Thanks for looking into this! Am 29.08.2005 um 20:33 schrieb Søren Schmidt: > On 29/08/2005, at 19:28, Stefan Bethke wrote: > >> Am 28.08.2005 um 18:46 schrieb Stefan Bethke: >> The most striking difference (to me) is that 5-stable thinks DMA66 >> is fine (and my tests show that at least reading is not a problem >> at all), while 6-beta wants to do PIO4, but seemingly gets stuck. >> > Uhm, both use PIO4 as far as I can tell.. > Sorry, no idea what I was looking at. > Enable debugging in ata-all.h and set ATA_R_DEBUG in request->flags > early in ata_queue_request() and see how far it gets. > > I'd suspect subtle timing in the lowest levels though.... > I still have the wrong entry in fstab, so this is from right after specifying the proper root device: After the final "ad2: req=0xc17fc000 READ completed callback/wakeup", nothing else happens. Stefan p.s. this is what I did to enable debugging: little:/usr/src/sys/dev/ata# cvs diff cvs diff: Diffing . Index: ata-all.h =================================================================== RCS file: /home/ncvs/src/sys/dev/ata/ata-all.h,v retrieving revision 1.103.2.2 diff -u -r1.103.2.2 ata-all.h --- ata-all.h 25 Aug 2005 16:21:05 -0000 1.103.2.2 +++ ata-all.h 29 Aug 2005 18:07:33 -0000 _at__at_ -391,7 +391,7 _at__at_ }; /* define this for debugging request processing */ -#if 0 +#if 1 #define ATA_DEBUG_RQ(request, string) \ { \ if (request->flags & ATA_R_DEBUG) \ Index: ata-queue.c =================================================================== RCS file: /home/ncvs/src/sys/dev/ata/ata-queue.c,v retrieving revision 1.50 diff -u -r1.50 ata-queue.c --- ata-queue.c 28 Jun 2005 09:06:52 -0000 1.50 +++ ata-queue.c 29 Aug 2005 18:57:15 -0000 _at__at_ -54,6 +54,7 _at__at_ ata_queue_request(struct ata_request *request) { struct ata_channel *ch = device_get_softc(device_get_parent (request->dev)); + request->flags |= ATA_R_DEBUG; /* mark request as virgin (this might be a ATA_R_REQUEUE) */ request->result = request->status = request->error = 0; -- Stefan Bethke <stb_at_lassitu.de> Fon +49 170 346 0140
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:42 UTC