Bug 2376
Summary: | Devices aren't recognized with Promise SATA 150 (20375-based) | ||
---|---|---|---|
Product: | IO/Storage | Reporter: | Jonathan Solomon (jsolomon) |
Component: | Serial ATA | Assignee: | Jeff Garzik (jgarzik) |
Status: | REJECTED INSUFFICIENT_DATA | ||
Severity: | blocking | CC: | bunk, netdragon |
Priority: | P2 | ||
Hardware: | i386 | ||
OS: | Linux | ||
Kernel Version: | 2.6.5-1.332 | Subsystem: | |
Regression: | --- | Bisected commit-id: |
Description
Jonathan Solomon
2004-03-26 11:10:20 UTC
Whoops. Forgot to mention that this combination of hardware worked fine with the source-code supplied by Promise under the 2.4 kernel but, of course, that code doesn't compile under 2.6. This bug persists in 2.6.5-1.315 I have the same card and have identified the same issue. Realize that the standard ATA connector still makes it detected as SATA by the BIOS. 2.6.5-1.358 I think you are correct that its not doing something right with the standard ATA headers. It is probably probing for just the SATA. It says in the manual I linked to at the bottom that Parallel ATA is port 3. I don't see anything about ata3 in the logs. I don't know much about device drivers on Linux either. This is also not a Fastrack card. It says that in the manual I linked to at the bottom. If you got this with your 250GB drive as I did, then I assume the same holds true for you. May 9 16:45:49 localhost kernel: SCSI subsystem initialized May 9 16:45:49 localhost kernel: ata1: SATA max UDMA/133 cmd 0x4283A200 ctl 0x4283A238 bmdma 0x0 irq 10 May 9 16:45:49 localhost kernel: ata2: SATA max UDMA/133 cmd 0x4283A280 ctl 0x4283A2B8 bmdma 0x0 irq 10 May 9 16:45:49 localhost kernel: ata1: no device found (phy stat 00000000) May 9 16:45:49 localhost kernel: scsi0 : sata_promise May 9 16:45:50 localhost kernel: ata2: no device found (phy stat 00000000) May 9 16:45:50 localhost kernel: scsi1 : sata_promise I noticed that message here: http://lxr.linux.no/source/drivers/scsi/libata-core.c?v=2.6.5#L1093 Promise code: http://lxr.linux.no/source/drivers/scsi/sata_promise.c?v=2.6.5 Maybe someone can find something useful: http://www.promise.com/support/file/manual/2_SATA%20150%20TX2%20TX4%20User%20v1.3.pdf These cards ship with the Maxtor 250GB hard drives. http://lxr.linux.no/source/drivers/scsi/sata_promise.c?v=2.6.5#L1751 Maybe this needs to be changed to: probe_ent->n_ports = 3; Re #4: Yes, those are exactly the messages I got. Re #5: Unfortunately that was the first thing that occurred to me. I increased the count and added the corequisite initialization but all it did was add additional "no device found" messages. I did indeed get the adapter with my HD, but it _is_ a 20375 card and I don't believe that the non-Fasttrack SATA 150 is. Jonathon: I got a kernel oops instead when I changed it to 4 :-) This might be of interest: http://linux.derkeiler.com/Mailing-Lists/Kernel/2004-03/6344.html I haven't tried it yet. http://www.kernel.org/pub/linux/kernel/people/jgarzik/libata/old/2.6.5-rc3-libata2.log ChangeSet@1.1673.7.9, 2004-03-26 01:13:20-05:00, jgarzik@redhat.com [libata sata_promise] minor initialization updates * remove incorrect PATA port check * enable undocumented bit 13 in flash control register, because the Promise driver does so. * wait 10 ms after setting TBG mode, for the same reason. See also: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=120680 I built and ran the 2.6.6-rc3 kernel hoping this would be fixed, but this still isn't fixed. :-( This problem occurs with the 2.6.6 stable kernel and 2.6.6rc3 kernel as well. I cannot fix it... but I have the card and drive attached to it working in 2.4.x with promise's drivers. BTW, the Promise SATA150 TX2 PDC20375 is NOT a promise fasttrak tx2 plus. I have the same exact card you have, with the same exact problem. It IS a problem with the kernel's driver for it. Hopefully they will fix it. I get a similar problem with my Asus A7V8X mobo, which has an on-board Promise 20376 connected to one standard IDE port and two SATA ports. I have a hard drive connected to the IDE port. The drive is recognized by the 2.4.x kernels with the Promise-supplied ft3xx.o driver, but is NOT recognized by the 2.6.5 kernels (in Knoppix 3.4 and Gentoo 2004.1) with the sata_promise.ko driver. Is this problem still present in recent vanilla ftp.kernel.org 2.6 kernels? Please reopen this bug if it's still present with recent kernels. |