On Wed, 8 Mar 2006, Sebastian Frei wrote: > I got during a backup the following messages in the system log: > > Mar 8 03:44:08 PINGUIN dc395x: sg_update_list: sg_to_virt failed > Mar 8 03:44:08 PINGUIN dc395x: sg_update_list: sg_to_virt failed > Mar 8 03:44:08 PINGUIN drivers/scsi/dc395x.c: Please, contact > <linux-scsi@xxxxxxxxxxxxxxx> to help improve support > for your system. Thanks for writing! You are the second one. The short background why this message was put in the driver - there is a known bug in the driver, which is hard to trigger, so, we couldn't test our bugfixes. Now that you wrote to the list, I see that you also have a 64-bit system, as well as the first reporter. Unlike him, you have a tape-drive on the bus, he had a scanner. But in both cases it's not a block device - whether or not that is relevant. For comparison - on my 32-bit system neither a scanner nor a tape trigger this path. > It is an opteron, 1GB RAM, a Tekram DC-315U controller with a HP DDS4 tape > drive, connected with a narrow->wide SCSI adaptor. > > Some info: > > Linux PINGUIN 2.6.15.6 #15 Mon Mar 6 11:30:52 CET 2006 x86_64 x86_64 x86_64 > GNU/Linux > > Mar 7 19:31:01 PINGUIN dc395x: Tekram DC395(U/UW/F), DC315(U) - ASIC > TRM-S1040 v2.05, 2004/03/08 > Mar 7 19:31:01 PINGUIN GSI 19 sharing vector 0xC1 and IRQ 19 > Mar 7 19:31:01 PINGUIN ACPI: PCI Interrupt 0000:00:0e.0[A] -> GSI 19 (level, > low) -> IRQ 19 > Mar 7 19:31:01 PINGUIN dc395x: Used settings: AdapterID=07, Speed=0(20.0MHz), > dev_mode=0x57 > Mar 7 19:31:01 PINGUIN dc395x: AdaptMode=0x0f, Tags=4(16), > DelayReset=1s > Mar 7 19:31:01 PINGUIN dc395x: Connectors: int50 Termination: Auto Low High > Mar 7 19:31:01 PINGUIN dc395x: Performing initial SCSI bus reset > Mar 7 19:31:01 PINGUIN scsi0 : Tekram DC395(U/UW/F), DC315(U) - ASIC > TRM-S1040 v2.05, 2004/03/08 > Mar 7 19:31:01 PINGUIN dc395x: Target 03: Sync: 48ns Offset 15 (20.8 MB/s) > Mar 7 19:31:01 PINGUIN Vendor: HP Model: C5683A Rev: C111 > Mar 7 19:31:01 PINGUIN Type: Sequential-Access ANSI SCSI > revision: 02 > Mar 7 19:31:01 PINGUIN st: Version 20050830, fixed bufsize 1048576, s/g segs > 256 > Mar 7 19:31:01 PINGUIN st 0:0:3:0: Attached scsi tape st0<4>st0: try direct > i/o: yes (alignment 512 B), max page re > achable by HBA 1048575 Do you have those messages "sg_to_virt failed" reproducibly? Does the tape work after that? > Some things I should do ? Sure. If you have some spare time to test a couple of patches we could try some patches. I think, we annoyed the list enough with the first reporter, so, we can as well take it off-list. The first reporter seems to have no time for further testing at the moment, so, if you want to play with the driver a bit just reply off-list to me. Thanks Guennadi --- Guennadi Liakhovetski - : send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
