Re: EXT2-fs error

From: John Heil (
Date: Sun Feb 25 2001 - 08:53:51 EST

On Sun, 25 Feb 2001, A E Lawrence wrote:
> A E Lawrence wrote:
> >
> > A E Lawrence wrote:
> > >
> > > Alan Cox wrote:
> > > >
> > > > > I have seen similar problems on stock 2.4.2 a machine which has not run
> > > > > 2.4.1.
> > > >
> > > > What disk controllers ? We really need that sort of info in order to see the
> > > > pattern in the odd reports of corruption we get
> >
> > Problems have just started to show up under 2.2.18, so it is likely that
> > the hardware has become flakey. Bit of a coincidence, unless it is a
> > side effect of upgrading one of the packages for 2.4.2 :-( or a damaged
> > library.
> >
> > So you had better discount this report. Apologies.
> Now investigated: the hardware has not changed. Rather the corruption
> under 2.2.18 only happens when hdparm -d1 is executed. I guess that is
> well reported, but I had forgotten if I ever knew :-(
> In contrast 2.4.2 corruptions happen whether dma is explicitly turned on
> by hdparm or not.
> [IDE interface: VIA Technologies, Inc. VT82C586 IDE [Apollo] (rev 10)]

The corruption I got seemed to be non-chipset related.

The 2.4.1 corruption I experienced spanned 2.2.14 (RH 6.2)...

The 2.4.1 IDE support was in an ALI 1535 south bridge in a Crusoe based
embedded system...

00:0f.0 IDE interface: Acer Laboratories Inc. [ALi] M5229 IDE (rev c3)
(prog-if b4)
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop-
ParErr- Stepping- SERR- FastB2B-
        Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
        Latency: 64 (500ns min, 1000ns max)
        Interrupt: pin A routed to IRQ 0
        Region 0: I/O ports at 01f0 [size=16]
        Region 1: I/O ports at 03f4
        Region 2: I/O ports at 0170 [size=8]
        Region 3: I/O ports at 0374
        Region 4: I/O ports at 1400 [size=16]
        Capabilities: [60] Power Management version 2
                Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA
                Status: D0 PME-Enable- DSel=0 DScale=0 PME-

Disk was...


 Model=HITACHI_DK23AA-60, FwRev=00XEA0A0, SerialNo=F77166
 Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
 RawCHS=12416/15/63, TrkSize=36477, SectSize=579, ECCbytes=4
 BuffType=3(DualPortCache), BuffSize=512kB, MaxMultSect=16, MultSect=off
 DblWordIO=no, OldPIO=2, DMA=yes, OldDMA=2
 CurCHS=12416/15/63, CurSects=142606515, LBA=yes, LBAsects=11733120
 tDMA={min:120,rec:120}, DMA modes: sword0 sword1 sword2 mword0 mword1
 IORDY=yes, tPIO={min:400,w/IORDY:120}, PIO modes: mode3 mode4
 UDMA modes: mode0 mode1 *mode2 mode3 mode4

When the above system couldn't adequately repair itself,
I fsck'd it on the following....

2.2.14-5.0 RH 6.2 desktop PII w PIIX4 IDE...
00:07.1 IDE interface: Intel Corporation 82371AB PIIX4 IDE (rev 01)
(prog-if 80 [Master])
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop-
ParErr- Stepping- SERR- FastB2B-
        Status: Cap- 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
        Latency: 64 set
        Region 4: I/O ports at 1460

Initially, the 2.2.14 RH 6.2 had e2fsprogs 1.18, but I ugraded it to
e2fsck 1.19, thinking that might be the problem but it did _not_ help.

The 2 systems, both with e2fsck 1.19, would do fsck repairs
after the other had touched it. It seemed impossible to keep it
error free.

This cross-system disk transfering was error free in 2.4.0-test10-pre3.

2.4.1-final was the first post 2.4.0-final kernel after t10p3, that
saw any significant activity.

I've since upgraded the RH to 2.2.18 and the embedded to -ac19.

So far, no more problems but I still haven't used the ac19's disk
on 2.2.18 very much yet but it will have to happen soon...
I am a bit gun-shy of it still.

I didn't/don't have much time to spend on this and I'm hoping ac19
is bug free is this regard... but I'm interested in how this
turns out.

My guess was that I may have hit the bug Russel King fixed for his ARM
system. IRC, that fix hit the -ac tree from Linus's 2.4.2-pre2.

John Heil
South Coast Software
Custom systems software for UNIX and IBM MVS mainframes

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

This archive was generated by hypermail 2b29 : Wed Feb 28 2001 - 21:00:10 EST