[U-Boot] [PATCH] CFI: geometry reversal for STMicro M29W320DT
Follow up to the flash_fixup_stm to fix geometry reversal on STMicro M29W320ET flash chip. The M29W320DT has 4 erase region.
Signed-off-by: Richard Retanubun RichardRetanubun@RuggedCom.com
--- My baseline is 2009.03-rc1. If there are STMicro/Numonyx guys on this mailing list, can you help provide a more 'generic' way of detection? this way the exception list does not have to grow a lot.
drivers/mtd/cfi_flash.c | 5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/mtd/cfi_flash.c b/drivers/mtd/cfi_flash.c index a66feac..391d169 100644 --- a/drivers/mtd/cfi_flash.c +++ b/drivers/mtd/cfi_flash.c @@ -1806,8 +1806,9 @@ static void flash_fixup_stm(flash_info_t *info, struct cfi_qry *qry) if (qry->num_erase_regions > 1) { /* reverse geometry if top boot part */ if (info->cfi_version < 0x3131) { - /* CFI < 1.1, guess by device id (only M29W320ET now) */ - if (info->device_id == 0x2256) { + /* CFI < 1.1, guess by device id (M29W320{DT,ET} only) */ + if (info->device_id == 0x22CA || + info->device_id == 0x2256) { cfi_reverse_geometry(qry); } }
On Friday 06 March 2009, Richard Retanubun wrote:
Follow up to the flash_fixup_stm to fix geometry reversal on STMicro M29W320ET flash chip. The M29W320DT has 4 erase region.
Signed-off-by: Richard Retanubun RichardRetanubun@RuggedCom.com
My baseline is 2009.03-rc1. If there are STMicro/Numonyx guys on this mailing list, can you help provide a more 'generic' way of detection? this way the exception list does not have to grow a lot.
Did you take a look at the Linux driver to see how this is handled there? Maybe they have a more "generic" solution...
Best regards, Stefan
===================================================================== DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@denx.de =====================================================================
Stefan Roese wrote:
On Friday 06 March 2009, Richard Retanubun wrote:
Follow up to the flash_fixup_stm to fix geometry reversal on STMicro M29W320ET flash chip. The M29W320DT has 4 erase region.
Signed-off-by: Richard Retanubun RichardRetanubun@RuggedCom.com
My baseline is 2009.03-rc1. If there are STMicro/Numonyx guys on this mailing list, can you help provide a more 'generic' way of detection? this way the exception list does not have to grow a lot.
Did you take a look at the Linux driver to see how this is handled there? Maybe they have a more "generic" solution...
Hi Stefan,
Thanks for the comments. I took a quick look around /drivers/mtd, but can't find a similar geometry reversal.
Unfortunately, I found this in:
/driver/mtd/chips/cfi_cmdset_0001.c [snip] { CFI_MFR_ST, 0x00ba, /* M28W320CT */ fixup_st_m28w320ct, NULL }, { CFI_MFR_ST, 0x00bb, /* M28W320CB */ fixup_st_m28w320cb, NULL }, [/snip]
By no means am I a CFI expert, but my impression is, if there has to be a part number specific fixup function for any reason, then it is a good indication that there are just some part specific peculiarity.
What utilities exist in Linux-world to do the equivalent of "flinfo" in u-boot, just so I can see if Linux has a more generic way.
And just in case it matters, our board with these flash is MMU-less, so if there is a different tool for uCLinux that will help too :)
- Richard
On Friday 06 March 2009, Richard Retanubun wrote:
My baseline is 2009.03-rc1. If there are STMicro/Numonyx guys on this mailing list, can you help provide a more 'generic' way of detection? this way the exception list does not have to grow a lot.
Did you take a look at the Linux driver to see how this is handled there? Maybe they have a more "generic" solution...
Hi Stefan,
Thanks for the comments. I took a quick look around /drivers/mtd, but can't find a similar geometry reversal.
Unfortunately, I found this in:
/driver/mtd/chips/cfi_cmdset_0001.c [snip] { CFI_MFR_ST, 0x00ba, /* M28W320CT */ fixup_st_m28w320ct, NULL }, { CFI_MFR_ST, 0x00bb, /* M28W320CB */ fixup_st_m28w320cb, NULL }, [/snip]
By no means am I a CFI expert, but my impression is, if there has to be a part number specific fixup function for any reason, then it is a good indication that there are just some part specific peculiarity.
Right.
What utilities exist in Linux-world to do the equivalent of "flinfo" in u-boot, just so I can see if Linux has a more generic way.
At this level you really need to "read" the code and/or use the embedded debug output (if available) or add some new code to see what's going on.
And just in case it matters, our board with these flash is MMU-less, so if there is a different tool for uCLinux that will help too :)
I don't think that uCLinux has different "tools" here.
Best regards, Stefan
===================================================================== DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@denx.de =====================================================================
Hi Stefan,
Just following up on this patch, the only other action I've seen on with regards to this is to fake the device identification from top-boot to bottom boot, but I think this is more 'masking' the problem and I don't have enough visibility to see if it is a safe thing to do.
In any case, even the reversal from top-boot to bottom-boot still needs identification on a chip-per-chip basis.
In any case, I'd like to know if this patch is [ACK|NACK]?
Many thanks for your time.
- Richard
Hi Richard,
On Thursday 19 March 2009, Richard Retanubun wrote:
Just following up on this patch, the only other action I've seen on with regards to this is to fake the device identification from top-boot to bottom boot, but I think this is more 'masking' the problem and I don't have enough visibility to see if it is a safe thing to do.
In any case, even the reversal from top-boot to bottom-boot still needs identification on a chip-per-chip basis.
In any case, I'd like to know if this patch is [ACK|NACK]?
It's an ACK. Sorry, I forgot to take care of this patch. I'll apply it in quickly and ask Wolfgang to pull shortly so that this fix can be included in the upcoming 2009-03 release.
Thanks.
Best regards, Stefan
===================================================================== DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@denx.de =====================================================================
participants (2)
-
Richard Retanubun -
Stefan Roese