Re: 2.5.19 (and earlier) IDE (+EXT3+???) bugs

[Martin Dalecki]

David Brownell wrote:
> I'm trying to use a slightly elderly laptop for some testing
> on the 2.5 kernels.  (As an ultralight, it's got some hardware
> that tweaks some interesting USB/APM codepaths that don't
> otherwise show up.)  It's run Linux (mostly) since I got it,
> and I can install and use RH 7.3 on it, no troubles.
> 
> But it doesn't seem to want to run any recent 2.5 kernels.
> I first tried with 2.5.15, and kernels up to and including
> the latest (2.5.19 as I write) have the same overall failure
> mode ... which does not happen with any of the 2.4 kernels
> I've tried.  (No recent ones other than the RH 7.3 code,
> but many earlier ones.)  Basically, I see:
> 
> - kernel loads OK ... I attach "dmesg" output.
> - runs init, which runs init scripts.
> - everything's fine, disk fscks as right, UNTIL ...
> - ...it  blows up when remounting the root filesystem r/w
>   * Takes a *long* time, if it even succeeds
>   * Most of that time is evidently used to scribble over
>     as much of the disk as it can!
>   * If I powerdown the system very quickly, "fsck" can mostly
>     recover.  If not, then both root and /boot get trashed.
> - Next step is to re-install the OS again.
> 
> As a stock RH 7.3 install, this root filesystem uses ext3.
> 
> I was able to boot with "init=/bin/sh" and do some basic
> testing with a read-only root FS.  Reading files works ok,
> "hdparm -I" gives the same info it did under the RH7.3 kernel,
> and I can use DD to read and write to the disk.  (USB works OK;
> I can bring it up by hand using the "ohci-hcd" driver, which
> is how I could transfer the dmesg info off this system.)
> 
> So far the only really suggestive thing I've come up with is
> that if I do much disk I/O, I start to see "hda: lost interrupt"
> and the operation seems to become timeout-driven.  I first
> noticed that with DD, but then "fsck" of the root FS (5+MBytes)
> turned up the same failure.  (The fsck took so much time I had
> to kill it; running on 2.4, it quickly reported no problems.)
> 
> Does anyone know what might be going on here?  Or better yet,
> have a fix to whatever it is that's wrong? :)  Seems to me there
> is a clear IDE problem: lost interrupts were not an issue on
> the 2.4 kernels.  Whether fixing that would make that "scribble
> on the disk" problem go away, I couldn't say.
> 
> - Dave
> 
> p.s. Hardware is a Toshiba Portege 3020ct, pci host bridge
>      is a "Toshiba America Info Systems 601 (rev a2)"
>      according to lspci.
> 
> 
> Linux version 2.5.19 (root@neon) (gcc version 2.96 20000731 (Red Hat 
> Linux 7.3 2.96-110)) #1 Thu May 30 19:40:52 PDT 2002
> Video mode to be used for restore is f03
> BIOS-provided physical RAM map:
>  BIOS-e820: 0000000000000000 - 000000000009fc00 (usable)
>  BIOS-e820: 000000000009fc00 - 00000000000a0000 (reserved)
>  BIOS-e820: 00000000000f0000 - 0000000000100000 (reserved)
>  BIOS-e820: 0000000000100000 - 0000000004010000 (usable)
>  BIOS-e820: 0000000004010000 - 0000000004020000 (ACPI data)
>  BIOS-e820: 0000000004020000 - 0000000004040000 (reserved)
>  BIOS-e820: 00000000fef80000 - 00000000ff000000 (reserved)
>  BIOS-e820: 00000000ffee0000 - 00000000ffee6e00 (reserved)
>  BIOS-e820: 00000000ffee6e00 - 00000000ffee7000 (ACPI NVS)
>  BIOS-e820: 00000000ffee7000 - 00000000ffef0000 (reserved)
>  BIOS-e820: 00000000ffff0000 - 0000000100000000 (reserved)
> 64MB LOWMEM available.
> On node 0 totalpages: 16400
> zone(0): 4096 pages.
> zone(1): 12304 pages.
> zone(2): 0 pages.
> Kernel command line: init=/bin/sh ro root=/dev/hda2 vga=0x0f03
> Initializing CPU#0
> Detected 299.947 MHz processor.
> Console: colour VGA+ 80x28
> Calibrating delay loop... 598.01 BogoMIPS
> Memory: 62912k/65600k available (1004k kernel code, 2300k reserved, 244k 
> data, 216k init, 0k highmem)
> Dentry-cache hash table entries: 16384 (order: 5, 131072 bytes)
> Inode-cache hash table entries: 8192 (order: 4, 65536 bytes)
> Mount-cache hash table entries: 512 (order: 0, 4096 bytes)
> CPU: Before vendor init, caps: 008001bf 00000000 00000000, vendor = 0
> Intel Pentium with F0 0F bug - workaround enabled.
> CPU: After vendor init, caps: 008001bf 00000000 00000000 00000000
> CPU:     After generic, caps: 008001bf 00000000 00000000 00000000
> CPU:             Common caps: 008001bf 00000000 00000000 00000000
> CPU: Intel Mobile Pentium MMX stepping 02
> Checking 'hlt' instruction... OK.
> POSIX conformance testing by UNIFIX
> Linux NET4.0 for Linux 2.4
> Based upon Swansea University Computer Society NET3.039
> Initializing RT netlink socket
> PCI: PCI BIOS revision 2.10 entry at 0xfd84f, last bus=21
> PCI: Using configuration type 1
> isapnp: Scanning for PnP cards...
> isapnp: No Plug & Play device found
> PnPBIOS: Found PnP BIOS installation structure at 0xc00f9020
> PnPBIOS: PnP BIOS version 1.0, entry 0xf0000:0x9563, dseg 0x0
> PnPBIOS: 17 nodes reported by PnP BIOS; 17 recorded by driver
> PnPBIOS: PNP0c02: ioport range 0x1882-0x1885 has been reserved
> PCI: Probing PCI hardware
> PCI: Probing PCI hardware (bus 00)
> apm: BIOS version 1.2 Flags 0x02 (Driver version 1.16)
> Starting kswapd
> BIO: pool of 256 setup, 14Kb (56 bytes/bio)
> biovec: init pool 0, 1 entries, 12 bytes
> biovec: init pool 1, 4 entries, 48 bytes
> biovec: init pool 2, 16 entries, 192 bytes
> biovec: init pool 3, 64 entries, 768 bytes
> biovec: init pool 4, 128 entries, 1536 bytes
> biovec: init pool 5, 256 entries, 3072 bytes
> Journalled Block Device driver loaded
> pty: 512 Unix98 ptys configured
> Real Time Clock Driver v1.11
> block: 256 slots per queue, batch=32
> Floppy drive(s): fd0 is 1.44M
> FDC 0 is an 8272A
> ATA/ATAPI device driver v7.0.0
> ATA: PCI bus speed 33.3MHz
> hda: TOSHIBA MK6411MAT, ATA DISK drive
> ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
>  hda: 12685680 sectors, CHS=13424/15/63
>  hda: [PTBL] [789/255/63] hda1 hda2 hda3
> mice: PS/2 mouse device common for all mice
> NET4: Linux TCP/IP 1.0 for NET4.0
> IP Protocols: ICMP, UDP, TCP
> IP: routing cache hash table of 512 buckets, 4Kbytes
> TCP: Hash tables configured (established 4096 bind 4096)
> NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
> kjournald starting.  Commit interval 5 seconds
> EXT3-fs: mounted filesystem with ordered data mode.
> VFS: Mounted root (ext3 filesystem) readonly.
> Freeing unused kernel memory: 216k freed

Thoese are indeed most propably symptoms of the multi mode write
probles, I have and I know about :-(.
The the drives are apparently comming up in PIO mode.
Well slowly it's time to do something about it.