that stupid aic7xxx AHC_NSEGS bug
We simply forget to initialize scb->sg_count in the non use_sg case, so if the previous usage of that scb has sg_count==AHC_NSEGS then we'd hit that panic erroneously. Here is the fix below.
"It can't possibly be my driver, something broke in some Linux subsystem which is making my driver break", sheesh get over it Justin...
scsi/aic7xxx/aic7xxx_linux.c renamed from scsi/aic7xxx/aic7xxx_linux.c.~1~ +1 -0
@@ -1699,6 +1699,7 @@ cmd->request_buffer, cmd->request_bufflen, scsi_to_pci_dma_dir(cmd->sc_data_direction)); + scb->sg_count =0 ; scb->sg_count =ahc_linux_map_seg (ahc, scb, sg, addr, cmd->request_bufflen);
Quoted 4 lines
We simply forget to initialize
scb->sg_countin the nonuse_sgcase, so if the previous usage of that scb has sg_count==AHC_NSEGS then we'd hit that panic erroneously. Here is the fix below.
Which just goes to show you have useful Steve Lord's report was. Even someone unfamiliar with the driver could figure this out once someone bothered to provide decend debuggin info.
Quoted 3 lines
"It can't possibly be my driver, something broke in some Linux subsystem which is making my driver break", sheesh get over it Justin...
That's a bit unfair David, and its also not an acurate quote. I did say to Nick Pasich, "I guarantee you though, it is not the aic7xxx driver's fault" on 11/26, but ever since it became apparent that I was wrong (11/29 perhaps?), I've been working with Nick to try and reproduce the problem. We all make mistakes David, even you. Get over it. 8-)
--
Justin
From: "Justin T. Gibbs" <[email protected]>
Date: Thu, 13 Dec 2001 13:57:43 -0700
We all make mistakes David, even you.
Some of us are just a little bit more of a pain in the ass about it, and this was my point.
Quoted 7 lines
From: "Justin T. Gibbs" <[email protected]>
Date: Thu, 13 Dec 2001 13:57:43 -0700We all make mistakes David, even you.
Some of us are just a little bit more of a pain in the ass about it, and this was my point.
Well, you made it brilliantly. Bravo!
--
Justin