Hide Forgot
Description of problem: Machine is Lenovo D20 with Marvell MV64460 SAS controller and two Seagate Cheetah 15k.7 SAS disks. Disks are detected properly by the controller itself, but not all disks are detected by the mvsas driver. Testing has shown that current Seagate Cheetah 15K.7 Series (300GB/450GB/600 GB) are not detected in most of the cases. Other disks such as older Seagate disks or Hitachi drives are detected in all tests. During installation, "/tmp/syslog" shows: <7>libata version 3.00 loaded. <6>ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 169 <7>PCI: Setting latency timer of device 0000:01:00.0 to 64 <7>mvsas 0000:01:00.0: No sig fis <7>mvsas 0000:01:00.0: No sig fis <6>scsi1 : mvsas <4>ERROR: Unidentified device type 0 <6>mvsas 0000:01:00.0: version 0.5.4 <6>mvsas 0000:01:00.0: 8 phys, addr 5005043011ab0000 <7>ahci 0000:00:1f.2: version 3.0 <6>ACPI: PCI Interrupt 0000:00:1f.2[B] -> GSI 17 (level, low) -> IRQ 177 <6>ahci 0000:00:1f.2: AHCI 0001.0200 32 slots 6 ports 3 Gbps 0x9 impl SATA mode <6>ahci 0000:00:1f.2: flags: 64bit ncq sntf led clo pmp pio slum part <7>PCI: Setting latency timer of device 0000:00:1f.2 to 64 <6>scsi2 : ahci <6>scsi3 : ahci <6>scsi4 : ahci <6>scsi5 : ahci <6>scsi6 : ahci <6>scsi7 : ahci Version-Release number of selected component (if applicable): Reproduced with RHEL5.4, RHEL5.5 and RHEL5.6beta (mvsas driver 0.5.4) Not reproduced with RHEL 6.0 (mvsas-Driver 0.8.2) Additional info: System is certified for RHEL5.4 here: https://hardware.redhat.com/show.cgi?id=561188 Support for this Controller was introduced with RHEL5.4 in: https://bugzilla.redhat.com/show_bug.cgi?id=485126
Hello Rob & other old Tru64 & HPUX mates! I'm supporting computational biochemistry at Pfizer research in Cambridge; our internal production images are RHEL5.2 and RHEL5.5. I've been wrestling with this same problem with Lenovo D20 and C20 systems, both using the Marvell 88SE63xx SATA RAID controller. Confirmed it works perfectly using RHEL6.0. I was hoping it would be possible to get patches/drivers to support these configurations. Here is additional output that may be of use, captured in an #sh session during the %pre phase of the kickstart of an isolinux boot. The install device itself is seen here, but never discovered when it comes time to download an image. I thought I could install a driver at this point, and then again into the running kernel once the install's done. Lenovo support includes a device driver floppy image which I tried during the install, to provide this bootstrap support. It definitely loaded, but either it didn't work for my case, or it got overwritten by another ahci or mvsas. The patch name name from Lenovo is "ldsas05ws14", available from url: http://www-307.ibm.com/pc/support/site.wss/MIGR-72070.html ## Captured during isolinux... kickstart... %pre phase: # uname -a Linux localhost.localdomain 2.6.18-92.el5 #1 SMP Tue Apr 29 13:16:15 EDT 2008 x86_64 unknown # lspci | egrep -i "raid|sata|ahci" 00:1f.2 SATA controller: Intel Corporation ICH10 6 port SATA AHCI Controller 01:00.0 RAID bus controller: Marvell Technology Group Ltd. MV64460/64461/64462 System Controller, Revision B (rev 01) # dmesg | egrep -i "raid|sata|ahci" <7>ahci 0000:00:1f.2: version 3.0 <4>ahci 0000:00:1f.2: nr_ports (6) and implemented port map (0x9) don't match, using nr_ports <4>ahci 0000:00:1f.2: forcing PORTS_IMPL to 0x3f <6>ahci 0000:00:1f.2: AHCI 0001.0200 32 slots 6 ports 3 Gbps 0x3f impl SATA mode <6>ahci 0000:00:1f.2: flags: 64bit ncq sntf led clo pmp pio slum part <6>scsi2 : ahci <6>scsi3 : ahci <6>scsi4 : ahci <6>scsi5 : ahci <6>scsi6 : ahci <6>scsi7 : ahci <6>ata1: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606100 irq 82 <6>ata2: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606180 irq 82 <6>ata3: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606200 irq 82 <6>ata4: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606280 irq 82 <6>ata5: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606300 irq 82 <6>ata6: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606380 irq 82 <6>ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 300) <6>ata2: SATA link down (SStatus 0 SControl 0) <6>ata3: SATA link down (SStatus 0 SControl 0) <6>ata4: SATA link down (SStatus 0 SControl 300) <6>ata5: SATA link down (SStatus 0 SControl 0) <6>ata6: SATA link down (SStatus 0 SControl 0) <6>md: raid0 personality registered for level 0 <6>md: raid1 personality registered for level 1 <6>raid5: automatically using best checksumming function: generic_sse <4>raid5: using function: generic_sse (9728.000 MB/sec) <4>raid6: int64x1 2433 MB/s <4>raid6: int64x2 2808 MB/s <4>raid6: int64x4 2238 MB/s <4>raid6: int64x8 1976 MB/s <4>raid6: sse2x1 5468 MB/s <4>raid6: sse2x2 6609 MB/s <4>raid6: sse2x4 7546 MB/s <4>raid6: using algorithm sse2x4 (7546 MB/s) <6>md: raid6 personality registered for level 6 <6>md: raid5 personality registered for level 5 <6>md: raid4 personality registered for level 4 Thanks! -josh
(In reply to comment #4) > Hello Rob & other old Tru64 & HPUX mates! Hi Josh... > > I'm supporting computational biochemistry at Pfizer research in Cambridge; our > internal production images are RHEL5.2 and RHEL5.5. The mvsas driver was include for the first time in RHEL5.4. It will not be backported to RHEL5.2. Also, a patch to enable new disks will not be backported to RHEL5.5 at this point. > > I've been wrestling with this same problem with Lenovo D20 and C20 systems, > both using the Marvell 88SE63xx SATA RAID controller. Confirmed it works > perfectly using RHEL6.0. > > I was hoping it would be possible to get patches/drivers to support these > configurations. > > Here is additional output that may be of use, captured in an #sh session during > the %pre phase of the kickstart of an isolinux boot. The install device itself > is seen here, but never discovered when it comes time to download an image. I > thought I could install a driver at this point, and then again into the running > kernel once the install's done. > > Lenovo support includes a device driver floppy image which I tried during the > install, to provide this bootstrap support. It definitely loaded, but either it > didn't work for my case, or it got overwritten by another ahci or mvsas. The > patch name name from Lenovo is "ldsas05ws14", available from url: > > http://www-307.ibm.com/pc/support/site.wss/MIGR-72070.html > > > ## Captured during isolinux... kickstart... %pre phase: > > # uname -a > Linux localhost.localdomain 2.6.18-92.el5 #1 SMP Tue Apr 29 13:16:15 EDT 2008 > x86_64 unknown > > # lspci | egrep -i "raid|sata|ahci" > 00:1f.2 SATA controller: Intel Corporation ICH10 6 port SATA AHCI Controller > 01:00.0 RAID bus controller: Marvell Technology Group Ltd. MV64460/64461/64462 > System Controller, Revision B (rev 01) > > # dmesg | egrep -i "raid|sata|ahci" > <7>ahci 0000:00:1f.2: version 3.0 > <4>ahci 0000:00:1f.2: nr_ports (6) and implemented port map (0x9) don't match, > using nr_ports > <4>ahci 0000:00:1f.2: forcing PORTS_IMPL to 0x3f > <6>ahci 0000:00:1f.2: AHCI 0001.0200 32 slots 6 ports 3 Gbps 0x3f impl SATA > mode > <6>ahci 0000:00:1f.2: flags: 64bit ncq sntf led clo pmp pio slum part > <6>scsi2 : ahci > <6>scsi3 : ahci > <6>scsi4 : ahci > <6>scsi5 : ahci > <6>scsi6 : ahci > <6>scsi7 : ahci > <6>ata1: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606100 irq 82 > <6>ata2: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606180 irq 82 > <6>ata3: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606200 irq 82 > <6>ata4: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606280 irq 82 > <6>ata5: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606300 irq 82 > <6>ata6: SATA max UDMA/133 abar m2048@0xf4606000 port 0xf4606380 irq 82 > <6>ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 300) > <6>ata2: SATA link down (SStatus 0 SControl 0) > <6>ata3: SATA link down (SStatus 0 SControl 0) > <6>ata4: SATA link down (SStatus 0 SControl 300) > <6>ata5: SATA link down (SStatus 0 SControl 0) > <6>ata6: SATA link down (SStatus 0 SControl 0) > <6>md: raid0 personality registered for level 0 > <6>md: raid1 personality registered for level 1 > <6>raid5: automatically using best checksumming function: generic_sse > <4>raid5: using function: generic_sse (9728.000 MB/sec) > <4>raid6: int64x1 2433 MB/s > <4>raid6: int64x2 2808 MB/s > <4>raid6: int64x4 2238 MB/s > <4>raid6: int64x8 1976 MB/s > <4>raid6: sse2x1 5468 MB/s > <4>raid6: sse2x2 6609 MB/s > <4>raid6: sse2x4 7546 MB/s > <4>raid6: using algorithm sse2x4 (7546 MB/s) > <6>md: raid6 personality registered for level 6 > <6>md: raid5 personality registered for level 5 > <6>md: raid4 personality registered for level 4 > > > Thanks! -josh Unfortunately Marvell seems to have dropped support of the mvsas controller driver in linux. This leaves a situation where qualification of any new functionality is likely non-existant. Additionally any changes to the driver introduce the risk of regression due to lack of thorough ongoing qualification by Marvell. It is theoretically possible for the code that enables other disk drives to be distilled out of the changes that occurred in the libsas/mvsas driver between rhel5 and rhel6, and for this change to be backported to rhel5. If you or anyone else are will to take on the task of doing the distillation, creating a rhel5 version of the patch, and testing it, I would consider taking the patch depending on how intrusive it is. This would only be available in the next rhel5.x release depending on schedule. Are you up for that? If not, the likely easiest solution is to go with a seperate disk controller and configure your disks using that. Rob
Rob, I don't think I'm "up for" distilling out an rhel5 patch, but thanks for the "offer", so to speak, and for the additional background details. As I referenced, Lenovo is providing a driver (ldsas05ws14) for 5.3 and 5.4; is that an mvsas for other (older) Marvell controllers, or for other older disks? Do you know what support that's supposed to provide? Thanks a lot. -josh
(In reply to comment #6) > Rob, I don't think I'm "up for" distilling out an rhel5 patch, but thanks for > the "offer", so to speak, and for the additional background details. > > As I referenced, Lenovo is providing a driver (ldsas05ws14) for 5.3 and 5.4; is > that an mvsas for other (older) Marvell controllers, or for other older disks? > Do you know what support that's supposed to provide? Unfortunately I have no visibility into what Lenovo is providing :(