On Thu, Jun 06, 2002 at 05:07:07PM -0600, Richard Gooch wrote:
> A. Guy Called Tyketto writes:
> > On Thu, Jun 06, 2002 at 11:38:54AM -0400, Pavel Roskin wrote:
> > > Hello!
> > >
> > > Maybe you forwarded a wrong message? I don't understand how a failure of
> > > the kernel to mount the root filesystem is related to devfs. The
> > > workaround is probably related, but shouldn't we care about the proper
> > > fix?
> > >
> > > By the way, is the root device specified correctly in the bootblock
> > > (using
> > > rdev)? Can you omit the "root" parameter? What happens with other
> > > bootloaders (lilo, syslinux, loadlin)? What do you see in /proc/cmdline?
> >
> > I had the same problem as the person who had originally posted this
> > to
> > the LKML. I have /dev/hda1 set up to be the root partition, and with
> > booting
> > 2.5.19 and higher (2.5.20{dj1,2,3}) I got that /dev/hda1 was not a valid
> > partition, and dropped me into single user mode to fix the problem, then
> > reboot.
> >
> > with kernels < 2.5.19, devfs maps /dev/hda1 to
> > /dev/ide/host0/bus0/target0/lun0/part1. with kernels 2.5.19 and higher,
> > that
> > mapping changes to /dev/ata/host0/bus0/target0/lun0/part1. The difference
> > is
> > 'ide' and 'ata'.
> >
> > > > This feature was introduced in 2.5.19
> > >
> > > Feature? Shouldn't we call it a bug? Maybe I don't understand what you
> > > mean.
> >
> > It culd be a feature; wanting to change or accent the fact
> > that it's on an ATA bus, or that it's an ATA/IDE drive. We don't
> > know which it is.
>
> It's a stupidity. /dev/ide is the published interface and shouldn't be
> changed. Someone want to code up a patch to correct this?
>
> Regards,
>
> Richard....
Heya Richard..
Found something that might be of interest to this problem.
I'm currently running 2.5.21 with Martin Dalecki's IDE 86 patch. it
boots successfully, with this coming from dmesg at boot time:
--snip--
block: 512 slots per queue, batch=32
ATA/ATAPI device driver v7.0.0
ATA: PCI bus speed 33.3MHz
ATA: VIA Technologies, Inc. Bus Master IDE, PCI slot 00:07.1
ATA: chipset rev.: 6
ATA: non-legacy mode: IRQ probe delayed
VP_IDE: VIA vt82c686b (rev 40) ATA UDMA100 controller on PCI 00:07.1
ide0: BM-DMA at 0xd000-0xd007, BIOS settings: hda:DMA, hdb:DMA
ide1: BM-DMA at 0xd008-0xd00f, BIOS settings: hdc:DMA, hdd:pio
hda: Maxtor 91531U3, DISK drive
hdb: WDC WD200AB-00BVA0, DISK drive
hdc: CD-W54E, ATAPI CD/DVD-ROM drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide1 at 0x170-0x177,0x376 on irq 15
hda: 30015216 sectors w/512KiB Cache, CHS=29777/16/63, UDMA(66)
/dev/ata/host0/bus0/target0/lun0: [PTBL] [1868/255/63] p1 p2 p3 p4 < p5 p6 p7
p8 >
hdb: 39102336 sectors w/2048KiB Cache, CHS=38792/16/63, UDMA(100)
/dev/ata/host0/bus0/target1/lun0: [PTBL] [2434/255/63] p1 p2 p3 p4 < p5 p6 p7
p8 p9 p10 p11 >
--snip--
Doing a search for the ATA/ATAPI driver (drivers/ide/main.c) and
searching through the code, I come up with this:
--snip--
linux/drivers/ide/main.c, line 1286:
/*
* This is the global initialization entry point.
*/
static int __init ata_module_init(void)
{
printk(KERN_INFO "ATA/ATAPI device driver v" VERSION "\n");
ide_devfs_handle = devfs_mk_dir(NULL, "ata", NULL);
devfs_mk_symlink(NULL, "ide", DEVFS_FL_DEFAULT, "ata", NULL, NULL);
--snip--
this looks like it would be the problem causing /dev/ide/* to be
renamed to /dev/ata/*. Your thoughts?
BL.
|