2003-03-21 19:19:17

by Richard B. Johnson

[permalink] [raw]
Subject: Ded-Fat 8.0 and ext3


Hello all. Greetings from fix-up land.

I installed RedHat 8.0 of an existing system. By default it
makes ext3 file-systems with journaling enabled.

That distribution used Linux-2.4.18-14.
I copied /boot/config-2.4.18-14 to /usr/src/linux-2.4.18-15/.config
as a ".config" file and executed:

`make oldconfig`

It asked me some more questions so I fear that /boot/config-2.4.18-14
was not used in the distribution configuration for linux-2.4.18-14,
but I continued anyway.

Then I executed `make dep`. This failed to find the math emulation
library, but continued anyway.

Then I executed make bzImage

The result was screens and screens full of errors, to lengthy to
describe. When it failed, it was obvious that this version,
supplied by our respected, and soon to be unloved, vendor had
never, ever been built from the supplied sources.

This has always happened to me when attempting to build whatever
Red Hat distributed. Undaunted, I grabbed a linux-2.4.20 tarball
and built a system using the same .config file. I also built the
modules, but many of them had errors. Not to worry, I learned a
long time ago to execite `make -i` when building modules. I got
the modules I needed.

The problem is:
It took 14:37:xx minutes to build a kernel that took 00:12:xx to
|___ 14 hours, 37 minutes |___ 12 min.
build on an identical system. The only difference is the "fast"
system uses ext2, and the slooooooooow system uses ext3. This
14 hour stuff is reminisicent of the days of x386 and ST-506
drives!

I just timed reading the whole physical drive and I can do that
at 24 megabytes per second. If I read a 24 megabyte file, however
it takes over 1 minute, about 400k per second! What is going
on with ext3 file-systems or is something else stealing CPU
cycles?

It there a way to convert to ext2 without having to start from
scratch again? I know you can mount it as an ext2, would that
help?

When trying to create this report, I executed `script`. Something
bad has happened because I get this:

The simple command `cat /proc/cpuinfo` results in:

#
#
# c
# c
# ca
# ca
# cat
# cat
# cat
# cat
# cat /
# cat /
# cat /p
# cat /p
# cat /pr
# cat /pr
# cat /pro
# cat /pro
# cat /proc
# cat /proc
# cat /proc/
# cat /proc/
# cat /proc/c
# cat /proc/c
# cat /proc/cp
# cat /proc/cp
# cat /proc/cpu
# cat /proc/cpu
# cat /proc/cpui
# cat /proc/cpui
# cat /proc/cpuin
# cat /proc/cpuin
# cat /proc/cpuinf
# cat /proc/cpuinf
# cat /proc/cpuinfo
# cat /proc/cpuinfo

So whatever `script` is interfacing with is broken in Red Hat 8.0 also.
I wish they actually tested this stuff.

Script started on Fri Mar 21 13:54:23 2003

#
#
# c
# c
# ca
# ca
# cat
# cat
# cat
# cat
# cat /
# cat /
# cat /p
# cat /p
# cat /pr
# cat /pr
# cat /pro
# cat /pro
# cat /proc
# cat /proc
# cat /proc/
# cat /proc/
# cat /proc/c
# cat /proc/c
# cat /proc/cp
# cat /proc/cp
# cat /proc/cpu
# cat /proc/cpu
# cat /proc/cpui
# cat /proc/cpui
# cat /proc/cpuin
# cat /proc/cpuin
# cat /proc/cpuinf
# cat /proc/cpuinf
# cat /proc/cpuinfo
# cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 5
model : 2
model name : Pentium 75 - 200
stepping : 12
cpu MHz : 199.437
fdiv_bug : no
hlt_bug : no
f00f_bug : yes
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr mce cx8
bogomips : 398.13


#
#
# l
# l
# ls
# ls
# lsp
# lsp
# lspc
# lspc
# lspci
# lspci
00:00.0 Host bridge: Intel Corp. 430VX - 82437VX TVX [Triton VX] (rev 02)
00:07.0 ISA bridge: Intel Corp. 82371SB PIIX3 ISA [Natoma/Triton II] (rev 01)
00:07.1 IDE interface: Intel Corp. 82371SB PIIX3 IDE [Natoma/Triton II]
00:0d.0 SCSI storage controller: BusLogic BT-946C (BA80C30) [MultiMaster 10] (rev 08)
00:0e.0 VGA compatible controller: S3 Inc. 86c325 [ViRGE] (rev 06)
00:10.0 Ethernet controller: 3Com Corporation 3c905 100BaseTX [Boomerang]

#
#
# c
# c
# ca
# ca
# cat
# cat
# cat
# cat
# cat /
# cat /
# cat /p
# cat /p
# cat /pr
# cat /pr
# cat /pro
# cat /pro
# cat /proc
# cat /proc
# cat /proc/
# cat /proc/
# cat /proc/m
# cat /proc/m
# cat /proc/me
# cat /proc/me
# cat /proc/mem
# cat /proc/mem
# cat /proc/memi
# cat /proc/memi
# cat /proc/memin
# cat /proc/memin
# cat /proc/meminf
# cat /proc/meminf
# cat /proc/meminfo
# cat /proc/meminfo
total: used: free: shared: buffers: cached:
Mem: 64098304 54976512 9121792 0 5009408 29765632
Swap: 148013056 4980736 143032320
MemTotal: 62596 kB
MemFree: 8908 kB
MemShared: 0 kB
Buffers: 4892 kB
Cached: 28468 kB
SwapCached: 600 kB
Active: 20828 kB
Inactive: 15084 kB
HighTotal: 0 kB
HighFree: 0 kB
LowTotal: 62596 kB
LowFree: 8908 kB
SwapTotal: 144544 kB
SwapFree: 139680 kB

#
#
# c
# c
# ca
# ca
# cat
# cat
# cat
# cat
# cat /
# cat /
# cat /p
# cat /p
# cat /pr
# cat /pr
# cat /pro
# cat /pro
# cat /proc
# cat /proc
# cat /proc/
# cat /proc/
# cat /proc/m
# cat /proc/m
# cat /proc/mo
# cat /proc/mo
# cat /proc/mod
# cat /proc/mod
# cat /proc/modu
# cat /proc/modu
# cat /proc/modul
# cat /proc/modul
# cat /proc/module
# cat /proc/module
# cat /proc/modules
# cat /proc/modules
nfsd 73360 8 (autoclean)
lockd 54064 1 (autoclean) [nfsd]
sunrpc 77308 1 (autoclean) [nfsd lockd]
autofs 12244 0 (autoclean) (unused)
3c59x 29360 1
iptable_filter 2316 0 (autoclean) (unused)
ip_tables 14488 1 [iptable_filter]
ext3 64192 5
jbd 47828 5 [ext3]
loop 10936 0 (unused)
sr_mod 16824 0 (unused)
cdrom 31008 0 [sr_mod]
BusLogic 93724 7
sd_mod 12652 14
scsi_mod 102836 3 [sr_mod BusLogic sd_mod]

#
#
# c
# c
# ca
# ca
# cat
# cat
# cat
# cat
# cat /
# cat /
# cat /p
# cat /p
# cat /pr
# cat /pr
# cat /pro
# cat /pro
# cat /proc
# cat /proc
# cat /proc/
# cat /proc/
# cat /proc/s
# cat /proc/s
# cat /proc/sc
# cat /proc/sc
# cat /proc/scs
# cat /proc/scs
# cat /proc/scsi
# cat /proc/scsi
# cat /proc/scsi/
# cat /proc/scsi/
# cat /proc/scsi/s
# cat /proc/scsi/s
# cat /proc/scsi/sc
# cat /proc/scsi/sc
# cat /proc/scsi/scs
# cat /proc/scsi/scs
# cat /proc/scsi/scsi
# cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
Vendor: Quantum Model: XP34300W Rev: L912
Type: Direct-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 01 Lun: 00
Vendor: QUANTUM Model: FIREBALL_TM1280S Rev: 300X
Type: Direct-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 02 Lun: 00
Vendor: SEAGATE Model: ST19101W Rev: 0014
Type: Direct-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 03 Lun: 00
Vendor: SONY Model: CD-ROM CDU-76S Rev: 1.1c
Type: CD-ROM ANSI SCSI revision: 02

#
#
# e
# e
# ex
# ex
# exi
# exi
# exit
# exit
Script done on Fri Mar 21 13:57:47 2003


I fear that the "new" system might be slow because there is some
loop going on in STDIN that's eating CPU time. That may be
the reason why `script` is broken in that distribution.
Anyway, sombody should check the evidence.


Cheers,
Dick Johnson
Penguin : Linux version 2.4.20 on an i686 machine (797.90 BogoMips).
Why is the government concerned about the lunatic fringe? Think about it.


2003-03-22 08:57:49

by Dumitru Ciobarcianu

[permalink] [raw]
Subject: Re: Ded-Fat 8.0 and ext3

On Vi, 2003-03-21 at 21:31, Richard B. Johnson wrote:
> Hello all. Greetings from fix-up land.
>
> I installed RedHat 8.0 of an existing system. By default it
> makes ext3 file-systems with journaling enabled.
>
> That distribution used Linux-2.4.18-14.
> I copied /boot/config-2.4.18-14 to /usr/src/linux-2.4.18-15/.config
> as a ".config" file and executed:
>
> `make oldconfig`
>
[stuff deleted]

AFAIR, you must do an "make mrproper" before copying the config and the
screenfull of errors will go away.


//Cioby


2003-03-24 14:00:40

by Richard B. Johnson

[permalink] [raw]
Subject: Re: Ded-Fat 8.0 and ext3

On Sat, 22 Mar 2003, Dumitru Ciobarcianu wrote:

> On Vi, 2003-03-21 at 21:31, Richard B. Johnson wrote:
> > Hello all. Greetings from fix-up land.
> >
> > I installed RedHat 8.0 of an existing system. By default it
> > makes ext3 file-systems with journaling enabled.
> >
> > That distribution used Linux-2.4.18-14.
> > I copied /boot/config-2.4.18-14 to /usr/src/linux-2.4.18-15/.config
> > as a ".config" file and executed:
> >
> > `make oldconfig`
> >
> [stuff deleted]
>

No. It isn't that easy. And if they keep rejecting
what I tell them, without actually checking and
testing what's in their distributions, they will
go the way of other arrogant companies who refuse
to listen to their customers.

As Red Hat "matures", it's distributions become
more and more like Microsoft in "look and feel".
This replaces functionality with glamor. This
it not good. You can add whatever glamor the
marketing departments require as long as you don't
remove long-time accepted functionality. Once
you throw away the functionality, you end up
loosing market share.

It is important that Red Hat remain in business so
that the "nay-sayers" do not win their arguments and
there remains several viable companies employing
kernel developers. We all have a stake in the success
of these distributors. If we we fail to report problems
with distributions, we actually fail to help support
Linux. On the other hand, when a distributor rejects
these reports out-of-hand, the whole Linux community
will eventually suffer the consequences.

Here is the output of standard-error while executing
`make dep`. It's informative:

find: /usr/src/linux-2.4.18-14/include/math-emu: No such file or directory
suspend.c:54:27: linux/compile.h: No such file or directory
In file included from suspend.c:56:
/usr/src/linux-2.4.18-14/include/linux/reboot.h:55:34: macro "software_suspend" passed 1 arguments, but takes just 0
suspend.c:981:27: macro "software_suspend" passed 1 arguments, but takes just 0
suspend.c:1277:26: macro "software_resume" passed 1 arguments, but takes just 0
serial.c:1524:1: directives may not be used inside a macro argument
serial.c:1522:39: unterminated argument list invoking macro "serial_outp"
au1000_gpio.c:41:24: asm/au1000.h: No such file or directory
au1000_gpio.c:42:29: asm/au1000_gpio.h: No such file or directory
In file included from i2c-algo-ite.c:50:
i2c-ite.h:36:31: asm/it8172/it8172.h: No such file or directory
rtc.c:27:25: asm/machdep.h: No such file or directory
rtc.c:29:22: asm/time.h: No such file or directory
via-pmu.c:40:22: asm/prom.h: No such file or directory
via-pmu.c:41:25: asm/machdep.h: No such file or directory
via-pmu.c:45:26: asm/sections.h: No such file or directory
via-pmu.c:48:30: asm/pmac_feature.h: No such file or directory
via-pmu.c:51:26: asm/sections.h: No such file or directory
via-pmu.c:52:26: asm/cputable.h: No such file or directory
via-pmu.c:53:22: asm/time.h: No such file or directory
i2o_pci.c:392:1: warning: no newline at end of file
audio.c:41:25: asm/audioio.h: No such file or directory
amd7930.c:95:26: asm/openprom.h: No such file or directory
amd7930.c:96:23: asm/oplib.h: No such file or directory
amd7930.c:100:22: asm/sbus.h: No such file or directory
amd7930.c:102:25: asm/audioio.h: No such file or directory
dbri.c:53:26: asm/openprom.h: No such file or directory
dbri.c:54:23: asm/oplib.h: No such file or directory
dbri.c:58:22: asm/sbus.h: No such file or directory
dbri.c:61:25: asm/audioio.h: No such file or directory
su.c:78:23: asm/oplib.h: No such file or directory
su.c:80:22: asm/ebus.h: No such file or directory
bbc_i2c.c:16:23: asm/oplib.h: No such file or directory
bbc_i2c.c:17:22: asm/ebus.h: No such file or directory
bbc_i2c.c:18:26: asm/spitfire.h: No such file or directory
bbc_i2c.c:19:21: asm/bbc.h: No such file or directory
In file included from bbc_i2c.c:21:
bbc_i2c.h:5:22: asm/ebus.h: No such file or directory
In file included from 53c700.c:134:
53c700.h:40:2: #error "Config.in must define either CONFIG_53C700_IO_MAPPED or CONFIG_53C700_MEM_MAPPED to use this scsi core."
53c700.c:155:22: 53c700_d.h: No such file or directory
fas216.c:55:23: asm/ecard.h: No such file or directory
newport.c:11:21: asm/gfx.h: No such file or directory
newport.c:12:21: asm/ng1.h: No such file or directory
rrm.c:15:21: asm/rrm.h: No such file or directory
shmiq.c:57:23: asm/shmiq.h: No such file or directory
shmiq.c:58:21: asm/gfx.h: No such file or directory
usema.c:38:25: asm/usioctl.h: No such file or directory
tc.c:15:27: asm/addrspace.h: No such file or directory
tc.c:17:30: asm/dec/machtype.h: No such file or directory
tc.c:18:28: asm/dec/tcinfo.h: No such file or directory
tc.c:19:30: asm/dec/tcmodule.h: No such file or directory
tc.c:20:32: asm/dec/interrupts.h: No such file or directory
sa1100fb.c:166:26: asm/hardware.h: No such file or directory
sa1100fb.c:169:28: asm/mach-types.h: No such file or directory
sa1100fb.c:171:30: asm/arch/assabet.h: No such file or directory
In file included from zorro.c:17:
/usr/src/linux-2.4.18-14/include/linux/zorro.h:158:23: asm/zorro.h: No such file or directory
zorro.c:20:25: asm/amigahw.h: No such file or directory

Now; I have attached the script, showing what I executed
to actually generate this. As previously reported by myself,
Version 8.0 even breaks the execution of `script` so there
is a lot of wrong console activity recorded.

I did not bother to go any further than `make oldconfig` and
`make dep` in this "Show-and-tell". As previously reported,
continuing with `make bzImage` will result in many errors and
the failure to compile. Further, what used to take a few minutes
to compile on that machine, now takes a whole day. This is
because something went wrong with the ext3 file-system which
Red Hat installed by default. The Red Hat installation never
asked me if I wanted to use ext3. It also didn't ask me if
I wanted a journal file-system. Instead, its arrogance
resulted in an unusable system that will have to be redone
as a long, time-consuming project.



Cheers,
Dick Johnson
Penguin : Linux version 2.4.20 on an i686 machine (797.90 BogoMips).
Why is the government concerned about the lunatic fringe? Think about it.


Attachments:
typescript.gz (59.44 kB)

2003-03-24 14:10:25

by Arjan van de Ven

[permalink] [raw]
Subject: Re: Ded-Fat 8.0 and ext3

On Mon, 2003-03-24 at 15:11, Richard B. Johnson wrote:

> I did not bother to go any further than `make oldconfig` and
> `make dep` in this "Show-and-tell". As previously reported,

as previously report to YOU: you have to do a make mrproper first.
Then it just works.

I've not received a SINGLE report where starting with make mrproper
didn't fix this issue. You can claim I ignore this issue, but I don't.
It's just not an issue at all so far!


Attachments:
signature.asc (189.00 B)
This is a digitally signed message part

2003-03-24 14:20:46

by Richard B. Johnson

[permalink] [raw]
Subject: Re: Ded-Fat 8.0 and ext3

On Mon, 24 Mar 2003, Arjan van de Ven wrote:

> On Mon, 2003-03-24 at 15:11, Richard B. Johnson wrote:
>
> > I did not bother to go any further than `make oldconfig` and
> > `make dep` in this "Show-and-tell". As previously reported,
>
> as previously report to YOU: you have to do a make mrproper first.
> Then it just works.
>

Look at the damn script. It does every possible:
make clean
make distclean
make mrproper (line 71)

It's hard to find because script got broken in that distribution,
too.

Again look at the evidence, rather than just barking back a
retort.

> I've not received a SINGLE report where starting with make mrproper
> didn't fix this issue. You can claim I ignore this issue, but I don't.
> It's just not an issue at all so far!
>
Well you got more than a "SINGLE" report now. Please check it out.



Cheers,
Dick Johnson
Penguin : Linux version 2.4.20 on an i686 machine (797.90 BogoMips).
Why is the government concerned about the lunatic fringe? Think about it.