Public bug reported: This bug is to confirm that the Marvell 88SE9128 chipset bug presents under AMD's IOMMU implementation as well as Intel's VT-D. I've tested this with a Syba SD-PEX40035 and a Highpoint Rocket 620 (both use the same Marvell chipset) PCIe card, under both the stock Quantal kernel and the 3.7.0rc8 kernel from the mainline kernel PPA. Both cards work fine as long as I disable the IOMMU in the bios. It sounds like the preferred solution is PCI quirk, and some of the infrastructure is in place, but it hasn't been implemented yet. Symptoms consist of a much longer boot process, and failure to identify any devices attached to the Marvell controllers.
My system consists of an AMD A10-5800K in a Gigabyte GA-F2A85X-D3H motherboard, 5x Seagate 1.5 TB drives in software Raid5, 4x Seagate 3 TB drives in a native ZFS raidz (not yet used), a Sata BD-ROM, and a 500 GB sata drive (not yet used). https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1005226 http://lists.linuxfoundation.org/pipermail/iommu/2012-September/004551.html https://bugzilla.kernel.org/show_bug.cgi?id=42679 https://bugzilla.redhat.com/show_bug.cgi?id=757166 pci 02:00.1 is the IDE endpoint, but there are no PATA devices in the system, which matches the behavior of the other chips (emitting DMA operations from the wrong PCI endpoint, which the IOMMU rejects.) ########################################################## 02:00.1 IDE interface: Marvell Technology Group Ltd. 88SE9128 IDE Controller (rev 11) (prog-if 8f [Master SecP SecO PriP PriO]) Subsystem: Marvell Technology Group Ltd. 88SE9128 IDE Controller Flags: bus master, fast devsel, latency 0, IRQ 18 I/O ports at d040 [size=8] I/O ports at d030 [size=4] I/O ports at d020 [size=8] I/O ports at d010 [size=4] I/O ports at d000 [size=16] Memory at fea20000 (32-bit, non-prefetchable) [size=16] Expansion ROM at fea00000 [disabled] [size=64K] Capabilities: [40] Power Management version 3 Capabilities: [50] MSI: Enable- Count=1/1 Maskable- 64bit- Capabilities: [70] Express Legacy Endpoint, MSI 00 Capabilities: [100] Advanced Error Reporting Kernel driver in use: pata_marvell Kernel modules: pata_marvell ########################################################## [ 1.225703] AMD-Vi: Event logged [IO_PAGE_FAULT device=02:00.1 domain=0x001d address=0x00000000001003c0 flags=0x0000] <snip> [ 6.207818] ata16.00: qc timeout (cmd 0xa1) [ 6.215674] ata9.00: qc timeout (cmd 0xa1) [ 6.215681] ata9.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 6.223640] ata24.00: qc timeout (cmd 0xa1) [ 6.231606] ata17.00: qc timeout (cmd 0xec) [ 6.231627] ata24.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 6.239574] ata17.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 6.343207] ata18.00: qc timeout (cmd 0xec) [ 6.351204] ata18.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 6.534578] ata9: SATA link up 1.5 Gbps (SStatus 113 SControl 300) [ 6.558489] ata24: SATA link up 1.5 Gbps (SStatus 113 SControl 300) [ 6.709935] ata16.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 7.531002] ata16: SATA link up 1.5 Gbps (SStatus 113 SControl 300) [ 11.588451] ata17: link is slow to respond, please be patient (ready=0) [ 11.700054] ata18: link is slow to respond, please be patient (ready=0) [ 16.219859] ata17: COMRESET failed (errno=-16) [ 16.331454] ata18: COMRESET failed (errno=-16) [ 16.498853] ata9.00: qc timeout (cmd 0xa1) [ 16.498865] ata9.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 16.498870] ata9: limiting SATA link speed to 1.5 Gbps [ 16.522792] ata24.00: qc timeout (cmd 0xa1) [ 16.522804] ata24.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 16.522807] ata24: limiting SATA link speed to 1.5 Gbps [ 16.817725] ata9: SATA link up 1.5 Gbps (SStatus 113 SControl 310) [ 16.849613] ata24: SATA link up 1.5 Gbps (SStatus 113 SControl 310) [ 17.495260] ata16.00: qc timeout (cmd 0xa1) [ 17.997499] ata16.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 17.997504] ata16: limiting SATA link speed to 1.5 Gbps [ 18.818556] ata16: SATA link up 1.5 Gbps (SStatus 113 SControl 310) [ 21.568700] ata17: link is slow to respond, please be patient (ready=0) [ 21.680300] ata18: link is slow to respond, please be patient (ready=0) [ 26.200104] ata17: COMRESET failed (errno=-16) [ 26.311693] ata18: COMRESET failed (errno=-16) [ 31.548931] ata17: link is slow to respond, please be patient (ready=0) [ 31.660534] ata18: link is slow to respond, please be patient (ready=0) [ 46.710611] ata9.00: qc timeout (cmd 0xa1) [ 46.710618] ata9.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 46.742498] ata24.00: qc timeout (cmd 0xa1) [ 46.750464] ata24.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 47.029478] ata9: SATA link up 1.5 Gbps (SStatus 113 SControl 310) [ 47.077301] ata24: SATA link up 1.5 Gbps (SStatus 113 SControl 310) [ 48.711438] ata16.00: qc timeout (cmd 0xa1) [ 49.213639] ata16.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 50.034704] ata16: SATA link up 1.5 Gbps (SStatus 113 SControl 310) [ 61.122962] ata17: COMRESET failed (errno=-16) [ 61.122971] ata17: limiting SATA link speed to 1.5 Gbps [ 61.234560] ata18: COMRESET failed (errno=-16) [ 61.234569] ata18: limiting SATA link speed to 3.0 Gbps [ 66.137000] ata17: COMRESET failed (errno=-16) [ 66.137009] ata17: reset failed, giving up [ 66.248595] ata18: COMRESET failed (errno=-16) [ 66.248603] ata18: reset failed, giving up <snip> -R C ** Affects: linux (Ubuntu) Importance: Undecided Status: New ** Tags: iommu kernel-bug marvell sata ** Attachment added: "3.5 kernel dmesg with iommu enabled" https://bugs.launchpad.net/bugs/1089768/+attachment/3458917/+files/dmesg.3.5-iommu-on.log -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1089768 Title: Marvell Sata controller and AMD-Vi problems To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1089768/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
