NFS support patch
I uploaded NFS support patch to my web site.
http://www.sonare.it/u-boot/u-boot-nfs.patch.gz
If you are interested in, please evaluate.
The most part of this is taken from etherboot.
To use NFS support, add the definition CFG_CMD_NFS to CONFIG_COMMANDS defined in the configration file of your board.
Thank you for the any advices and suggestions.
With best regards.
Masami Komiya at tokyo.jp
Hello,
Quoted 3 lines · in message <[email protected]> you wrote:
I uploaded NFS support patch to my web site.
Well done, thanks a lot!
Checked in.
Best regards,
Wolfgang Denk
Signature · 3 lines
Software Engineering: Embedded and Realtime Systems, Embedded Linux Phone: (+49)-8142-4596-87 Fax: (+49)-8142-4596-88 Email: wd at denx.de You can only live once, but if you do it right, once is enough.
Hi
Quoted 7 lines
Quoted 3 lines
I uploaded NFS support patch to my web site.
Well done, thanks a lot!
Checked in.
I think CFG_CMD_NFS should be considered "non-standard". It does not even compile for AT91RM9200.
Index: include/cmd_confdefs.h
===================================================================
RCS file: /cvsroot/u-boot/u-boot/include/cmd_confdefs.h,v retrieving revision 1.16
cmd_confdefs.h 24 Feb 2004 11:28:21 -0000 renamed from cmd_confdefs.h 24 Feb 2004 02:00:13 -0000 1.16 +1 -0
@@ -119,6 +119,7 @@ CFG_CMD_MII | \ CFG_CMD_MMC | \ CFG_CMD_NAND | \ + CFG_CMD_NFS | \ CFG_CMD_PCI | \ CFG_CMD_PCMCIA | \ CFG_CMD_PING | \
Signature · 6 lines
Steven Scholz imc Measurement & Control imc Me?systeme GmbH Voltastr. 5 Voltastr. 5 13355 Berlin 13355 Berlin Germany Deutschland
I have no objection.
Masami Komiya
Quoted 33 lines · Steven Scholz
Hi
Quoted 8 lines
Quoted 3 lines
I uploaded NFS support patch to my web site.
Well done, thanks a lot!
Checked in.
I think
CFG_CMD_NFSshould be considered "non-standard". It does not even compile for AT91RM9200.Index:
include/cmd_confdefs.h
===================================================================
RCS file: /cvsroot/u-boot/u-boot/include/cmd_confdefs.h,v retrieving revision 1.16
diff -p -u -r1.16cmd_confdefs.hcmd_confdefs.h 24 Feb 2004 11:28:21 -0000 renamed from cmd_confdefs.h 24 Feb 2004 02:00:13 -0000 1.16 +1
-0@@ -119,6 +119,7 @@ CFG_CMD_MII | \ CFG_CMD_MMC | \ CFG_CMD_NAND | \ + CFG_CMD_NFS | \ CFG_CMD_PCI | \ CFG_CMD_PCMCIA | \ CFG_CMD_PING | \
Hi Masami & Steven,
Quoted 17 lines
I have no objection.
Masami Komiya
Quoted 12 lines · Steven Scholz
Hi
Quoted 8 lines
Quoted 3 lines
I uploaded NFS support patch to my web site.
Well done, thanks a lot!
Checked in.
I think
CFG_CMD_NFSshould be considered "non-standard". It does not even compile for AT91RM9200.
As long as the NFS support does not compile on all platforms it should probably be non-standard - _iff_ it compiles on all platforms I would certainly reconsider this as the code impact is minimal but the feature will be very helpful (IMHO).
But until then, Steven can you share the compile problems with us so we can fix them? There should really be no general reason why this should not work on ARM, or am I missing something?
Thanks!
Detlev
Signature · 3 lines
Directories are added, deleted, and rearranged much as you would
expect, even if you don't know it's what you'd expect.
-- Tom Lord in TLA DocumentationHi Detlev,
Quoted 11 lines
Quoted 2 lines
Quoted 2 lines
I think
CFG_CMD_NFSshould be considered "non-standard". It does not even compile for AT91RM9200.As long as the NFS support does not compile on all platforms it should probably be non-standard - _iff_ it compiles on all platforms I would certainly reconsider this as the code impact is minimal but the feature will be very helpful (IMHO).
But until then, Steven can you share the compile problems with us so we can fix them? There should really be no general reason why this should not work on ARM, or am I missing something?
No big deal.
There's no get_ticks() yet for AT91RM9200. I fixed that and I am about to test it and to provide a patch real soon now.
Signature · 6 lines
Steven Scholz imc Measurement & Control imc Me?systeme GmbH Voltastr. 5 Voltastr. 5 13355 Berlin 13355 Berlin Germany Deutschland
Quoted 11 lines · Masami Komiya
I uploaded NFS support patch to my web site.
http://www.sonare.it/u-boot/u-boot-nfs.patch.gz
If you are interested in, please evaluate.
The most part of this is taken from etherboot.
To use NFS support, add the definition
CFG_CMD_NFStoCONFIG_COMMANDSdefined in the configration file of your board.
I tried it but it does not work on my ARM (AT91RM9200) board:
multiIO> nfs 0x20500000 10.0.2.9:/opt/eldk/arm_920TDI/uImage File transfer via NFS from server 10.0.2.9; our IP address is 10.0.9.155 Filename '/opt/eldk/arm_920TDI/uImage'.
Load address: 0x20500000
Loading: #################################################################
#############################################
done
Bytes transferred = 560892 (88efc hex)
Fair enough so far. The length is correct. But apparently I did not get the correct data:
multiIO> md.w 0x20500000
20500000: 0500 4c89 3b40 5f7c 0000 0000 3b40 b77a ...L@;|_....@;z. 20500010: 0000 0000 3b40 b77a 0000 0000 ffff ffff ....@;z......... 20500020: ffff ffff ffff ffff ffff ffff ffff ffff ................ 20500030: ffff ffff ffff ffff ffff ffff ffff ffff ................ 20500040: ffff ffff ffff ffff ffff ffff ffff ffff ................ 20500050: ffff ffff ffff ffff ffff ffff ffff ffff ................ 20500060: ffff ffff ffff ffff ffff ffff ffff ffff ................ 20500070: ffff ffff ffff ffff ffff ffff ffff ffff ................
But it should look like this:
scholz at Pinguin:/opt/eldk/arm_920TDI> hexdump uImage | head 0000000 0527 5619 1fb4 eeca 1640 146c 0800 bc8e 0000010 0020 0080 0020 0080 455f 8430 0205 0102 0000020 0000 0000 0000 0000 0000 0000 0000 0000
*
0000040 8b1f 0808 6c13 4016 0302 696c 756e 2e78 0000050 6962 006e bdec 7c0d 555c 3f99 ee7e a4bc 0000060 6493 dc9a 1264 6d08 b7a0 806d 0180 da6e 0000070 a100 74a6 8b4a 5b54 da65 a806 3558 d140
What's the problem?
Signature · 6 lines
Steven Scholz imc Measurement & Control imc Me?systeme GmbH Voltastr. 5 Voltastr. 5 13355 Berlin 13355 Berlin Germany Deutschland
Steven,
Quoted 2 lines
Fair enough so far. The length is correct. But
apparently I did not get the correct data:
...
Quoted 1 line
What's the problem?
The packet wasn't copied correctly. Try this
patch. I tested with this patch on a PQ2FADS
board and everything looked good.
Regards,
--Scott
=> tftpboot 4000 u-boot.bin
Using FCC2 ETHERNET device
TFTP from server 192.168.0.40; our IP address is 192.168.0.99
Filename 'u-boot.bin'.
Load address: 0x4000
Loading: ########################################
done
Bytes transferred = 209720 (33338 hex)
=> nfs 40000 192.168.0.40:/tftpboot/u-boot.bin
Using FCC2 ETHERNET device
File transfer via NFS from server 192.168.0.40; our IP address is 192.168.0.99
Filename '/tftpboot/u-boot.bin'.
Load address: 0x40000
Loading: ########################################
done
Bytes transferred = 209720 (33338 hex)
=> cmp.b 4000 40000 33338
Total of 209720 bytes were the same
=>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: nfs-reply-len.patch
Url: http://lists.denx.de/pipermail/u-boot/attachments/20040224/8309d3c3/attachment.txt
Quoted 3 lines · In message <[email protected]> you wrote:
I think
CFG_CMD_NFSshould be considered "non-standard". It does not even compile for AT91RM9200.
Actually I want to make this "standard". If the AT91RM9200 is missing some function it should be fixed.
Best regards,
Wolfgang Denk
Signature · 3 lines
Software Engineering: Embedded and Realtime Systems, Embedded Linux Phone: (+49)-8142-4596-87 Fax: (+49)-8142-4596-88 Email: wd at denx.de Bus error -- please leave by the rear door.
Thank you for your evaluation and advise.
The cause of the problem is the pointer's value for store_block(). Scott's patch is the one of the solution.
The attached patch reduces the copy count in nfs_read_reply() and store_block(). This patch is for current CVS.
Thank you for your comment.
With best regards.
Masami Komiya
-------------- next part --------------
A non-text attachment was scrubbed...
Name: nfs-test-040225.patch.gz
Type: application/x-gzip
Size: 550 bytes
Desc: not available
Url : http://lists.denx.de/pipermail/u-boot/attachments/20040225/56547b3b/attachment.bin
Quoted 7 lines · Masami Komiya
Thank you for your evaluation and advise.
The cause of the problem is the pointer's value for
store_block(). Scott's patch is the one of the solution.The attached patch reduces the copy count in
nfs_read_reply()andstore_block(). This patch is for current CVS.
Patch works for me:
multiIO> nfs $loadaddr 10.0.2.9:/opt/eldk/arm_920TDI/uImage File transfer via NFS from server 10.0.2.9; our IP address is 10.0.9.155 Filename '/opt/eldk/arm_920TDI/uImage'.
Load address: 0x20500000
Loading: #################################################################
#############################################
done
Bytes transferred = 560892 (88efc hex)
multiIO> imi
## Checking Image at 20500000 ...
Image Name:
Image Type: ARM Linux Kernel Image (gzip compressed)
Data Size: 560828 Bytes = 547.7 kB
Load Address: 20008000
Entry Point: 20008000
Verifying Checksum ... OKThanks.
Quoted 2 lines · Wolfgang Denk
Actually I want to make this "standard". If the AT91RM9200 is missing some function it should be fixed.
I do _not_ object any longer. I'll submit a patch for AT91RM9200.
Signature · 6 lines
Steven Scholz imc Measurement & Control imc Me?systeme GmbH Voltastr. 5 Voltastr. 5 13355 Berlin 13355 Berlin Germany Deutschland
Hi Steven
Quoted 1 line · Steven Scholz
Patch works for me:
:
Quoted 1 line
Thanks.
Thank you for your report.
I have the peace of mind.
# I don't realize why the only my machine could get the success ....
Quoted 3 lines
Quoted 2 lines · Wolfgang Denk
Actually I want to make this "standard". If the AT91RM9200 is missing some function it should be fixed.
It's my pleasure.
Quoted 1 line
I do _not_ object any longer. I'll submit a patch for AT91RM9200.
I wish the problem won't be find out more :-)
Masami Komiya @ Tokyo.JP
Quoted 2 lines · In message <[email protected]> you wrote:
Thank you for your evaluation and advise.
Thanks a lot for your contribution(s).
Quoted 5 lines
The cause of the problem is the pointer's value for
store_block(). Scott's patch is the one of the solution.The attached patch reduces the copy count in
nfs_read_reply()andstore_block(). This patch is for current CVS.
Checked in.
Best regards,
Wolfgang Denk
Signature · 7 lines
Software Engineering: Embedded and Realtime Systems, Embedded Linux Phone: (+49)-8142-4596-87 Fax: (+49)-8142-4596-88 Email: wd at denx.de "Science makes godlike -- it is all over with priests and gods when man becomes scientific. Moral: science is the forbidden as such -- it alone is forbidden. Science is the *first* sin, the *original* sin. *This alone is morality.* ``Thou shalt not know'' -- the rest follows." - Friedrich Nietzsche