Re: Port Multiplier Patch 2.6.22.1 on Alpha
Tom Evans wrote: I do not see any console or syslog messages at the time - perhaps I do not have a debug level set appropriately? Executing dmesg after the event should give you the messages if the kernel has spoken anything. Also, as an aside, I noticed that the device scan on the remote array shows this device: Sep 8 00:15:12 chronictown kernel: [4194044.507921] ata4.05: ATA-6: Config Disk, PLS12052, max UDMA/133 Sep 8 00:15:12 chronictown kernel: [4194044.507921] ata4.05: 640 sectors, multi 1: LBA Sep 8 00:15:12 chronictown kernel: [4194044.507921] ata4.05: device is on DMA blacklist, disabling DMA Sep 8 00:15:12 chronictown kernel: [4194044.507921] ata4.05: configured for PIO4 I couldn't find any reference - is this firmware for the PMP or something like that? That's the pseudo device for configuration. It can be safely ignored unless it causes problems. Thanks. -- tejun - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Port Multiplier Patch 2.6.22.1 on Alpha
Tom Evans wrote: No real performance change, but the # of unaligned kernel goes from ~14 to zero on my system. It's really just cosmetic I believe. Care to post signed-off patch? -- tejun - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Port Multiplier Patch 2.6.22.1 on Alpha
Tom Evans wrote: I decided to put the Norco 4618 card into the first PCI slot of my DS20. It and the PMP support seem to like that location much better - I'm successfully creating an array at the moment. The system still responds, but accessing currently running filesystems (another array and the local scsi boot disk) seems to freeze for 20-30 seconds. Does kernel complain anything during the freeze? I was able to move the graphics card to another slot (which for a period in the past was also limited to the first PCI slot). Maybe this will make sense to some folks out there - thanks! To me, not much yet. :-( -- tejun - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Port Multiplier Patch 2.6.22.1 on Alpha
No real performance change, but the # of unaligned kernel goes from ~14 to zero on my system. It's really just cosmetic I believe. ...tom Tejun Heo wrote: Tom Evans wrote: In sil24_config_port, the writel's for /* zero error counters */ should be writew's - they are 16bit control registers. Only PORT_DECODE_ERR_THRESH and PORT_CRC_ERR_THRESH appear to be unaligned, but why risk a stl when a stw (on platforms that support it) can be used. That part of the code was taken verbatim from the Silicon Image sample driver, which I'm sure have been tested only on x86s. Those error threshold feature isn't really used and using writew's should be fine. Do things work better after changing to writew's? - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Port Multiplier Patch 2.6.22.1 on Alpha
Tejun Heo wrote: Does kernel complain anything during the freeze? I do not see any console or syslog messages at the time - perhaps I do not have a debug level set appropriately? Also, as an aside, I noticed that the device scan on the remote array shows this device: Sep 8 00:15:12 chronictown kernel: [4194044.507921] ata4.05: ATA-6: Config Disk, PLS12052, max UDMA/133 Sep 8 00:15:12 chronictown kernel: [4194044.507921] ata4.05: 640 sectors, multi 1: LBA Sep 8 00:15:12 chronictown kernel: [4194044.507921] ata4.05: device is on DMA blacklist, disabling DMA Sep 8 00:15:12 chronictown kernel: [4194044.507921] ata4.05: configured for PIO4 I couldn't find any reference - is this firmware for the PMP or something like that? Thanks, ...tom - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Port Multiplier Patch 2.6.22.1 on Alpha
I decided to put the Norco 4618 card into the first PCI slot of my DS20. It and the PMP support seem to like that location much better - I'm successfully creating an array at the moment. The system still responds, but accessing currently running filesystems (another array and the local scsi boot disk) seems to freeze for 20-30 seconds. I was able to move the graphics card to another slot (which for a period in the past was also limited to the first PCI slot). Maybe this will make sense to some folks out there - thanks! ...tom Jeff Garzik wrote: Tom Evans wrote: Have you heard of such and issue with the 3124-2/4276 combination? Is anyone else trying this driver on an Alpha? Alpha should not be operationally much different from other little endian 64-bit platforms. The PCI swizzle might be weird (what is your Alpha platform?), is the only Alpha-specific thing that comes to mind. I have similar such behavior before when using drivers that generate many unaligned accesses - I only have seen 2 places they occur in sata_sil24 and have fixed them myself, but that made no difference. We definitely want to fix all these you find. Please post more info... Also, if you are highly motivated, I would love to see if other SATA cards have similar problems on Alpha... Jeff - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Port Multiplier Patch 2.6.22.1 on Alpha
I decided to put the Norco 4618 card into the first PCI slot of my DS20. It and the PMP support seem to like that location much better - I'm successfully creating an array at the moment. The system still responds, but accessing currently running filesystems (another array and the local scsi boot disk) seems to freeze for 20-30 seconds. I was able to move the graphics card to another slot (which for a period in the past was also limited to the first PCI slot). Maybe this will make sense to some folks out there - thanks! ...tom Jeff Garzik wrote: Tom Evans wrote: Have you heard of such and issue with the 3124-2/4276 combination? Is anyone else trying this driver on an Alpha? Alpha should not be operationally much different from other little endian 64-bit platforms. The PCI swizzle might be weird (what is your Alpha platform?), is the only Alpha-specific thing that comes to mind. I have similar such behavior before when using drivers that generate many unaligned accesses - I only have seen 2 places they occur in sata_sil24 and have fixed them myself, but that made no difference. We definitely want to fix all these you find. Please post more info... Also, if you are highly motivated, I would love to see if other SATA cards have similar problems on Alpha... Jeff - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Port Multiplier Patch 2.6.22.1 on Alpha
I decided to put the Norco 4618 card into the first PCI slot of my DS20. It and the PMP support seem to like that location much better - I'm successfully creating an array at the moment. The system still responds, but accessing currently running filesystems (another array and the local scsi boot disk) seems to freeze for 20-30 seconds. I was able to move the graphics card to another slot (which for a period in the past was also limited to the first PCI slot). Maybe this will make sense to some folks out there - thanks! ...tom Jeff Garzik wrote: Tom Evans wrote: Have you heard of such and issue with the 3124-2/4276 combination? Is anyone else trying this driver on an Alpha? Alpha should not be operationally much different from other little endian 64-bit platforms. The PCI swizzle might be weird (what is your Alpha platform?), is the only Alpha-specific thing that comes to mind. I have similar such behavior before when using drivers that generate many unaligned accesses - I only have seen 2 places they occur in sata_sil24 and have fixed them myself, but that made no difference. We definitely want to fix all these you find. Please post more info... Also, if you are highly motivated, I would love to see if other SATA cards have similar problems on Alpha... Jeff - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Port Multiplier Patch 2.6.22.1 on Alpha
Sorry about the spam - said system is my mail server and the mail program, although I though had timed-out the sends still got the messages thru. ...tom - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Port Multiplier Patch 2.6.22.1 on Alpha
I have included my boot messages and an lspci, in any of the info is useful. I do have a VIA based sata board that works well also installed - any potential for conflict? ...tom - Boot Log - [4194001.852669] Linux version 2.6.22.1-ds20 ([EMAIL PROTECTED]) (gcc version 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)) #1 SMP Wed Aug 29 03:40:59 EDT 2007 [4194001.852669] Booting on Tsunami variation Goldrush using machine vector DP264 from SRM [4194001.852669] Major Options: SMP LEGACY_START VERBOSE_MCHECK MAGIC_SYSRQ [4194001.852669] Command line: ro root=/dev/sda3 ide=reverse single [4194001.852669] memcluster 0, usage 1, start0, end 256 [4194001.852669] memcluster 1, usage 0, start 256, end 130976 [4194001.852669] memcluster 2, usage 1, start 130976, end 131072 [4194001.852669] freeing pages 256:384 [4194001.852669] freeing pages 999:130976 [4194001.852669] reserving pages 999:1001 [4194001.852669] 4096K Bcache detected; load hit latency 20 cycles, load miss latency 106 cycles [4194001.852669] Console graphics on hose 0 [4194001.852669] SMP: 2 CPUs probed -- cpu_present_map = 3 [4194001.852669] On node 0 totalpages: 130976 [4194001.852669] DMA zone: 895 pages used for memmap [4194001.852669] DMA zone: 0 pages reserved [4194001.852669] DMA zone: 130081 pages, LIFO batch:15 [4194001.852669] Normal zone: 0 pages used for memmap [4194001.852669] Built 1 zonelists. Total pages: 130081 [4194001.852669] Kernel command line: ro root=/dev/sda3 ide=reverse single [4194001.852669] PID hash table entries: 4096 (order: 12, 32768 bytes) [4194001.852669] Using epoch = 1952 [4194001.855599] Console: colour VGA+ 80x25 [4194001.858529] Dentry cache hash table entries: 131072 (order: 7, 1048576 bytes) [4194001.861459] Inode-cache hash table entries: 65536 (order: 6, 524288 bytes) [4194001.907357] Memory: 1029240k/1047808k available (3251k kernel code, 15888k reserved, 225k data, 168k init) [4194001.908333] Calibrating delay loop... 996.00 BogoMIPS (lpj=486400) [4194001.926888] Mount-cache hash table entries: 512 [4194001.926888] SMP starting up secondaries. [4194001.935677] Calibrating delay loop... 996.00 BogoMIPS (lpj=486400) [4194001.954232] Brought up 2 CPUs [4194001.954232] SMP: Total of 2 processors activated (1998.25 BogoMIPS). [4194002.035287] migration_cost=2929 [4194002.036263] NET: Registered protocol family 16 [4194002.041146] PCI: Bridge: :00:08.0 [4194002.041146] IO window: 8000-8fff [4194002.041146] MEM window: 0a00-0a0f [4194002.042122] PREFETCH window: 0a10-0a1f [4194002.042122] SMC37c669 Super I/O Controller found @ 0x3f0 [4194002.046029] Linux Plug and Play Support v0.97 (c) Adam Belay [4194002.047005] SCSI subsystem initialized [4194002.047005] libata version 3.00 loaded. [4194002.048958] NET: Registered protocol family 2 [4194002.061654] IP route cache hash table entries: 8192 (order: 3, 65536 bytes) [4194002.061654] TCP established hash table entries: 32768 (order: 6, 786432 bytes) [4194002.063607] TCP bind hash table entries: 32768 (order: 6, 524288 bytes) [4194002.065560] TCP: Hash tables configured (established 32768 bind 32768) [4194002.065560] TCP reno registered [4194002.069466] srm_env: version 0.0.6 loaded successfully [4194002.071419] io scheduler noop registered [4194002.071419] io scheduler anticipatory registered (default) [4194002.071419] io scheduler deadline registered [4194002.071419] isapnp: Scanning for PnP cards... [4194002.427864] isapnp: No Plug Play device found [4194002.562630] rtc: Digital UNIX epoch (1952) detected [4194002.562630] Real Time Clock Driver v1.12ac [4194002.562630] Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing disabled [4194002.562630] serial8250: ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A [4194002.563607] serial8250: ttyS1 at I/O 0x2f8 (irq = 3) is a 16550A [4194002.565560] Floppy drive(s): fd0 is 2.88M [4194002.581185] FDC 0 is a post-1991 82077 [4194002.583138] nbd: registered device at major 43 [4194002.587044] Loading iSCSI transport class v2.0-724. [4194002.747200] sym0: 875 rev 0x4 at pci :01:00.0 irq 27 [4194002.834114] sym0: Symbios NVRAM, ID 7, Fast-20, SE, parity checking [4194002.834114] sym0: open drain IRQ line driver, using on-chip SRAM [4194002.834114] sym0: using LOAD/STORE-based firmware. [4194002.835091] sym0: SCSI BUS has been reset. [4194002.841927] scsi0 : sym-2.2.3 [4194003.003059] sym1: 875 rev 0x4 at pci :01:01.0 irq 26 [4194003.089974] sym1: Symbios NVRAM, ID 7, Fast-20, SE, parity checking [4194003.089974] sym1: open drain IRQ line driver, using on-chip SRAM [4194003.089974] sym1: using LOAD/STORE-based firmware. [4194003.089974] sym1: SCSI BUS has been reset. [4194003.097786] scsi1 : sym-2.2.3 [4194003.097786] sata_via 0001:02:08.0: version 2.2 [4194003.097786] sata_via 0001:02:08.0: routed to hard irq line 11 [4194003.098763] scsi2 : sata_via [4194003.098763] scsi3 : sata_via [4194003.098763] scsi4 :
Port Multiplier Patch 2.6.22.1 on Alpha
Hi Tejun and all - I have applied the libata-tj-stable patch to the 2.6.22.1 kernel and am having some issues. I'm not certain if it is the PMP patches, or my platform. I have a Silicon Images 3124-2 based PCI-X card (Norco 4618) for my Alpha system . I realize this is not the usual test platform and I understand if you decide to blame it on the platform. I also have a StorBox which is a 5-bay eSATA box based with the Si4276 port multiplier, which is noted to be quirky in the README. The problem arises when I try to create/assemble a linux software raid using 4 disks (level 6). At the time the system becomes very very slow - basically unresponsive - I can get an occasional keystroke in on the console, but the machine is essentially dead. If I pull the eSATAcable out of the card, the sata_si24 driver does its thing and the machine comes back to life. The unit is plugged into the port designated as the first - all drive bays contain a drive. I am able to individually address and to format individual drives without performance issue, but the creation of a raid is deadly Have you heard of such and issue with the 3124-2/4276 combination? Is anyone else trying this driver on an Alpha? I have similar such behavior before when using drivers that generate many unaligned accesses - I only have seen 2 places they occur in sata_sil24 and have fixed them myself, but that made no difference. Are there any logs or output that might help? Should I attempt to upgrade the firmware of the devices? Thanks, ...tom - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Port Multiplier Patch 2.6.22.1 on Alpha
Tom Evans wrote: Have you heard of such and issue with the 3124-2/4276 combination? Is anyone else trying this driver on an Alpha? Alpha should not be operationally much different from other little endian 64-bit platforms. The PCI swizzle might be weird (what is your Alpha platform?), is the only Alpha-specific thing that comes to mind. I have similar such behavior before when using drivers that generate many unaligned accesses - I only have seen 2 places they occur in sata_sil24 and have fixed them myself, but that made no difference. We definitely want to fix all these you find. Please post more info... Also, if you are highly motivated, I would love to see if other SATA cards have similar problems on Alpha... Jeff - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Port Multiplier Patch 2.6.22.1 on Alpha
Thanks for the response Jeff ! Jeff Garzik wrote: Alpha should not be operationally much different from other little endian 64-bit platforms. The PCI swizzle might be weird (what is your Alpha platform?), is the only Alpha-specific thing that comes to mind. I am using a DS20 (Tsunami / Goldrush). I have similar such behavior before when using drivers that generate many unaligned accesses - I only have seen 2 places they occur in sata_sil24 and have fixed them myself, but that made no difference. We definitely want to fix all these you find. Please post more info... In sil24_config_port, the writel's for /* zero error counters */ should be writew's - they are 16bit control registers. Only PORT_DECODE_ERR_THRESH and PORT_CRC_ERR_THRESH appear to be unaligned, but why risk a stl when a stw (on platforms that support it) can be used. Also, if you are highly motivated, I would love to see if other SATA cards have similar problems on Alpha... I'll need to find a computer store with a decent return policy, but I would consider doing it ...tom - To unsubscribe from this list: send the line unsubscribe linux-ide in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
