On Thu, Oct 13, 2005 at 04:14:03PM -0600, Moore, Eric Dean wrote: > > I'm also wondering if pci_disable_device() should be added to > > the error handling code. > > I guess that sounds good adding pci_disable_device(). > Can you resend with patch that handles this(failure of pci_set_dma_mask, > and the kmalloc right below it); I can look at this. > as well as calling pci_disable_device() from mpt_detach() ? I think pci_disable_device() changes should be handled seperately. This can have nasty side effects for driver unload/load and needs real testing before being committed to the tree. I'm willing to do this...but not right away. First I want to see mpt "sort of" working on parisc-linux and then I can bang on load/unload paths. > > +mpterr_freeirq: ... > Howabout using out_free_irq, and out_free_ioc, instead > of mpterr_freeirq, and mpterr_freeioc, respectively. Sure. I just wanted to use unique names - easier to locate when they does show up in "grep free_irq *.c" output. I'll resubmit after I look at the two additional error cases. cheers, grant - : 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
