Not my best posting.
The issue is the standalone boot FAILS if the environment variable is not set. All of the default settings are correct. If I simple execute the "saveenv" command once, subsequent boots work fine. This issue will only occur if starting with a NAND Flash that has been entirely erased, then reprogrammed using CCS tools.
On Fri, 24 Apr 2009 14:37:41 -0500
"Bob Simning" <[email protected]> wrote:
>
> Hello,
>
> I have an issue with u-boot-1.2.0 using NAND FLASH when
>booting a standalone system w/o the environment block
>programmed. If the environment block has been updated
>with a "saveenv" command, the standalone booting process
>succeeds.
>
> The default behavior within env_nand.c,
>env_relocate_spec(), when the reading of the environment
>block fails should be to continue using the values
>specified within the default environment. What occurs
>when booting a stand alone system is a failure when
>reading the ramdisk. If I modify env_relocate_spec() so
>the logic performs a read of a good block subsequent to
>the failure of the reading of the environmental block,
>the reading of the RAMDISK works.
>
> I tested this behavior both by modifying the function
>and by performing read/writes at the command line in
>u-boot and the behavior appears consistent.
>
> My bootargs were:"nand read.e 0x82000000 0x1860000
>0x400000; nboot 0x80700000 0 0xa60000; bootm".
>
> If the environment has been saved with a 'saveenv"
>command, the standalone boot works. If the environment
>has not been saved, the part of the command reading
>RAMDISK fails due to "too many erros" (mispelling part
>of the message). The subsequent "nboot ..." works,
>eventually panicing when trying to mount the RAMDISK that
>failed to be read earlier.
>
> Simply put, with version 1.2.0 u-boot, the environment
>block has to be programmed for the stand alone boot
>process to work.
>
> Thanks,
> Bob
The issue is the standalone boot FAILS if the environment variable is not set. All of the default settings are correct. If I simple execute the "saveenv" command once, subsequent boots work fine. This issue will only occur if starting with a NAND Flash that has been entirely erased, then reprogrammed using CCS tools.
On Fri, 24 Apr 2009 14:37:41 -0500
"Bob Simning" <[email protected]> wrote:
>
> Hello,
>
> I have an issue with u-boot-1.2.0 using NAND FLASH when
>booting a standalone system w/o the environment block
>programmed. If the environment block has been updated
>with a "saveenv" command, the standalone booting process
>succeeds.
>
> The default behavior within env_nand.c,
>env_relocate_spec(), when the reading of the environment
>block fails should be to continue using the values
>specified within the default environment. What occurs
>when booting a stand alone system is a failure when
>reading the ramdisk. If I modify env_relocate_spec() so
>the logic performs a read of a good block subsequent to
>the failure of the reading of the environmental block,
>the reading of the RAMDISK works.
>
> I tested this behavior both by modifying the function
>and by performing read/writes at the command line in
>u-boot and the behavior appears consistent.
>
> My bootargs were:"nand read.e 0x82000000 0x1860000
>0x400000; nboot 0x80700000 0 0xa60000; bootm".
>
> If the environment has been saved with a 'saveenv"
>command, the standalone boot works. If the environment
>has not been saved, the part of the command reading
>RAMDISK fails due to "too many erros" (mispelling part
>of the message). The subsequent "nboot ..." works,
>eventually panicing when trying to mount the RAMDISK that
>failed to be read earlier.
>
> Simply put, with version 1.2.0 u-boot, the environment
>block has to be programmed for the stand alone boot
>process to work.
>
> Thanks,
> Bob
_______________________________________________ Davinci-linux-open-source mailing list [email protected] http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
