MMC_SWITCH command doesn't update the ext_csd data structure

Commit Message

Dirk Behme Sept. 3, 2013, 11:04 a.m. UTC

Hi,

using the MMC_SWITCH command via the ioctl to write registers of the 
EXT_CSD, it looks to us that in this case the internal ext_csd data 
structure isn't updated. Resulting in a mismatch of what the ext_csd 
data structure contains and what's written to the real hardware.

We are using the mmc utils [1] to e.g. enable the bootpart

mmc bootpart enable 1 0 /dev/mmcblk1

This correctly writes to the EXT_CSD structure of the device, but 
doesn't update

card->ext_csd.part_config

Issuing additional MMC_SWITCH commands afterwards use the then (wrong) 
card->ext_csd.part_config. Resulting in a undo of the mmc bootpart enable.

What we think might be necessary here is anything like

Best regards

Dirk

[1] https://git.kernel.org/cgit/linux/kernel/git/cjb/mmc-utils.git/

--
To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Chris Ball Sept. 3, 2013, 11:26 a.m. UTC | #1

Hi,

On Tue, Sep 03 2013, Dirk Behme wrote:
> +       if ((cmd.opcode == MMC_SWITCH) && ((cmd.arg >> 24) & 0x3)) {
> +               /* In case the IOCTL has modified the EXT_CSD, update
> it, i.e. re-read the EXT_CSD */
> +		mmc_update_ext_csd(card->ext_csd);
> +       }
> +

Your analysis looks good, please go ahead and submit a full patch.
Thanks!

- Chris.

Dirk Behme Sept. 3, 2013, 11:34 a.m. UTC | #2

On 03.09.2013 13:26, Chris Ball wrote:
> Hi,
>
> On Tue, Sep 03 2013, Dirk Behme wrote:
>> +       if ((cmd.opcode == MMC_SWITCH) && ((cmd.arg >> 24) & 0x3)) {
>> +               /* In case the IOCTL has modified the EXT_CSD, update
>> it, i.e. re-read the EXT_CSD */
>> +		mmc_update_ext_csd(card->ext_csd);
>> +       }
>> +
>
> Your analysis looks good, please go ahead and submit a full patch.

The issue here is that something like mmc_update_ext_csd() doesn't 
exist, yet?

So we have to implement a mmc_update_ext_csd() in core/mmc.c and call it 
from mmc_blk_ioctl_cmd() in card/block.c?

Or any better proposal?

Thanks

Dirk
--
To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

--- a/drivers/mmc/card/block.c
+++ b/drivers/mmc/card/block.c
@@ -542,11 +544,17 @@  static int mmc_blk_ioctl_cmd(struct block_device 
*bdev,
                                         __func__, status, err);
         }

+       if ((cmd.opcode == MMC_SWITCH) && ((cmd.arg >> 24) & 0x3)) {
+               /* In case the IOCTL has modified the EXT_CSD, update 
it, i.e. re-read the EXT_CSD */
+		mmc_update_ext_csd(card->ext_csd);
+       }
+
  cmd_rel_host:
         mmc_release_host(card->host);

?

Opinions?