RE: [ECOS] [Fwd: How to enable redboots CRC checking]

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Winkler Andreas]

> I think he's just trying to test RedBoot's checking of the CRC,
> by corrupting the data and showing that the Linux de-compressor
> detects it properly.

Yes, exact.

Regards
Andreas

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Winkler Andreas]

> From: ecos-discuss-owner@ecos.sourceware.org
> [mailto:ecos-discuss-owner@ecos.sourceware.org]On Behalf Of 
> Gary Thomas
> It would be nice to know what their schedule is...

The snapshot should be updated every sunday (maybe except in may :-) )
 
> Attached is the [very simple] patch that should fix your problem.  You
> should not have to update anything else.

Thanks a lot for your really very quick reply. I'll try at once.

Regards
Andreas

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Gary Thomas]

[-- Attachment #1: Type: text/plain, Size: 779 bytes --]

On Wed, 2004-05-26 at 07:15, Winkler Andreas wrote:
> > Sorry, but I don't know anything about the eCosCentric snapshots.
> > 
> > Perhaps you can update via CVS?
> 
> Unfortunately our proxy server doesn't support the access to the CVS
> repository.
> So I downloaded the sources from
> http://www.ecoscentric.com/devzone/snapshots.shtml
> which where said to be updated every sunday.
> But they are not. I just downloaded them again, but the newes changelog
> entry still
> is from 2004-04-29.
> I'll try to ask the eCosCentric people for more recent sources.

It would be nice to know what their schedule is...

Attached is the [very simple] patch that should fix your problem.  You
should not have to update anything else.

-- 
Gary Thomas <gary@mlbassoc.com>
MLB Associates

[-- Attachment #2: diffs --]
[-- Type: text/x-patch, Size: 1458 bytes --]

Index: redboot/current/src/flash.c
===================================================================
RCS file: /misc/cvsfiles/ecos/packages/redboot/current/src/flash.c,v
retrieving revision 1.65
retrieving revision 1.66
diff -u -5 -p -r1.65 -r1.66
--- redboot/current/src/flash.c	23 Apr 2004 20:38:18 -0000	1.65
+++ redboot/current/src/flash.c	5 May 2004 10:44:31 -0000	1.66
@@ -981,15 +981,15 @@ fis_load(int argc, char *argv[])
     struct fis_image_desc *img;
     CYG_ADDRESS mem_addr;
     bool mem_addr_set = false;
     bool show_cksum = false;
     struct option_info opts[3];
-#ifdef CYGPKG_REDBOOT_FIS_CRC_CHECK
+#if defined(CYGSEM_REDBOOT_FIS_CRC_CHECK)
     unsigned long cksum;
 #endif
     int num_options;
-#ifdef CYGPRI_REDBOOT_ZLIB_FLASH
+#if defined(CYGPRI_REDBOOT_ZLIB_FLASH) ||  defined(CYGSEM_REDBOOT_FIS_CRC_CHECK)
     bool decompress = false;
 #endif
     void *err_addr;
 
     init_opts(&opts[0], 'b', true, OPTION_ARG_TYPE_NUM, 
@@ -1064,11 +1064,11 @@ fis_load(int argc, char *argv[])
         load_address = mem_addr;
         load_address_end = mem_addr + img->data_length;
     }
     entry_address = (unsigned long)img->entry_point;
 
-#ifdef CYGPKG_REDBOOT_FIS_CRC_CHECK
+#ifdef CYGSEM_REDBOOT_FIS_CRC_CHECK
     cksum = cyg_crc32((unsigned char *)mem_addr, img->data_length);
     if (show_cksum) {
         diag_printf("Checksum: 0x%08lx\n", cksum);
     }
     // When decompressing, leave CRC checking to decompressor


[-- Attachment #3: Type: text/plain, Size: 148 bytes --]

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Winkler Andreas]

> Sorry, but I don't know anything about the eCosCentric snapshots.
> 
> Perhaps you can update via CVS?

Unfortunately our proxy server doesn't support the access to the CVS
repository.
So I downloaded the sources from
http://www.ecoscentric.com/devzone/snapshots.shtml 
which where said to be updated every sunday.
But they are not. I just downloaded them again, but the newes changelog
entry still
is from 2004-04-29.
I'll try to ask the eCosCentric people for more recent sources.

Regards
Andreas

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Gary Thomas]

On Wed, 2004-05-26 at 06:50, Winkler Andreas wrote:
> > From: Gary Thomas [mailto:gary@mlbassoc.com]
> > Also, make sure that your sources are new enough.  In 
> > .../redboot/current/ChangeLog:
> > 
> > 2004-05-05  Gary Thomas  <gary@mlbassoc.com>
> > 
> >         * src/flash.c (fis_load): CRC check on load was 
> > broken (the CDL
> >         option changed names)
> 
> The newest entry in my ChangeLog is from 2004-04-29.
> But I loaded this snapshot on 2004-05-19.
> I thought it would be built every Sunday.
> Does it help when i load it now again?

Sorry, but I don't know anything about the eCosCentric snapshots.

Perhaps you can update via CVS?

-- 
Gary Thomas <gary@mlbassoc.com>
MLB Associates


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Gary Thomas]

On Wed, 2004-05-26 at 06:52, Mark Salter wrote:
> On Wed, 2004-05-26 at 06:15, Winkler Andreas wrote:
> 
> > What i get is
> > 
> > RedBoot> loa -r -b 0x01008000 jffs2k
> > Using default protocol (TFTP)
> > Raw file loaded 0x01008000-0x010c27c7, assumed entry at 0x01008000
> > RedBoot> fi cr kernel
> > ... Erase from 0x50040000-0x50100000: ......
> > ... Program from 0x01008000-0x010c27c8 at 0x50040000: ......
> > ... Unlock from 0x50fe0000-0x51000000: .
> > ... Erase from 0x50fe0000-0x51000000: .
> > ... Program from 0x0ffe0000-0x10000000 at 0x50fe0000: .
> > ... Lock from 0x50fe0000-0x51000000: .
> > RedBoot> fi wr -f 0x50060000 -b 0x1008000 -l 100
> > * CAUTION * about to program FLASH
> >             at 0x50060000..0x5007ffff from 0x01008000 - continue (y/n)? y
> > ... Erase from 0x50060000-0x50080000: .
> > ... Program from 0x01008000-0x01028000 at 0x50060000: .
> > RedBoot> fi loa kernel
> > RedBoot> e
> > Using base address 0x01008000 and length 0x000ba7c8
> > Uncompressing Linux...........
> > 
> > invalid compressed format (err=2)
> > 
> >  -- System halted
> > 
> > Any idea, what could be my fault?
> 
> Could you explain what you are trying to accomplish by the "fi wr"
> command above? 
> 
> You download the kernel image and write it to flash from 0x50040000
> to 0x500fa7c8. Then you write the first 100 bytes of the kernel
> image to 0x50060000 which is in the middle of the kernel image you
> wrote in the previous step. So of course the kernel decompresses
> until it hits the area you corrupted, at which point it complains
> about the invalid compressed image.

I think he's just trying to test RedBoot's checking of the CRC,
by corrupting the data and showing that the Linux de-compressor
detects it properly.
-- 
Gary Thomas <gary@mlbassoc.com>
MLB Associates


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Mark Salter]

On Wed, 2004-05-26 at 06:15, Winkler Andreas wrote:

> What i get is
> 
> RedBoot> loa -r -b 0x01008000 jffs2k
> Using default protocol (TFTP)
> Raw file loaded 0x01008000-0x010c27c7, assumed entry at 0x01008000
> RedBoot> fi cr kernel
> ... Erase from 0x50040000-0x50100000: ......
> ... Program from 0x01008000-0x010c27c8 at 0x50040000: ......
> ... Unlock from 0x50fe0000-0x51000000: .
> ... Erase from 0x50fe0000-0x51000000: .
> ... Program from 0x0ffe0000-0x10000000 at 0x50fe0000: .
> ... Lock from 0x50fe0000-0x51000000: .
> RedBoot> fi wr -f 0x50060000 -b 0x1008000 -l 100
> * CAUTION * about to program FLASH
>             at 0x50060000..0x5007ffff from 0x01008000 - continue (y/n)? y
> ... Erase from 0x50060000-0x50080000: .
> ... Program from 0x01008000-0x01028000 at 0x50060000: .
> RedBoot> fi loa kernel
> RedBoot> e
> Using base address 0x01008000 and length 0x000ba7c8
> Uncompressing Linux...........
> 
> invalid compressed format (err=2)
> 
>  -- System halted
> 
> Any idea, what could be my fault?

Could you explain what you are trying to accomplish by the "fi wr"
command above? 

You download the kernel image and write it to flash from 0x50040000
to 0x500fa7c8. Then you write the first 100 bytes of the kernel
image to 0x50060000 which is in the middle of the kernel image you
wrote in the previous step. So of course the kernel decompresses
until it hits the area you corrupted, at which point it complains
about the invalid compressed image.

--Mark


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Winkler Andreas]

> From: Gary Thomas [mailto:gary@mlbassoc.com]
> Also, make sure that your sources are new enough.  In 
> .../redboot/current/ChangeLog:
> 
> 2004-05-05  Gary Thomas  <gary@mlbassoc.com>
> 
>         * src/flash.c (fis_load): CRC check on load was 
> broken (the CDL
>         option changed names)

The newest entry in my ChangeLog is from 2004-04-29.
But I loaded this snapshot on 2004-05-19.
I thought it would be built every Sunday.
Does it help when i load it now again?

Regards
Andreas

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Gary Thomas]

On Wed, 2004-05-26 at 06:15, Winkler Andreas wrote:
> > The 'fis load' with CRC checking function in RedBoot had suffered some
> > bitrot (the CDL option changed names, but the code did not).  This has
> > now been repaired, as seen below:
> > 
> > RedBoot> fi lo JFFS2                               
> > ** Warning - checksum failure.  stored: 0x78d127ea, computed: 
> > 0xf08ba2c8
> > 
> > -- 
> > Gary Thomas <gary@chez-thomas.org>
> 
> Hi,
> 
> I just built a new redboot for my IXDP425 evaluation board from Intel
> with an ecos snapshot from 19. May 2004.
> (Is it OK, that i get a lot of warnings, e. g. signed/unsigned mismatch
> while building?)
> The CRC check doesn't work yet.

Do you have the CRC package included & enabled in your build?
Look in your build directory, you should see something like this:
  [gthomas@hermes rt]$ grep CRC install/include/pkgconf/*
  install/include/pkgconf/crc.h:#ifndef CYGONCE_PKGCONF_CRC_H
  install/include/pkgconf/crc.h:#define CYGONCE_PKGCONF_CRC_H
  install/include/pkgconf/hal.h:#define CYGDBG_HAL_CRCTABLE_LOCATION RAM
  install/include/pkgconf/hal.h:#define CYGDBG_HAL_CRCTABLE_LOCATION_RAM
  install/include/pkgconf/redboot.h:#define CYGSEM_REDBOOT_FIS_CRC_CHECK 1
  install/include/pkgconf/system.h:#define CYGPKG_CRC current
  install/include/pkgconf/system.h:#define CYGPKG_CRC_current
  install/include/pkgconf/system.h:#define CYGNUM_CRC_VERSION_MAJOR CYGNUM_VERSION_CURRENT
  install/include/pkgconf/system.h:#define CYGNUM_CRC_VERSION_MINOR -1
  install/include/pkgconf/system.h:#define CYGNUM_CRC_VERSION_RELEASE -1

Also, make sure that your sources are new enough.  In .../redboot/current/ChangeLog:

2004-05-05  Gary Thomas  <gary@mlbassoc.com>

        * src/flash.c (fis_load): CRC check on load was broken (the CDL
        option changed names)


> What i get is
> 
> RedBoot> loa -r -b 0x01008000 jffs2k
> Using default protocol (TFTP)
> Raw file loaded 0x01008000-0x010c27c7, assumed entry at 0x01008000
> RedBoot> fi cr kernel
> ... Erase from 0x50040000-0x50100000: ......
> ... Program from 0x01008000-0x010c27c8 at 0x50040000: ......
> ... Unlock from 0x50fe0000-0x51000000: .
> ... Erase from 0x50fe0000-0x51000000: .
> ... Program from 0x0ffe0000-0x10000000 at 0x50fe0000: .
> ... Lock from 0x50fe0000-0x51000000: .
> RedBoot> fi wr -f 0x50060000 -b 0x1008000 -l 100
> * CAUTION * about to program FLASH
>             at 0x50060000..0x5007ffff from 0x01008000 - continue (y/n)? y
> ... Erase from 0x50060000-0x50080000: .
> ... Program from 0x01008000-0x01028000 at 0x50060000: .
> RedBoot> fi loa kernel
> RedBoot> e
> Using base address 0x01008000 and length 0x000ba7c8
> Uncompressing Linux...........
> 
> invalid compressed format (err=2)
> 
>  -- System halted
> 
> Any idea, what could be my fault?
> Which source has been changed to repair the bitrot bug?
> 
> Regards
> Andreas
-- 
Gary Thomas <gary@mlbassoc.com>
MLB Associates


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] [Fwd: How to enable redboots CRC checking]

[Andrew Lunn]

> Hi,
> 
> I just built a new redboot for my IXDP425 evaluation board from Intel
> with an ecos snapshot from 19. May 2004.
> (Is it OK, that i get a lot of warnings, e. g. signed/unsigned mismatch
> while building?)
> The CRC check doesn't work yet.

Are you sure the CRC check does not work? Look at the actual error message

> RedBoot> fi loa kernel
> RedBoot> e
> Using base address 0x01008000 and length 0x000ba7c8
> Uncompressing Linux...........
> 
> invalid compressed format (err=2)

I think this is an error from the compression library, not the CRC
checking code. What format did you compress the image with? Do you
have the correct endianness etc...

        Andrew

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Winkler Andreas]

> The 'fis load' with CRC checking function in RedBoot had suffered some
> bitrot (the CDL option changed names, but the code did not).  This has
> now been repaired, as seen below:
> 
> RedBoot> fi lo JFFS2                               
> ** Warning - checksum failure.  stored: 0x78d127ea, computed: 
> 0xf08ba2c8
> 
> -- 
> Gary Thomas <gary@chez-thomas.org>

Hi,

I just built a new redboot for my IXDP425 evaluation board from Intel
with an ecos snapshot from 19. May 2004.
(Is it OK, that i get a lot of warnings, e. g. signed/unsigned mismatch
while building?)
The CRC check doesn't work yet.
What i get is

RedBoot> loa -r -b 0x01008000 jffs2k
Using default protocol (TFTP)
Raw file loaded 0x01008000-0x010c27c7, assumed entry at 0x01008000
RedBoot> fi cr kernel
... Erase from 0x50040000-0x50100000: ......
... Program from 0x01008000-0x010c27c8 at 0x50040000: ......
... Unlock from 0x50fe0000-0x51000000: .
... Erase from 0x50fe0000-0x51000000: .
... Program from 0x0ffe0000-0x10000000 at 0x50fe0000: .
... Lock from 0x50fe0000-0x51000000: .
RedBoot> fi wr -f 0x50060000 -b 0x1008000 -l 100
* CAUTION * about to program FLASH
            at 0x50060000..0x5007ffff from 0x01008000 - continue (y/n)? y
... Erase from 0x50060000-0x50080000: .
... Program from 0x01008000-0x01028000 at 0x50060000: .
RedBoot> fi loa kernel
RedBoot> e
Using base address 0x01008000 and length 0x000ba7c8
Uncompressing Linux...........

invalid compressed format (err=2)

 -- System halted

Any idea, what could be my fault?
Which source has been changed to repair the bitrot bug?

Regards
Andreas

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Winkler Andreas]

> Unfortunately my proxy doesn't allow me to access the cvs server.
> Is cvs the only way to get a recent version or are there tar files?

Please excuse my request.
I found the weekly snapshots.

Regards
Andreas

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Gary Thomas]

On Wed, 2004-05-19 at 04:02, Winkler Andreas wrote:
> > From: Gary Thomas [mailto:gary@chez-thomas.org]
> > The 'fis load' with CRC checking function in RedBoot had suffered some
> > bitrot (the CDL option changed names, but the code did not).  This has
> > now been repaired, as seen below:
> 
> Finally I've got time to check an new Redboot version from Intel for my
> IXDP425 evaluation board and in this version the CRC checking doesn't work.
> 
> Unfortunately my proxy doesn't allow me to access the cvs server.
> Is cvs the only way to get a recent version or are there tar files?
> 

Look at http://www.ecoscentric.com/devzone/snapshots.shtml

-- 
Gary Thomas <gary@mlbassoc.com>
MLB Associates


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] [Fwd: How to enable redboots CRC checking]

[Winkler Andreas]

> From: Gary Thomas [mailto:gary@chez-thomas.org]
> The 'fis load' with CRC checking function in RedBoot had suffered some
> bitrot (the CDL option changed names, but the code did not).  This has
> now been repaired, as seen below:

Finally I've got time to check an new Redboot version from Intel for my
IXDP425 evaluation board and in this version the CRC checking doesn't work.

Unfortunately my proxy doesn't allow me to access the cvs server.
Is cvs the only way to get a recent version or are there tar files?

Regards
Andreas

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] [Fwd: How to enable redboots CRC checking]

[Gary Thomas]

On Wed, 2004-05-05 at 04:13, Jonathan Larmour wrote:
> This was misaddressed...

The 'fis load' with CRC checking function in RedBoot had suffered some
bitrot (the CDL option changed names, but the code did not).  This has
now been repaired, as seen below:

RedBoot> lo -r -b 0x100000 -h 192.168.1.125 test.JFFS2
Using default protocol (TFTP)
Raw file loaded 0x00100000-0x00298483, assumed entry at 0x00100000
RedBoot> fi cr JFFS2
An image named 'JFFS2' exists - continue (y/n)? y
RedBoot> fi lo JFFS2
RedBoot> fi li      
Name              FLASH addr  Mem addr    Length      Entry point
RedBoot           0xC0000000  0x00100000  0x00080000  0x00100100
JFFS2             0xC0080000  0x00100000  0x00400000  0x00100000
FIS directory     0xC7FFC000  0xC7FFC000  0x00004000  0x00000000
RedBoot> fi wr -f 0xc0080000 -b 0x200000 -l 0x80000
* CAUTION * about to program FLASH
            at 0xc0080000..0xc00fffff from 0x00200000 - continue (y/n)? y
RedBoot> fi lo JFFS2                               
** Warning - checksum failure.  stored: 0x78d127ea, computed: 0xf08ba2c8

-- 
Gary Thomas <gary@chez-thomas.org>


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[ECOS] [Fwd: How to enable redboots CRC checking]

[Jonathan Larmour]

[-- Attachment #1: Type: text/plain, Size: 179 bytes --]

This was misaddressed...
-- 
eCosCentric    http://www.eCosCentric.com/    The eCos and RedBoot experts
--["No sense being pessimistic, it wouldn't work anyway"]-- Opinions==mine

[-- Attachment #2: How to enable redboots CRC checking --]
[-- Type: message/rfc822, Size: 3385 bytes --]

From: Winkler Andreas <andreas.aw.winkler@siemens.com>
To: "'ecos-discuss-owner@ecos.sourceware.org'" <ecos-discuss-owner@ecos.sourceware.org>
Subject: How to enable redboots CRC checking
Date: Wed, 5 May 2004 09:28:40 +0200
Message-ID: <BCB494E8B7BF2D40B56D1164C67BC2A801B16872@kher443a.ww004.siemens.net>

Hi,

some weeks ago I asked about the CRC checking of redboot. From Thomas Gary i
got the following answer:

"When images are written to the FIS, a CRC will be calculated and saved with
the image. When you read the image into RAM ("fis load"), the CRC is
verified and if it fails, then the entry point address is suppressed. This
is normally used to prevent the "go" command from executing."

Now I made a test. for the CRC checking.
With Redboot I wrote a Linux kernel to flash (fis create). I could load that
kernel to RAM and start it.
With Linux I wrote something into that flash image.
Then I booted Redboot. Again I loaded the kernel image to RAM and could
start it without any fault message.
Of course the kernel was not runnable and crashed.

I had expected a fault message when trying to run the corrumpted kernel
image.
What could be the reason that it doesn't work?

Regards
Andreas


[-- Attachment #3: Type: text/plain, Size: 148 bytes --]

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss