Messages in this thread | | | Date | Wed, 08 Dec 1999 01:19:44 +0100 | From | Martin Dalecki <> | Subject | Re: PATCH for 2.3.30: block device setup cleanup. |
| |
Andries.Brouwer@cwi.nl wrote: > > Marcin: > > > So what's the problem? I'm starting to eat this fish from the back, so > > you could start doing it at *the* *same* time from the head. > > No, it is your turn now.
There it is ;-). However I still didn't extend the semantics of the struct blk_dev_size blk_dev[] array into a conglomerate of per device not just per major object, becouse basically:
1. The speed optimizations I have applyed to ll_rw_blk.c are interresting for theyr own sake. All the setup done in blk_dev_init where blk_dev just get's initialized in one run will later be moved to some kind of per device allocation function. They help to reduce the number of direct refferences of blk_dev there too... (BTW. Just compare the size of ll_rw_blk.o befor and after and you will basically scream about the exhaustive switches on hardcoded major numbers found there....)
2. I Have to think over about what exactly to do with the similiar static arrays blkdevs and chrdevs in fs/devices.c. (Fortunately all the lookup done on this isn't in any place timinig crucial.
3. I wonted to give this patch out, since it's entierly clean in itself, so can be easly tested. (Or even permanently applyed.)
4. I still need time to dig a bit deeper into the loop and raid driver *mess*.
BTW. Surprisingly the transition for 2.3.29 to 2.3.30 went entierly smooth for me...
--Marcin[unhandled content-type:application/x-gzip] | |