in the latest Fedora Core 2 update kernel, please try to reproduce it under Fedora Core 3, and reopen if necessary, changing the product version accordingly. Thank you.
It looks like some funky hardware limitations: "When using a non Windows XP system, only 4 harddisks/IDE devices are supported at the same time on IDE 1,2 and SATA 1,2. Windows XP systems can support all 6 devices on these ports (2 SATA, 4 parallel ATA/ATAPI devices)." http://www.soyogroup.com/kb/kbdesc.php?id=160
*** Bug 156392 has been marked as a duplicate of this bug. ***