Messages in this thread | | | Subject | Re: [ATM] first pass at fixing atm spinlock | Date | Wed, 19 Mar 2003 10:13:45 -0500 | From | chas williams <> |
| |
In message <20030319025734.C35024@sfgoth.com>,Mitchell Blank Jr writes: >> i dont know. i believe all of the adapters do a synchronous close. >I'm really not sure it's that safe. At the very least the drivers all >need to make sure that their ->close() excludes their interrupt/bh work
i suppose i could be he-centric. when the rx and tx close complete you know you wont be getting anymore traffic in the queues regarding that vpi/vcc. i suppose other cards might now have this feature. some drivers just keep a pointer to vcc in a table. so it they dont use atm_vcc_lookup() each time they need a receieve then i guess the vcc's will need some sort of ref counting. i am generally against this scheme, but it could be made to work.
>Great - now we just have to do the same thing for vcc's :-)
actually i think the vcc related code is 'broken'. since the vcc is really attached to a struct sock, the vcc code should be more sock-centric. sock provides ref counting, locks, linking, hash support. at the time the atm code was initially written, struct sock might not have been generic enough. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |