[U-Boot-Users] Re: JFFS2: CRC Problem with Kernel Image

[U-Boot-Users] Re: JFFS2: CRC Problem with Kernel Image

[Martin, Vitali]

> Which version of U-Boot is this?

0.4.1

> Are you sure the image is still OK?

My target mounts root filesystem over NFS. So running on my host system "mkimage -l /mnt/pImage" works:

	Target:

	bash-2.05# mount -t jffs2 /dev/mtdblock1 /mnt/
	bash-2.05# cd /mnt/
	bash-2.05# ll
	total 514
	-rw-r--r--    1 root     root       525815 Jul 22 13:31 pImage
	bash-2.05# cp pImage /pImage

	Host:

	vm at pc241:/nfs/eldk/ppc_8xx> mkimage -l pImage
	Image Name:   Linux-2.4.4
	Created:      Tue Jul 22 15:28:19 2003
	Image Type:   PowerPC Linux Kernel Image (gzip compressed)
	Data Size:    525751 Bytes = 513.43 kB = 0.50 MB
	Load Address: 0x00000000
	Entry Point:  0x00000000


With the same image "tftpboot" works fine.

	->tftp 100000 pImage
	TFTP from server 172.22.20.193; our IP address is 172.22.10.33
	Filename 'pImage'.
	Load address: 0x100000
	Loading: T ##################################
	done
	Bytes transferred = 525815 (805f7 hex)
	->imi
	## Checking Image at 00100000 ...
	   Image Name:   Linux-2.4.4
	   Created:      2003-07-22  13:28:19 UTC
	   Image Type:   PowerPC Linux Kernel Image (gzip compressed)
	   Data Size:    525751 Bytes = 513.4 kB
	   Load Address: 00000000
	   Entry Point:  00000000
	   Verifying Checksum ... OK
	->

So I am sure that my image file is OK, I suppose something goes wrong in "fsload", when pImage will copied to memory.

[U-Boot-Users] Re: JFFS2: CRC Problem with Kernel Image

[Mark Rakes]

How did the pImage file get to /dev/mtdblock1 in the first place?

We had a problem a while back where using scp to copy an image to
a jffs2 flash partition would corrupt the image (happened like 12% of
the time). Finally the cause was found to be a timer in scp used  to
update the progress bar to the user (timer alarm was handled mid-write
to flash). We ended up masking the timer alarm in scp until the write
was finished, and all was OK.

In reality, we're working around some bug in the kernel jffs2 code...

hth,
-mark


On Tuesday, July 22, 2003, at 06:44  AM, Martin, Vitali wrote:
>> Which version of U-Boot is this?
> 0.4.1
>> Are you sure the image is still OK?
> My target mounts root filesystem over NFS. So running on my host 
> system "mkimage -l /mnt/pImage" works:
> 	Target:
> 	bash-2.05# mount -t jffs2 /dev/mtdblock1 /mnt/
> 	bash-2.05# cd /mnt/
> 	bash-2.05# ll
> 	total 514
> 	-rw-r--r--    1 root     root       525815 Jul 22 13:31 pImage
> 	bash-2.05# cp pImage /pImage
>
> 	Host:
> 	vm at pc241:/nfs/eldk/ppc_8xx> mkimage -l pImage
> 	Image Name:   Linux-2.4.4
> 	Created:      Tue Jul 22 15:28:19 2003
> 	Image Type:   PowerPC Linux Kernel Image (gzip compressed)
> 	Data Size:    525751 Bytes = 513.43 kB = 0.50 MB
> 	Load Address: 0x00000000
> 	Entry Point:  0x00000000
>
> With the same image "tftpboot" works fine.
> 	->tftp 100000 pImage
> 	TFTP from server 172.22.20.193; our IP address is 172.22.10.33
> 	Filename 'pImage'.
> 	Load address: 0x100000
> 	Loading: T ##################################
> 	done
> 	Bytes transferred = 525815 (805f7 hex)
> 	->imi
> 	## Checking Image at 00100000 ...
> 	   Image Name:   Linux-2.4.4
> 	   Created:      2003-07-22  13:28:19 UTC
> 	   Image Type:   PowerPC Linux Kernel Image (gzip compressed)
> 	   Data Size:    525751 Bytes = 513.4 kB
> 	   Load Address: 00000000
> 	   Entry Point:  00000000
> 	   Verifying Checksum ... OK
> 	->
> So I am sure that my image file is OK, I suppose something goes wrong 
> in "fsload", when pImage will copied to memory.

[U-Boot-Users] Re: JFFS2: CRC Problem with Kernel Image

[Martin, Vitali]

> So what address are you loading to 1000 10000 or 100000

Sorry that I wasn't attentive im my last email.
The memory address is always 0x100000.

And data written by "tftpboot" and "fsload" to 0x100000 is not the same. There are differences from address 0x00101000, so I get Linux CRC Error running "iminfo" or "bootm".

	tftp   =  00101000: dfc958e0 5f66d603 e27f2fe3 0027f536
	fsload =  00101000: dfc9c970 de53a432 188d2417 af469b2b  

[U-Boot-Users] Re: JFFS2: CRC Problem with Kernel Image

[Kenneth Johansson]

On Thu, 2003-07-24 at 08:55, Martin, Vitali wrote:

> I checked the memory with "md 100000" in U-Boot and there seems to be
> not the same data on the memory address 0x100 when I using "tftp 1000
> pImage" and "fsload 10000 pImage". From address 0x101000 the data im the
> RAM memory is different???

So what address are you loading to 1000 10000 or 100000


> 
> 	->tftp 10000 pImage

So you load at 0x10000 add 500kB and you do not write any data at
0x100000.

> 	TFTP from server 172.22.20.193; our IP address is 172.22.10.33
> 	Filename 'pImage'.
> 	Load address: 0x10000
> 	Loading: T ###################T #TX timeout
> 	done
> 	Bytes transferred = 520149 (7efd5 hex)
> 	->md 100000 200000
> 	00100000: 27051956 52c9c624 3f1d4bb5 0007ef95
> '..VR..$?.K.....
> 	00100010: 00000000 00000000 5171b599 05070201
> ........Qq......
> 	00100020: 4c696e75 782d322e 342e3400 00000000
> Linux-2.4.4.....
> 	...
> 	00101000: dfc958e0 5f66d603 e27f2fe3 0027f536
> ..X._f..../..'.6
> 	00101010: c055f99e f5e16e85 bd761595 a9ce7bd2
> .U....n..v....{.
> 
> 	->fsload 100000 pImage
> 	### JFFS2 loading 'pImage' to 0x100000
> 	### JFFS2 load complete: 520149 bytes loaded to 0x100000
> 	->md 100000 200000
> 	00100000: 27051956 52c9c624 3f1d4bb5 0007ef95
> '..VR..$?.K.....
> 	00100010: 00000000 00000000 5171b599 05070201
> ........Qq......
> 	00100020: 4c696e75 782d322e 342e3400 00000000
> Linux-2.4.4.....
> 	...
> 	00101000: dfc9c970 de53a432 188d2417 af469b2b
> ...p.S.2..$..F.+
> 	00101010: 7dd3d9cf e1df4c64 65acdb75 83efc98b
> }.....Lde..u....

-- 
Kenneth Johansson	
Ericsson AB                       Tel: +46 8 719 70 20
Tellusborgsv?gen  90              Fax: +46 8 719 29 45
126 25 Stockholm                  ken at switchboard.ericsson.se

[U-Boot-Users] Re: JFFS2: CRC Problem with Kernel Image

[Martin, Vitali]

> How did the pImage file get to /dev/mtdblock1 in the first place?

	bash-2.05# eraseall /dev/mtd1
	bash-2.05# mount -t jffs2 /dev/mtdblock1 /mnt/
	bash-2.05# cp /pImage /mnt/

> We had a problem a while back where using scp to copy an image ...

I'm using "cp" from the linux console, so I think there is no interrupt
problem.
Because I can copy pImage to JFFS2 filesystem and my image is still OK.


I checked the memory with "md 100000" in U-Boot and there seems to be
not the same data on the memory address 0x100 when I using "tftp 1000
pImage" and "fsload 10000 pImage". From address 0x101000 the data im the
RAM memory is different???

	->tftp 10000 pImage
	TFTP from server 172.22.20.193; our IP address is 172.22.10.33
	Filename 'pImage'.
	Load address: 0x10000
	Loading: T ###################T #TX timeout
	done
	Bytes transferred = 520149 (7efd5 hex)
	->md 100000 200000
	00100000: 27051956 52c9c624 3f1d4bb5 0007ef95
'..VR..$?.K.....
	00100010: 00000000 00000000 5171b599 05070201
........Qq......
	00100020: 4c696e75 782d322e 342e3400 00000000
Linux-2.4.4.....
	...
	00101000: dfc958e0 5f66d603 e27f2fe3 0027f536
..X._f..../..'.6
	00101010: c055f99e f5e16e85 bd761595 a9ce7bd2
.U....n..v....{.

	->fsload 100000 pImage
	### JFFS2 loading 'pImage' to 0x100000
	### JFFS2 load complete: 520149 bytes loaded to 0x100000
	->md 100000 200000
	00100000: 27051956 52c9c624 3f1d4bb5 0007ef95
'..VR..$?.K.....
	00100010: 00000000 00000000 5171b599 05070201
........Qq......
	00100020: 4c696e75 782d322e 342e3400 00000000
Linux-2.4.4.....
	...
	00101000: dfc9c970 de53a432 188d2417 af469b2b
...p.S.2..$..F.+
	00101010: 7dd3d9cf e1df4c64 65acdb75 83efc98b
}.....Lde..u....