Final patch for -current
Ganbold
ganbold at micom.mng.net
Thu Feb 12 12:07:31 PST 2009
Sean,
How about following redundant loop near the end of fwohci.c?
--- sys/dev/firewire/fwohci.c 2008-11-22 14:39:45.000000000 +0800
+++ /var/current/mysrc/sys/dev/firewire/fwohci.c 2008-11-25 11:36:09.000000000 +0800
@@ -2973,7 +2973,7 @@
db_tr = STAILQ_NEXT(db_tr, link);
resCount = FWOHCI_DMA_READ(db_tr->db[0].db.desc.res)
& OHCI_COUNT_MASK;
- } while (resCount == 0)
+ }
printf(" done\n");
dbch->top = db_tr;
dbch->buf_offset = dbch->xferq.psize - resCount;
thanks,
Ganbold
Sean Bruno wrote:
> Allrighty, this patch wraps up my -current work and should establish a
> nice baseline for my debugging of problems.
>
> Introduce struct sbp_dev.bustgtlun to replace the overloaded
> show_sdev_info() in most places. This is an optimization of the debug
> case fortunately. Stash the BB:TT:LL string here for later
> device_printf() goodness.
>
> More debugging cleanup printf()-->device_printf()
>
> Explicity set the callout sdev->login_callout to acquire the sbp mtx.
>
> Rename sbp_alloc_target() variable i to new_target_id. Clarity is good,
> even in temporary variables.
>
> Cleanup and document the sbp_login() function. This function is a bit
> hairy, so document what it's supposed to do.
>
> Add some locking in sbp_probe_target()
>
> Assert the sbp mtx in sbp_write_cmd_locked()
>
> Avoid LoR in sbp_mgm_orb() and assert sbp mtx locked.
>
> Acquire sbp mtx before sbp_mgm_orb() is invoked.
>
> Sean
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> freebsd-firewire at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-firewire
> To unsubscribe, send any mail to "freebsd-firewire-unsubscribe at freebsd.org"
--
I have a friend whose a billionaire. He invented Cliff's notes. When I
asked him how he got such a great idea he said, "Well first I... I
just... to make a long story short..." -- Steven Wright
More information about the freebsd-firewire
mailing list