Missing bio_resid on error case in mfi_disk_complete()

Doug Ambrisko ambrisko at ambrisko.com
Fri Mar 30 22:25:25 UTC 2012


Andrew Boyer writes:
| Hello Doug,
| We noticed that the AIO system doesn't react to I/O failures that 
| come from mfi disks; the commands appear to complete.  The upper 
| layers are expecting bio->bio_resid to be non-zero if the bio failed.  
| The patch below matches what scsi_da.c does in this case and also what 
| mfi itself does in mfi_disk_strategy() if the I/O is invalid.
| 
| Does this make sense?  I meant to get this email sent before you 
| started on the big merge.
| 
| Thanks,
|  Andrew
| 
| Index: sys/dev/mfi/mfi_disk.c
| ===================================================================
| --- sys/dev/mfi/mfi_disk.c	(revision 233707)
| +++ sys/dev/mfi/mfi_disk.c	(working copy)
| @@ -274,6 +274,7 @@
|  	hdr = bio->bio_driver1;
|  
|  	if (bio->bio_flags & BIO_ERROR) {
| +		bio->bio_resid = bio->bio_bcount;
|  		if (bio->bio_error == 0)
|  			bio->bio_error = EIO;
|  		disk_err(bio, "hard error", -1, 1);

Looks good to me.  I can check it in after I finish the merge.
The driver is compile tested now.  So I'm about to check that in.

Thanks,

Doug A.


More information about the freebsd-scsi mailing list