[U-Boot] [BUG] cfi_flash can't work correctly on latest uboot

[Stefan Roese]

Hi Bob,

sorry for the late reply.

On Wednesday 02 November 2011 08:20:03 Bob Liu wrote:
> Hi, folks
> 
> I think this patch make cfi_flash can't work correctly.
> http://git.denx.de/?p=u-boot.git;a=commitdiff;h=df4e813b72bf07d9026b00455f5
> e7dffd694ae48
> 
> --- a/drivers/mtd/cfi_flash.c
> +++ b/drivers/mtd/cfi_flash.c
> @@ -1426,6 +1426,11 @@ int flash_real_protect (flash_info_t * info,
> long sector, int prot)
>  #endif
>         };
> 
> +       /*
> +        * Flash needs to be in status register read mode for
> +        * flash_full_status_check() to work correctly
> +        */
> +       flash_write_cmd(info, sector, 0, FLASH_CMD_READ_STATUS);
> 
> --------------------------------------
> without this line, flash can't be recognized. So it's correct here.
> 
> U-Boot 2011.09-svn2735 (ADI-2012R1-pre) (Nov 02 2011 - 15:05:23)

<snip>

> --------------------------------------
> 
>         if ((retcode =
>              flash_full_status_check (info, sector, info->erase_blk_tout,
>                                       prot ? "protect" : "unprotect")) ==
> 0) { @@ -1975,6 +1980,13 @@ ulong flash_get_size (phys_addr_t base, int banknum)
>                                 case CFI_CMDSET_INTEL_PROG_REGIONS:
>                                 case CFI_CMDSET_INTEL_EXTENDED:
>                                 case CFI_CMDSET_INTEL_STANDARD:
> +                                       /*
> +                                        * Set flash to read-id mode. Otherwise
> +                                        * reading protected status is not
> +                                        * guaranteed.
> +                                        */
> +                                       flash_write_cmd(info, sect_cnt, 0,
> +                                                       FLASH_CMD_READ_ID);
> ---------------------------------------------------------------------------
> - But with this two lines. flash ops are not correct.
> 
> I used old uboot version to read out data from flash.
> Then tftpboot and run with latest uboot version, you can see that the read
> out data from flash is different.
> And without set flash to read-id mode here, it works fine.
>
> U-Boot 2010.06 (ADI-2010R1-RC2) (Oct 20 2010 - 04:37:34)
> 
> CPU:   ADSP bf526-0.0 (Detected Rev: 0.2) (parallel flash boot)
> Board: ADI BF526 EZ-Board board
>        Support: http://blackfin.uclinux.org/
> Clock: VCO: 400 MHz, Core: 400 MHz, System: 80 MHz
> RAM:   64 MiB
> Flash: 4 MiB
> In:    serial
> Out:   serial
> Err:   serial
> KGDB:  [on serial] ready
> Net:   bfin_mac
> Hit any key to stop autoboot:  0
> bfin>
> bfin> md 0x20100000
> 20100000: 56190527 0b223987 d60baf4e 2e871c00    '..V.9".N.......
> 20100010: 00100000 c42c1c00 a86adb94 01021005    ......,...j.....
> 20100020: 32356662 2e302d36 2e332d30 2d382e30    bf526-0.0-3.0.8-
> 20100030: 2d494441 31313032 702d3152 732d6572    ADI-2011R1-pre-s
> 20100040: 00088b1f 00000000 fdc40302 4514780b    .............x.E
> 20100050: 570e38d6 3d33264f 24f48743 0649d303    .8.WO&3=C..$..I.
> 20100060: 93080932 834e9970 0931141a 648443a0    2...p.N...1..C.d
> 20100070: e0804212 12048762 56501a34 086bb2f1    .B..b...4.PV..k.
> 20100080: 66e0888a d9782808 d165049d 75049945    ...f.(x...e.E..u
> 20100090: 31415041 ba02cba2 aebbb2ae 8093d34b    APA1........K...
> 201000a0: 682f1a8a 9cffceb8 105ceeea 7f7df77d    ../h......\.}.}.
> 201000b0: 3c7cf3df 3aea984f 9d554ea7 4e753aaa    ..|<O..:.NU..:uN
> 201000c0: b1a94e5d b13ae91e cc436dbe 749b42c2    ]N....:..mC..B.t
> 201000d0: fa449283 2aa2ec29 a82e0846 b7089ae5    ..D.)..*F.......
> 201000e0: cc1f9dc4 a9920651 113e5122 b1122213    ....Q..."Q>.."..
> 201000f0: 308f15a9 b6496a6a 12562598 49a8c8af    ...0jjI..%V....I
> 
> bfin> tftpboot 0x1000 u-boot.bin
> Using bfin_mac device
> TFTP from server 10.100.4.174; our IP address is 10.100.4.50
> Filename 'u-boot.bin'.
> Load address: 0x1000
> Loading: ###############
> done
> Bytes transferred = 213828 (34344 hex)
> bfin> go 0x1000
> ## Starting application at 0x00001000 ...
> 
> 
> U-Boot 2011.09-svn2735 (ADI-2012R1-pre) (Nov 02 2011 - 15:15:05)
> 
> CPU:   ADSP bf526-0.0 (Detected Rev: 0.2) (parallel flash boot)
> Board: ADI BF526 EZ-Board board
>        Support: http://blackfin.uclinux.org/
> Clock: VCO: 400 MHz, Core: 400 MHz, System: 80 MHz
> RAM:   64 MiB
> Flash: 4 MiB
> In:    serial
> Out:   serial
> Err:   serial
> KGDB:  [on serial] ready
> Net:   bfin_mac
> Hit any key to stop autoboot:  0
> bfin> md 0x20100000
> 20100000: 88150020 ffff0001 bfcfffff ffffffff     ...............
> 20100010: ffffffff ffffffff ffffffff ffffffff    ................
> 20100020: 00520051 00030059 00390000 00000000    Q.R.Y.....9.....

Hmmm. Strange. Your flash still seems to be in read-id mode after
initial detection. But from looking at flash_get_size(),
the following flash reset command should bring it back to read-array
mode:

	flash_write_cmd (info, 0, 0, info->cmd_reset);

Did you check that cmd_reset is set correctly on your hardware (AMD vs.
Intel)? Which flash chip are you using?

Thanks,
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 at denx.de