September 16, 200916 yr When are new drivers added to unRaid. Is there a place on the forum to ask for added controller card driver support? It doesn't seem to fit under new feature request. Are drivers added periodically to new releases or on a request bases. I would like to use the Supermicro 8-Port SAS/SATA card AOC-SASLP-MV8 but it doesn't seem to be supported yet.
September 16, 200916 yr In the default situation: First and foremost, it depends on when the Linux Kernel adds drivers. Second, it depends on when Tom upgrades unRAID to the new Linux Kernel. Third, it depends on when YOU upgrade to the new unRAID release. In the power-user situation: First and foremost, it depends on when the Linux Kernel adds drivers. Second, it depends on when YOU upgrade to the new Linux Kernel.
September 16, 200916 yr I suppose the über power user option would be: When YOU add the drivers to the Linux kernel.
September 16, 200916 yr Author The documentation for the card says it is supported by RedHat Enterprise, Fedora Linus 9, and SuSE Linux Enterprise, is the kernel different for different Linus distributions?
September 17, 200916 yr The documentation for the card says it is supported by RedHat Enterprise, Fedora Linus 9, and SuSE Linux Enterprise, is the kernel different for different Linus distributions? Yes, the kernel is different (in compliation options) for different distributions and architecture.
October 8, 200916 yr Rick check out http://lime-technology.com/forum/index.php?topic=3382.45 and http://lime-technology.com/forum/index.php?topic=3109.0 The new linux core has a new driver that might help. We'll have to wait for Tom to get us another beta with the new linux on it. http://74.125.47.132/search?q=cache:pnQYwMiSrScJ:www.gossamer-threads.com/lists/linux/kernel/1131237+HDIO_GET_IDENTITY+2.6.31&cd=1&hl=en&ct=clnk&gl=us&client=firefox-a Has a user posting that 2.6.31 has the driver and supports it.
October 8, 200916 yr I do not have that hardware, but FWIW, my brief stint last weekend with Linux 2.6.31.1 was not pleasant. The initial boot and quick test worked fine. However, after half a day the kernel paniced and crashed. I didn't have a chance to debug the issue. Looking today, I see they have 2.6.31.3 released. After a quick look at the changelogs, this may have been the issue (patched in 2.6.31.2): commit 6177445fb4f5d65dde3c883b9987c840178fb86d Author: Kashyap, Desai <[email protected]> Date: Wed Sep 2 11:44:57 2009 +0530 mptsas : PAE Kernel more than 4 GB kernel panic commit c55b89fba9872ebcd5ac15cdfdad29ffb89329f0 upstream. This patch is solving problem for PAE kernel DMA operation. On PAE system dma_addr and unsigned long will have different values. Now dma_addr is not type casted using unsigned long. Signed-off-by: Kashyap Desai <[email protected]> Signed-off-by: James Bottomley <[email protected]> Cc: Jan Beulich <[email protected]> Signed-off-by: Greg Kroah-Hartman <[email protected]>
October 11, 200916 yr Initial indications with 2.6.31.3 is it behaves better than 2.6.31.1 did. The system has only been up for 7.5 hours, but it's been stable so far. It faired well with some Usenet / Torrent activity to a non-array drive. All 4 Gigs of memory have been utilized for cache and no crashes occurred. An additional 13Gigs of data written to the array seems fine so far. I'll perform a more extensive test later, but so far so good. I did have to do the typical patching of md.c and md_private.h to get the drivers to be compliant with kernels 2.6.30.x and above (see code blocks below) --- /cache/.root/unraid-4.5b6/usr/src/linux/drivers/md/md.c 2009-05-06 02:59:56.0000000 00 -0400 +++ linux-2.6.31.3-unraid/drivers/md/md.c 2009-10-08 18:16:49.000000000 -0400 @@ -481,7 +481,8 @@ */ static int import_device(dev_t dev, mdk_rdev_t *rdev, int unit) { - struct hd_driveid driveid; + u16 driveid[ATA_ID_WORDS]; struct hd_geometry geometry; int err = 0; @@ -508,13 +509,13 @@ /* get id information */ /* note: these strings are not null-terminated */ - err = ioctl_by_bdev( rdev->bdev, HDIO_GET_IDENTITY, (unsigned long)&driveid); + err = ioctl_by_bdev( rdev->bdev, HDIO_GET_IDENTITY, driveid); if (err) { printk("md: import disk%d: HDIO_GET_IDENTITY ioctl error: %d\n", unit, err) ; goto out; } - memcpy( rdev->model, driveid.model, sizeof(rdev->model)); - memcpy( rdev->serial_no, driveid.serial_no, sizeof(rdev->serial_no)); + memcpy( rdev->model, (char *)(driveid + ATA_ID_PROD), sizeof(rdev->model)); + memcpy( rdev->serial_no, (char *)(driveid + ATA_ID_SERNO), sizeof(rdev->serial_no)) ; /* get size (in sectors, ie, 512-byte blocks) */ err = ioctl_by_bdev( rdev->bdev, BLKGETSIZE, (unsigned long)&rdev->size); --- /cache/.root/unraid-4.5b6/usr/src/linux/drivers/md/md_private.h 2009-05-06 02:59:56.000000000 -0400 +++ linux-2.6.31.3-unraid/drivers/md/md_private.h 2009-10-08 18:16:49.000000000 -0400 @@ -45,6 +45,7 @@ #include <asm/unaligned.h> #include <linux/hdreg.h> #include <net/checksum.h> +#include <linux/libata.h> /* * Different major versions are not compatible.
October 12, 200916 yr BRiT, When you get a chance, Do we still need to manually edit the md_private.h and md.c files since Beta 7 is now using 2.6.30.8 kernel?
October 12, 200916 yr I'm pulling it down now, but wont get a chance to have a look at the changes until later. I'll let everyone know.
October 12, 200916 yr Good news. You DO NOT have to make changes to md.c and md_private.h anymore. The 4.5b7 source files are compliant with kernels 2.6.30.x and above. This should make things easier for the less experienced or non-experimental hackers.
Archived
This topic is now archived and is closed to further replies.