[40] in linux-scsi channel archive
Generic SCSI & patch77
daemon@ATHENA.MIT.EDU (jered@vorlon.mit.edu)
Mon Jan 30 23:45:14 1995
To: linux-scsi@vger.rutgers.edu
From: jered@vorlon.mit.edu
Date: Tue, 17 Jan 1995 11:52:32 -0500
In patch 1.1.77, after which generic SCSI support has not worked, there
were a number of changes which I do not understand. (Comments in the
patches are needed! :-) )
These seem to be related:
+++ linux/drivers/scsi/sg.c Thu Jan 5 09:24:45 1995
@@ -116,6 +116,8 @@
}
if (!scsi_generics[dev].users)
scsi_generics[dev].timeout=SG_DEFAULT_TIMEOUT;
+ if (scsi_generics[dev].device->host->hostt->usage_count)
+ (*scsi_generics[dev].device->host->hostt->usage_count)++;
scsi_generics[dev].users++;
return 0;
}
@@ -124,6 +126,8 @@
{
int dev=MINOR(inode->i_rdev);
scsi_generics[dev].users--;
+ if (scsi_generics[dev].device->host->hostt->usage_count)
+ (*scsi_generics[dev].device->host->hostt->usage_count)--;
scsi_generics[dev].exclude=0;
wake_up(&scsi_generics[dev].generic_wait);
}
The next two appear to be correct, checking for proper command size (although
maybe I should change them back and try it out...)
The next, and final, patch seems to be some sort of interface change:
@@ -291,7 +311,8 @@
printk("do cmd\n");
#endif
scsi_do_cmd (SCpnt,(void *) cmnd,
- (void *) device->buff,amt,sg_command_done,device->timeout,SG_DEF
AULT_RETRIES);
+ (void *) device->buff,device->header.pack_len-size-sizeof(struct
sg_header),
+ sg_command_done,device->timeout,SG_DEFAULT_RETRIES);
#ifdef DEBUG
printk("done cmd\n");
#endif
And...that's it. 5 patches total (Assuming the breakage is in sg.c). *sigh*
Jered
jered@mit.edu
