From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeremy Huddleston Subject: Re: Promise PDC20269 - PIO Only? Date: Mon, 20 Oct 2003 13:15:33 -0700 Sender: linux-ide-owner@vger.kernel.org Message-ID: <200310201315.34213.jeremyhu@uclink4.berkeley.edu> References: <200310181819.08930.jeremyhu@uclink4.berkeley.edu> <200310191603.40227.jeremyhu@uclink4.berkeley.edu> <200310201949.39826.bzolnier@elka.pw.edu.pl> Mime-Version: 1.0 Content-Type: Text/Plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Return-path: Received: from 12-233-44-179.client.attbi.com ([12.233.44.179]:10179 "EHLO terminator.outersquare.org") by vger.kernel.org with ESMTP id S262747AbTJTULD convert rfc822-to-8bit (ORCPT ); Mon, 20 Oct 2003 16:11:03 -0400 In-Reply-To: <200310201949.39826.bzolnier@elka.pw.edu.pl> Content-Description: clearsigned data Content-Disposition: inline List-Id: linux-ide@vger.kernel.org To: Bartlomiej Zolnierkiewicz Cc: linux-ide@vger.kernel.org -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 /var/log/messages doesn't contain all of the boot log because the kernel's buffer is too small. When syslog-ng starts up, the earliest thing in the log buffer is some raid stuff... I need to know how to increase the buffer size... anyone have any ideas? Oct 20 12:58:21 eradicator syslog-ng[1736]: Changing permissions on special file /dev/tty12 Oct 20 12:58:21 eradicator syslog-ng[1736]: syslog-ng version 1.6.0rc3 starting Oct 20 12:58:21 eradicator /part6's event counter: 0000015d Oct 20 12:58:21 eradicator md3: max total readahead window set to 744k Oct 20 12:58:21 eradicator md3: 3 data-disks, max readahead per data-disk: 248k Oct 20 12:58:21 eradicator raid5: device ide/host4/bus1/target0/lun0/part6 operational as raid disk 3 Oct 20 12:58:21 eradicator raid5: device ide/host4/bus0/target0/lun0/part6 operational as raid disk 2 I do have the ide-ribbon debug output though since it was after the controller initialization output: Oct 20 12:58:21 eradicator hdc: attached ide-cdrom driver. Oct 20 12:58:21 eradicator hdc: ATAPI 32X CD-ROM drive, 128kB Cache, UDMA(33) Oct 20 12:58:21 eradicator Uniform CD-ROM driver Revision: 3.12 Oct 20 12:58:21 eradicator ide2: 80-pin cable, 80-pin cable, 0 Oct 20 12:58:21 eradicator blk: queue c0349538, I/O limit 4095Mb (mask 0xffffffff) Oct 20 12:58:21 eradicator ide2: 80-pin cable, 80-pin cable, 0 Oct 20 12:58:21 eradicator ide3: 80-pin cable, 80-pin cable, 0 Oct 20 12:58:21 eradicator blk: queue c0349994, I/O limit 4095Mb (mask 0xffffffff) Oct 20 12:58:21 eradicator ide3: 80-pin cable, 80-pin cable, 0 Oct 20 12:58:21 eradicator ide3: 80-pin cable, 80-pin cable, 0 Oct 20 12:58:21 eradicator blk: queue c0349ad4, I/O limit 4095Mb (mask 0xffffffff) Oct 20 12:58:21 eradicator ide3: 80-pin cable, 80-pin cable, 0 Oct 20 12:58:21 eradicator ide4: 80-pin cable, 80-pin cable, 0 Oct 20 12:58:21 eradicator blk: queue c0349df0, I/O limit 4095Mb (mask 0xffffffff) Oct 20 12:58:21 eradicator ide4: 80-pin cable, 80-pin cable, 0 Oct 20 12:58:21 eradicator ide5: 80-pin cable, 80-pin cable, 0 Oct 20 12:58:21 eradicator blk: queue c034a24c, I/O limit 4095Mb (mask 0xffffffff) Oct 20 12:58:21 eradicator ide5: 80-pin cable, 80-pin cable, 0 Oct 20 12:58:21 eradicator PCI: Found IRQ 3 for device 00:0d.0 (12:49:44 Mon Oct 20 2003 root@eradicator) /usr/src/linux $ grep CONFIG_IDEDMA_PCI_AUTO .config CONFIG_IDEDMA_PCI_AUTO=y (13:08:17 Mon Oct 20 2003 root@terminator) /usr/src/linux $ grep CONFIG_IDEDMA_PCI_AUTO .config CONFIG_IDEDMA_PCI_AUTO=y The drives start in DMA mode by default on the other controllers, but not the PDC20267 or PDC20269. The PDC20267 is in a different system than the PDC20269. On Monday October 20 2003 10:49, Bartlomiej Zolnierkiewicz wrote: > On Monday 20 of October 2003 01:03, Jeremy Huddleston wrote: > > I have a lot of md output to the kernel logger when my system starts, so > > the IDE initialization output doesn't appear when I do my initial > > 'dmesg'... How do I increase the buffer size? I tried doing 'dmesg -s > > ', but I still can't get the output... > > /var/log/messages ? > > > Anyhow, I was trying the -X70 option to force udma6 on my 2.4.20 kernel > > (which was with the old driver), but when I do the force on 2.4.22-ac4, > > it works now, so thanks. > > > > What is the reason that the drive starts in pio mode when it is attatched > > to the promise controller, but it starts in DMA mode when it is on the > > other controller? > > Was the other controller on the same machine and with the same kernel? > If not check if you have CONFIG_IDEDMA_PCI_AUTO set to Y in your .config > file. Also please try to extract boot log... > > thanks, > --bartlomiej > > > Thanks again, > > Jeremy > > > > On Sunday October 19 2003 13:45, you wrote: > > > Here is a hint. UDMA2 is a max if you don't use 80-wire cable. > > > If you can't force mode by "hdparm -X" driver might be misdetecting > > > cable type. Please change "#define PDC202_DEBUG_CABLE 0" > > > to "#define PDC202_DEBUG_CABLE 1" in pdc202_new.c, recompile > > > and retest. It should give some useful information. > > > > > > thanks, > > > --bartlomiej > > > > > > On Sunday 19 of October 2003 21:01, Willy Gardiol wrote: > > > > Well, in this case, i cant help more. > > > > Lets hope some "guru" from this list will give an hint ! > > > > > > > > Alle Sunday 19 October 2003 21:00, Jeremy Huddleston ha scritto: > > > > > It's still in mode2 after the explicit -X69 or -X70. I have a > > > > > 20267 on my other system, and it goes to mode5 with just the -d1... > > > > > > > > > > 11:53:45 Sun Oct 19 2003 root@eradicator) > > > > > ~ $ hdparm -d1 -X69 /dev/hde > > > > > > > > > > /dev/hde: > > > > > setting using_dma to 1 (on) > > > > > setting xfermode to 69 (UltraDMA mode5) > > > > > using_dma = 1 (on) > > > > > > > > > > (11:53:51 Sun Oct 19 2003 root@eradicator) > > > > > ~ $ hdparm -i /dev/hde > > > > > > > > > > /dev/hde: > > > > > > > > > > Model=Maxtor 6Y120P0, FwRev=YAR41VW0, SerialNo=Y40BMM7E > > > > > Config={ Fixed } > > > > > RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=57 > > > > > BuffType=DualPortCache, BuffSize=7936kB, MaxMultSect=16, > > > > > MultSect=off CurCHS=16383/16/63, CurSects=16514064, LBA=yes, > > > > > LBAsects=240121728 IORDY=on/off, tPIO={min:120,w/IORDY:120}, > > > > > tDMA={min:120,rec:120} PIO modes: pio0 pio1 pio2 pio3 pio4 > > > > > DMA modes: mdma0 mdma1 mdma2 > > > > > UDMA modes: udma0 udma1 *udma2 udma3 udma4 udma5 udma6 > > > > > AdvancedPM=yes: disabled (255) WriteCache=enabled > > > > > Drive conforms to: (null): > > > > > > > > > > * signifies the current active mode > > > > > > > > > > (11:55:38 Sun Oct 19 2003 root@eradicator) > > > > > ~ $ hdparm -d1 -X70 /dev/hde > > > > > > > > > > /dev/hde: > > > > > setting using_dma to 1 (on) > > > > > setting xfermode to 70 (UltraDMA mode6) > > > > > using_dma = 1 (on) > > > > > > > > > > (11:55:44 Sun Oct 19 2003 root@eradicator) > > > > > ~ $ hdparm -i /dev/hde > > > > > > > > > > /dev/hde: > > > > > > > > > > Model=Maxtor 6Y120P0, FwRev=YAR41VW0, SerialNo=Y40BMM7E > > > > > Config={ Fixed } > > > > > RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=57 > > > > > BuffType=DualPortCache, BuffSize=7936kB, MaxMultSect=16, > > > > > MultSect=off CurCHS=16383/16/63, CurSects=16514064, LBA=yes, > > > > > LBAsects=240121728 IORDY=on/off, tPIO={min:120,w/IORDY:120}, > > > > > tDMA={min:120,rec:120} PIO modes: pio0 pio1 pio2 pio3 pio4 > > > > > DMA modes: mdma0 mdma1 mdma2 > > > > > UDMA modes: udma0 udma1 *udma2 udma3 udma4 udma5 udma6 > > > > > AdvancedPM=yes: disabled (255) WriteCache=enabled > > > > > Drive conforms to: (null): > > > > > > > > > > * signifies the current active mode > > > > > > > > > > On Sunday October 19 2003 02:39, Willy Gardiol wrote: > > > > > > Ok, you should also set explicitly the dma mode: > > > > > > hdparm -d1 -X69 > > > > > > (69=udma100...) > > > > > > > > > > > > Alle Sunday 19 October 2003 12:11, Jeremy Huddleston ha scritto: > > > > > > > I have done this, and the result is that it still doesn't use > > > > > > > the drive to its full potential: > > > > > > > > > > > > > > (03:07:19 Sun Oct 19 2003 root@eradicator) > > > > > > > ~ $ hdparm -d1 /dev/hde > > > > > > > > > > > > > > /dev/hde: > > > > > > > setting using_dma to 1 (on) > > > > > > > using_dma = 1 (on) > > > > > > > > > > > > > > (03:07:25 Sun Oct 19 2003 root@eradicator) > > > > > > > ~ $ hdparm -i /dev/hde > > > > > > > > > > > > > > /dev/hde: > > > > > > > > > > > > > > Model=Maxtor 6Y120P0, FwRev=YAR41VW0, SerialNo=Y40BMM7E > > > > > > > Config={ Fixed } > > > > > > > RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=57 > > > > > > > BuffType=DualPortCache, BuffSize=7936kB, MaxMultSect=16, > > > > > > > MultSect=off CurCHS=16383/16/63, CurSects=16514064, LBA=yes, > > > > > > > LBAsects=240121728 IORDY=on/off, tPIO={min:120,w/IORDY:120}, > > > > > > > tDMA={min:120,rec:120} PIO modes: pio0 pio1 pio2 pio3 pio4 > > > > > > > DMA modes: mdma0 mdma1 mdma2 > > > > > > > UDMA modes: udma0 udma1 *udma2 udma3 udma4 udma5 udma6 > > > > > > > AdvancedPM=yes: disabled (255) WriteCache=enabled > > > > > > > Drive conforms to: (null): > > > > > > > > > > > > > > * signifies the current active mode > > > > > > > > > > > > > > The PDC-Specific drivers are all compiled into the kernel > > > > > > > > > > > > > > On Sunday October 19 2003 00:08, Willy Gardiol wrote: > > > > > > > > Have you tried to enable uDMA using hdpard after boot? > > > > > > > > hdparm -d1 /dev/hdX > > > > > > > > > > > > > > > > There are also some options in kernel configurations which > > > > > > > > addresses when you have more than one promise comtrollers, > > > > > > > > have you tried them? Last, also in the ide section of kernel > > > > > > > > config you can find an option which should enable by default > > > > > > > > UDMA, have you enabled it? > > > > > > > > > > > > > > > > bye! > > > > > > > > > > > > > > > > Alle Sunday 19 October 2003 03:19, Jeremy Huddleston ha scritto: > > > > > > > > > Hi, > > > > > > > > > > > > > > > > > > I've been using two Promise PDC20269 cards for almost a > > > > > > > > > year now in my system. The cards are U133, but when the > > > > > > > > > kernel boots up, I see the message that all the drives are > > > > > > > > > operating in pio mode. All the drives are U133 drives, and > > > > > > > > > they are shown as operating in DMA mode when I put them on > > > > > > > > > another controller. > > > > > > > > > > > > > > > > > > I've noticed this deficiency with vanilla linux-2.4.19 > > > > > > > > > through linux-2.4.21, linux-2.4.22-ac4 as well as > > > > > > > > > 2.6.0-test6-mm4... > > > > > > > > > > > > > > > > > > Is there a patch to the 2.4 kernels availible to enable > > > > > > > > > this support? Is this problem being addressed? Is there > > > > > > > > > any way I can help out? > > > > > > > > > > > > > > > > > > Thanks, > > > > > > > > > Jeremy -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.3 (GNU/Linux) iD8DBQE/lEJmgKpk8srJOlIRAtlgAJwJRCxe/ZcJLRo6g3vVXd467a9TswCcDPaj ZFZ/k7aA+2P2XDbp0hUOxe0= =Gdox -----END PGP SIGNATURE-----