On Sat, 1 Apr 2006, Christoph Hellwig wrote: > On Sat, Apr 01, 2006 at 05:35:13PM +0300, Kai Makisara wrote: > > I have done some more investigation based on the message from Gerard > > Roudier that Jens found. I fetched the FreeBSD driver. It had a > > work-around for the 896 rev. 1 chips but I did not find that in the Linux > > version. The work-around does build two device s/g elements for any > > original s/g segment that crosses a 16 MB boundary. > > > > In Linux we can do that by setting the .dma_boundary field in the > > scsi_host_template. Would it be an acceptable solution to set the 16 MB > > limit for all sym53c8xx adapters? It would take care of the known errata > > and Gerards suspicions about possible problems with other chips. > > We can also set the flag in the scsi_host so not all adapters are punished. > Yes, but the question is, do we want to be conservative and set the mask for all adapters (including the ones possibly having errata we don't know), or less conservative and set it for the ones we know are buggy. For those who make decisions, below is an attempt at limiting dma only for the chip we know is having a problem. Builds and boots (with 53c1010). Kai --- linux-2.6.16/drivers/scsi/sym53c8xx_2/sym_glue.c 2006-03-20 23:13:18.000000000 +0200 +++ linux-2.6.16-k1/drivers/scsi/sym53c8xx_2/sym_glue.c 2006-04-01 23:16:04.000000000 +0300 @@ -1576,6 +1576,7 @@ static struct Scsi_Host * __devinit sym_ struct host_data *host_data; struct sym_hcb *np = NULL; struct Scsi_Host *instance = NULL; + struct sym_chip *chip = &dev->chip; struct pci_dev *pdev = dev->pdev; unsigned long flags; struct sym_fw *fw; @@ -1712,6 +1713,13 @@ static struct Scsi_Host * __devinit sym_ BUG_ON(sym2_transport_template == NULL); instance->transportt = sym2_transport_template; + /* + * Due to errata on the 896 rev. 1 chip, don't allow s/g segments to + * cross 16 MB boundary + */ + if (pdev->device == PCI_DEVICE_ID_NCR_53C896 && chip->revision_id < 2) + instance->dma_boundary = 0xFFFFFF; + spin_unlock_irqrestore(instance->host_lock, flags); return instance; @@ -1978,7 +1986,8 @@ static struct scsi_host_template sym2_te .eh_bus_reset_handler = sym53c8xx_eh_bus_reset_handler, .eh_host_reset_handler = sym53c8xx_eh_host_reset_handler, .this_id = 7, - .use_clustering = DISABLE_CLUSTERING, + .use_clustering = ENABLE_CLUSTERING, + .max_sectors = 0xFFFF, #ifdef SYM_LINUX_PROC_INFO_SUPPORT .proc_info = sym53c8xx_proc_info, .proc_name = NAME53C8XX, - : 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
