[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: kernel-image packages for alpha in need of testing



Hello Steve,
On Wed, Jun 16, 2004 at 04:27:39PM -0500, Steve Langasek wrote:
> On Mon, Jun 14, 2004 at 09:16:18AM +0200, Helge Kreutzmann wrote:
> 
> > * Please turn IDE-DMA off by default, at least for installs. On my alpha
> >   I see tons of scary messages, and trying to use the IDE drives fails 
> >   (on installed kernels), c.f. #241185 for details. Since for many alphas
> >   IDE-support is very poor (hardware-wise), I think keeping it off during
> >   installs is wise. Maybe the installer could ask/guess later on, if DMA
> >   should be attempted. Or it should always add "ide=nodma" in /etc/aboot.conf
> >   and an installation note should inform the user that (s)he can remove it.
> 
> I've been asked by the Debian kernel maintainers to get more information
> about the specific IDE controller you have (PCI ID, etc.), so that it
> can be appropriately blacklisted in the kernel images instead of being
> disabled with an overbroad boot-time option in the installer.
> 
> FWIW, DMA also doesn't work on my IDE controller, but the installer
> (kernel) figures this out on its own.

The releveant part of the boot log with DMA enabled is:
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
ide: late registration of driver.
CMD646: IDE controller at PCI slot 00:0b.0
CMD646: chipset revision 1
CMD646: not 100% native mode: will probe irqs later
CMD646: chipset revision 0x01, MultiWord DMA Limited, IRQ workaround enabled
    ide0: BM-DMA at 0x8800-0x8807, BIOS settings: hda:pio, hdb:pio
    ide1: BM-DMA at 0x8808-0x880f, BIOS settings: hdc:pio, hdd:pio
hda: IBM-DTTA-351010, ATA DISK drive
hdb: IC35L120AVVA07-0, ATA DISK drive
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
blk: queue fffffffc00249918, no I/O memory limit
blk: queue fffffffc00249af8, no I/O memory limit
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
hda: attached ide-disk driver.
hda: 19807200 sectors (10141 MB) w/466KiB Cache, CHS=19650/16/63, (U)DMA
hdb: attached ide-disk driver.
hdb: 241254720 sectors (123522 MB) w/1863KiB Cache, CHS=239340/16/63, (U)DMA
Partition check:
 /dev/ide/host0/bus0/target0/lun0:<4>hda: dma_timer_expiry: dma status == 0x62
hda: error waiting for DMA
hda: dma timeout retry: status=0x58 { DriveReady SeekComplete DataRequest }

blk: queue fffffffc00249918, no I/O memory limit
 p1 p2 p3 p4 p5 p6 p7
 /dev/ide/host0/bus0/target1/lun0:<4>hdb: dma_timer_expiry: dma status == 0x62
hdb: error waiting for DMA
hdb: dma timeout retry: status=0x58 { DriveReady SeekComplete DataRequest }

blk: queue fffffffc00249af8, no I/O memory limit
 p1 p2 p3
Journalled Block Device driver loaded
hda: dma_timer_expiry: dma status == 0x62
hda: error waiting for DMA
hda: dma timeout retry: status=0x5a { DriveReady SeekComplete DataRequest Index }

blk: queue fffffffc00249918, no I/O memory limit
hda: dma_timer_expiry: dma status == 0x62
hda: error waiting for DMA
hda: dma timeout retry: status=0x58 { DriveReady SeekComplete DataRequest }

hda: dma_timer_expiry: dma status == 0x62
hda: error waiting for DMA
hda: dma timeout retry: status=0x58 { DriveReady SeekComplete DataRequest }

I have provided lots of system information in #241185, please tell me which
exact information you need.

> > * Any NIC I chose results in a printed "Segmentation fault". I did not expect
> >   it to actually find any NIC, since my old one was doing strange things
> >   and a 3c509B I got is not recognized in my alpha (neither Jumpstart (2.2
> >   based), nor 2.4.21-pre1, nor 2.4.26). After the segfault, the hardware
> >   detection bar is shown briefly, and I am again at the choosing NIC
> >   selection.
> 
> Hmm, not good. :/   Never seen this before... does this also happen with
> the most recently-posted mini-ISO?

I have not tried the one you just informed me about, but the one from
sunday: yes.

> > * Detect NIC-Hardware is not available in german (or switching did not work).
> >   I currently do some test installs on my ibook (ppc) as well, where all
> >   screens are in german. If you can point me to the proper .po-files, I can
> >   translate them.
> 
> The package is the same for ppc and alpha (though I don't know off-hand
> which package this is).  I'm not sure why you would get translations
> with one, but not the other.

The image from sunday was translated the first round, but not the
second. Maybe I simply overlooked it in my first attempt. Will check
it carefully with the new ISO.

> > * I see that the list of modules is quite limited. Since no aic7xxx is 
> >   available on this ISO, I was unable to test if the aic7xxx-fix has gone
> >   in. I think aic7xxx should be included (if the fix is available, of course).
> >   These missing modules were also seen in the log (by hw-detect).
> >   (Update when sending: Was the module left out on purpose?)
> 
> If you point at unstable from this installer instead of testing
> (requires using DEBCONF_PRIORITY=medium or lower, IIRC), additional
> modules for 2.4.26 will be available.  These are from the official
> kernel image in the archive, which *should* match the module ABI of the
> PCI-fixed image.

Ah, ok. I only checked on the ISO, as I only have a dial-up (ppp) at
home, and some SCSI adapters are in /lib/modules ...

> > * Due to my NIC-troubles I tried out Firewire (I got an USB/Firewire card
> >   about 1.5 years ago). In 2.4.21 I had troubles using ethernet over firewire,
> >   but in 2.4.26 it worked flawlessly. Please consider adding eth1394 (and
> >   ohci1394, which needs to be loaded as well) in the NIC-Dialog. After insmodding
> >   these modules, the interface behaves exactly like a NIC. As I read, though,
> >   only connections via Linux-boxes are supported. This could be noted in the
> >   install logs. I think it would be cool having Linux be able to install
> >   over Firewire, and given that the support is so trivial to add.
> > 0000:01:05.0 FireWire (IEEE 1394): Texas Instruments TSB43AB22/A IEEE-1394a-2000 Controller (PHY/Link) (prog-if 10 [OHCI])
> >         Flags: bus master, medium devsel, latency 32, IRQ 24
> >         Memory at 0000000009006000 (32-bit, non-prefetchable) [size=2K]
> >         Memory at 0000000009000000 (32-bit, non-prefetchable) [size=16K]
> >         Capabilities: [44] Power Management version 2
> 
> Could you file a bug against debian-installer requesting this, please?
> I'd like to discuss with the maintainers of the linux-kernel-di packages
> for other architectures before making such a change.

Will do.

Greetings

         Helge

-- 
Helge Kreutzmann, Dipl.-Phys.               Helge.Kreutzmann@itp.uni-hannover.de
  gpg signed mail preferred    gpg-key: finger kreutzm@zibal.itp.uni-hannover.de
    64bit GNU powered                  http://www.itp.uni-hannover.de/~kreutzm
       Help keep free software "libre": http://www.freepatents.org/

Attachment: pgp6rXN25paAK.pgp
Description: PGP signature


Reply to: