The Linux Memory Technology Driver subsystem

?? | 1 Aug 2002 14:30

Picon

thanks , after this patch , i mount jffs2 successfully

dear all,

thanks , after this patch , i mount jffs2 successfully
(i also patched to make zlib_deflate/zlib_inflate inside kernel's lib
instead of module)

best regards,
kevin

----- Original Message -----
From: "Thomas Gleixner" <tglx <at> linutronix.de>
To: "??" <kevin <at> gv.com.tw>; "David Woodhouse" <dwmw2 <at> infradead.org>
Cc: <linux-mtd <at> lists.infradead.org>
Sent: Thursday, August 01, 2002 5:28 PM
Subject: Re: do i have to "eraseall" first?

> On Thursday, 1. August 2002 05:23, ?? wrote:
> > dear all,
> >
> > the only different config settings is CONFIG_JFFS2_FS_NAND
> > but there is no "CONFIG_JFFS2_FS_NAND" in source of jffs2/mtd here!!
> > (i'm using lastest jffs2/mtd source from cvs)
> > please help
> Sorry, CONFIG_JFFS2_FS_NAND is not patched in. Please add it manually.
>
> See updated information at
http://www.linux-mtd.infradead.org/tech/nand.html
>
> --
> Thomas

(Continue reading)

David Woodhouse | 5 Aug 2002 11:54

Re: error when writing to jffs2

kevin <at> gv.com.tw said:
>  error when writing to jffs2 any idea?

What version of fs/jffs2/scan.c do you have?

--
dwmw2

David Woodhouse | 5 Aug 2002 12:25

Re: error when writing to jffs2 (scan.c v1.79 20020725)

kevin <at> gv.com.tw said:
> scan.c v 1.79 20020725

OK, then I don't know -- could you get full logs with 
CONFIG_JFFS2_FS_DEBUG=1 from mounting the file system to the oops? Make 
sure you are logging over a serial console so you catch _all_ the messages, 
and remember they're at level KERN_DEBUG so echo 9 > /proc/sys/kernel/printk

--
dwmw2

?? | 6 Aug 2002 06:06

Picon

Re: error when writing to jffs2 (success in another evb)

by enabling CONFIG_JFFS2_FS_DEBUG=1
i discover a typo buggy in fs/jffs2/gc.c:
fs/fs.o: In function `jffs2_garbage_collect_hole':
gc.c(.text+0x45c9c): undefined reference to `unlikely'
gc.c(.text+0x45c9c): relocation truncated to fit: R_MIPS_26 unlikely
fixed it by add:
#include <linux/compiler.h>
to fs/jffs2/gc.c for macro<"unlikely">
---success
story----------------------------------------------------------------------
success story: in our evb241x, jffs2 is ported ok

# mount
/dev/root on / type jffs2 (rw)
none on /proc type proc (rw)
/dev/ram1 on /var type minix (rw)
# df -h
Filesystem                Size      Used Available Use% Mounted on
/dev/mtdblock1            6.7M      2.3M      4.4M  34% /
/dev/ram1              1009.0k     23.0k    986.0k   2% /var
#
--- faulting
log---------------------------------------------------------------------
the faulting kernel in our evb13x can't boot today, so log can't send :<
log show: "kernel BUG at traps.c:623!"
(probably hw problem or buggy mips 2.4.17 kernel)

POSIX conformance testing by UNIFIX
kernel BUG at traps.c:623!
Unable to handle kernel paging request at virtual address 00000000, epc ==

(Continue reading)

Thomas Gleixner | 6 Aug 2002 11:46

Picon

Re: error when writing to jffs2 (success in another evb)

On Tue, 2002-08-06 at 06:06, ?? wrote:
> the faulting kernel in our evb13x can't boot today, so log can't send :<
could you send me your hardware driver ? 
--

-- 
Thomas 
____________________________________________________
linutronix - competence in embedded & realtime linux
http://www.linutronix.de
mail: tglx <at> linutronix.de

nix.de

David Woodhouse | 6 Aug 2002 10:26

Re: what't the max partition(single file size) jffs2/mtd can support?(64MB? or 2GB? or 8GB?)

kevin <at> gv.com.tw said:
> what't the max partition(single file size) jffs2/mtd can support?
> 64MB? or 2GB? or 8GB? thanks in advanced!:) 

The maximum partition size is 4GiB, as is the maximum file size for JFFS2.
Theoretically -- I suspect the performance would suck too hard _long_ 
before you got there.

--
dwmw2

David Woodhouse | 6 Aug 2002 14:26

Re: fail to rm files when jffs2 is full

kevin <at> gv.com.tw said:
> when we test disk-full with jffs2/mtd, we fail to rm files when jffs2
> is full

Were there any other errors leading up to this? The flash should never get 
so full that it doesn't allow you to remove things.

--
dwmw2

David Woodhouse | 7 Aug 2002 12:05

Re: No free space left for GC(test2 with log)

kevin <at> gv.com.tw said:
> jffs2_g_c_deletion_dirent(): Short read (40 not 12) reading from
> obsolete node at 00633ae0

Ok, this was a stupid error on my part which prevented GC from working. It 
may well have been causing the problem. CVS update and try again.

--
dwmw2

David Woodhouse | 6 Aug 2002 15:07

Re: No free space left for GC(test2 with log)

kevin <at> gv.com.tw said:
> jffs2_g_c_deletion_dirent(): Short read (40 not 12) reading from
> obsolete node at 00665b68 

Er, that's broken. Thomas?

--
dwmw2

?? | 6 Aug 2002 15:03

Picon

No free space left for GC(test2 with log)

(console log of evb241x:)

# mkdir aaa

# df
Filesystem           1k-blocks      Used Available Use% Mounted on
/dev/mtdblock1            6904      1612      5292  23% /
/dev/ram1                 1009        17       992   2% /var
192.72.4.43:/home/test/mips   4134900    328600   3596252   8% /mnt

(/mnt is a mounted nfs directory on pc)

# cp /mnt/* /aaa/
jffs2_g_c_deletion_dirent(): Short read (40 not 12) reading from obsolete
node at 00633ae0
jffs2_g_c_deletion_dirent(): Short read (40 not 12) reading from obsolete
node at 00665b68

(after about 10 minutes, we do a ctrl-c here to interrupt the files copy)

#

(it begin to show "No free space left for GC" repeatly after ctrl-c)

Argh. No free space left for GC. nr_erasing_blocks is 0. nr_free_blocks is
0. (erasableempty: yes, erasingempty: yes, erasependingempty: yes)
jffs2_reserve_space_gc of 196 bytes for garbage_collect_dnode failed: -28
Argh. No free space left for GC. nr_erasing_blocks is 0. nr_free_blocks is
0. (erasableempty: yes, erasingempty: yes, erasependingempty: yes)
mjffs2_reserve_space_gc of 196 bytes for garbage_collect_dnode failed: -28

(Continue reading)

Thomas Gleixner | 6 Aug 2002 16:08

Picon

Re: No free space left for GC(test2 with log)

On Tue, 2002-08-06 at 15:03, ?? wrote:
> (console log of evb241x:)
> 
> # mkdir aaa
> 
> # df
> Filesystem           1k-blocks      Used Available Use% Mounted on
> /dev/mtdblock1            6904      1612      5292  23% /
> /dev/ram1                 1009        17       992   2% /var
> 192.72.4.43:/home/test/mips   4134900    328600   3596252   8% /mnt
> 
> (/mnt is a mounted nfs directory on pc)
> 
> # cp /mnt/* /aaa/
> jffs2_g_c_deletion_dirent(): Short read (40 not 12) reading from obsolete
> node at 00633ae0
> jffs2_g_c_deletion_dirent(): Short read (40 not 12) reading from obsolete
> node at 00665b68
We really need a complete log from mount to this point. Could you please
provide one ?

compile kernel with CONFIG_JFFS2_DEBUG = 1 and serial console

boot
# echo 9 >/proc/sys/kernel/printk
Switch on logging on your terminal program
# mount /dev/mtdblockx
# mkdir
# cp
Save logfile and compress it

(Continue reading)

?? | 6 Aug 2002 13:43

Picon

fail to rm files when jffs2 is full

dear all,

when we test disk-full with jffs2/mtd,
we fail to rm files when jffs2 is full
it keeps showing the following messages repeatly:
----fail to rm files when jffs2 is
full------------------------------------------------
jffs2_reserve_space_gc of 196 bytes for garbage_collect_dnode failed: -28
Argh. No free space left for GC. nr_erasing_blocks is 0. nr_free_blocks is
0. (e
rasableempty: yes, erasingempty: yes, erasependingempty: yes)
.....repeat ........
----------------------------------------------------------------------------
------
any idea?
thanks in advanced!:)

?? | 6 Aug 2002 08:13

Picon

what't the max partition(single file size) jffs2/mtd can support?(64MB? or 2GB? or 8GB?)

what't the max partition(single file size) jffs2/mtd can support?
64MB? or 2GB? or 8GB?
thanks in advanced!:)

?? | 5 Aug 2002 12:18

Picon

Re: error when writing to jffs2 (scan.c v1.79 20020725)

scan.c v 1.79 20020725

/*
 * JFFS2 -- Journalling Flash File System, Version 2.
 *
 * Copyright (C) 2001, 2002 Red Hat, Inc.
 *
 * Created by David Woodhouse <dwmw2 <at> cambridge.redhat.com>
 *
 * For licensing information, see the file 'LICENCE' in this directory.
 *
 * $Id: scan.c,v 1.79 2002/07/25 20:48:51 dwmw2 Exp $
 *
 */

----- Original Message -----
From: "David Woodhouse" <dwmw2 <at> infradead.org>
To: "??" <kevin <at> gv.com.tw>
Cc: <tglx <at> linutronix.de>; <linux-mtd <at> lists.infradead.org>; "? ??"
<djh <at> gv.com.tw>; <cms <at> gv.com.tw>
Sent: Monday, August 05, 2002 5:54 PM
Subject: Re: error when writing to jffs2

>
> kevin <at> gv.com.tw said:
> >  error when writing to jffs2 any idea?
>
> What version of fs/jffs2/scan.c do you have?
>
> --

(Continue reading)

?? | 2 Aug 2002 04:50

Picon

error when writing to jffs2

dear all,

error when writing to jffs2
any idea?

thanks in advanced!:)

--------- following message log -----------------------------
# cp -a bin bin2
ARGH. About to write node to 0x0004c020 on flash, but there's data already
there
:
0x0004c020: 85 19 01 e0 2d 00 00 00 3a 31 4d 58 04 00 00 00
argh. node added in wrong place
ARGH. About to write node to 0x0004c064 on flash, but there's data already
there
:
0x0004c064: 80 21 00 00 00 00 00 00 00 00 00 00 7d 7c 76 3a
argh. node added in wrong place
nand_read_ecc: Failed ECC read, page 0x00001260
ARGH. About to write node to 0x0004c020 on flash, but there's data already
there
:
0x0004c020: 85 19 00 e0 04 00 00 00 18 31 45 18 00 00 00 00
argh. node added in wrong place
Calculated used size 00003fc8 != stored used size 00003f84
kernel BUG at nodemgmt.c:320!
Unable to handle kernel paging request at virtual address 00000000, epc ==
80086
154, ra == 80086154

(Continue reading)

Return

Return to gmane.linux.drivers.mtd.

Project Web Page

The Linux Memory Technology Driver subsystem

Search Archive

Language

Change language

Options

Current view: Threads only / Showing only 20 lines / Not hiding cited text.
Change to All messages, whole messages, or hide cited text.

Post a message
NNTP Newsgroup
Classic Gmane web interface
XML

XML

RSS Feed
List Information

About Gmane