OMAP platform specific patch.
- Adjust the frame size for DMA transfers.
Signed-off-by: Anderson Briglia <anderson.briglia <at> indt.org.br>
Signed-off-by: Carlos Eduardo Aguiar <carlos.aguiar <at> indt.org.br>
Index: linux-omap-2.6.git/drivers/mmc/omap.c
===================================================================
--- linux-omap-2.6.git.orig/drivers/mmc/omap.c 2006-11-22 09:07:25.000000000 -0400
+++ linux-omap-2.6.git/drivers/mmc/omap.c 2006-11-22 09:19:03.000000000 -0400
@@ -629,6 +629,14 @@ mmc_omap_prepare_dma(struct mmc_omap_hos
data_addr = host->phys_base + OMAP_MMC_REG_DATA;
frame = data->blksz;
+
+#ifdef CONFIG_MMC_PASSWORDS
+ /* MMC LOCK/UNLOCK: Do frame size multiple of two. This is
+ * needed for DMA transfers to work properly, once
+ * the block size depends on MMC password length.
+ */
+ frame += frame&0x1;
+#endif
count = sg_dma_len(sg);
if ((data->blocks == 1) && (count > (data->blksz)))
Anderson Briglia wrote:
> OMAP platform specific patch.
> - Adjust the frame size for DMA transfers.
>
> Signed-off-by: Anderson Briglia <anderson.briglia <at> indt.org.br>
> Signed-off-by: Carlos Eduardo Aguiar <carlos.aguiar <at> indt.org.br>
>
This looks like a generic patch for OMAP and shouldn't be in this set.
> Index: linux-omap-2.6.git/drivers/mmc/omap.c
> ===================================================================
> --- linux-omap-2.6.git.orig/drivers/mmc/omap.c 2006-11-22
> 09:07:25.000000000 -0400
> +++ linux-omap-2.6.git/drivers/mmc/omap.c 2006-11-22
> 09:19:03.000000000 -0400
> @@ -629,6 +629,14 @@ mmc_omap_prepare_dma(struct mmc_omap_hos
>
> data_addr = host->phys_base + OMAP_MMC_REG_DATA;
> frame = data->blksz;
> +
> +#ifdef CONFIG_MMC_PASSWORDS
> + /* MMC LOCK/UNLOCK: Do frame size multiple of two. This is
> + * needed for DMA transfers to work properly, once
> + * the block size depends on MMC password length.
> + */
> + frame += frame&0x1;
> +#endif
> count = sg_dma_len(sg);
>
> if ((data->blocks == 1) && (count > (data->blksz)))
>
Now this you're going to have to explain to me. Especially the part
where why this is specific to the lock commands.
--
-- Pierre Ossman
Linux kernel, MMC maintainer http://www.kernel.org
PulseAudio, core developer http://pulseaudio.org
rdesktop, core developer http://www.rdesktop.org
ext Pierre Ossman wrote:
> Anderson Briglia wrote:
>> OMAP platform specific patch.
>> - Adjust the frame size for DMA transfers.
>>
>> Signed-off-by: Anderson Briglia <anderson.briglia <at> indt.org.br>
>> Signed-off-by: Carlos Eduardo Aguiar <carlos.aguiar <at> indt.org.br>
>>
>
> This looks like a generic patch for OMAP and shouldn't be in this set.
This patch is needed only for lock/unlock commands. So, it's necessary to
make MMC omap works when using that feature. It's not a generic patch.
But I can take off this one from the series and send after (if) the series
is integrated.
>
>> Index: linux-omap-2.6.git/drivers/mmc/omap.c
>> ===================================================================
>> --- linux-omap-2.6.git.orig/drivers/mmc/omap.c 2006-11-22
>> 09:07:25.000000000 -0400
>> +++ linux-omap-2.6.git/drivers/mmc/omap.c 2006-11-22
>> 09:19:03.000000000 -0400
>> @@ -629,6 +629,14 @@ mmc_omap_prepare_dma(struct mmc_omap_hos
>>
>> data_addr = host->phys_base + OMAP_MMC_REG_DATA;
>> frame = data->blksz;
>> +
>> +#ifdef CONFIG_MMC_PASSWORDS
>> + /* MMC LOCK/UNLOCK: Do frame size multiple of two. This is
>> + * needed for DMA transfers to work properly, once
>> + * the block size depends on MMC password length.
>> + */
>> + frame += frame&0x1;
>> +#endif
>> count = sg_dma_len(sg);
>>
>> if ((data->blocks == 1) && (count > (data->blksz)))
>>
>
> Now this you're going to have to explain to me. Especially the part
> where why this is specific to the lock commands.
frame depends on data->blksz. When we were using data->blksz_bits everything was
ok because we always had a multiple of 16 bits (2 bytes). Once a pwd can has a size
not multiple of 2, the value must be rounded.
According to MMC OMAP Technical Reference Manual, because of each DMA transfer is of
equal size, it is necessary to have the block size of the transfer be a multiple of
the DMA write access size (which is 2 bytes).
I'm not a DMA expert, If I misunderstand something, please, correct me.
Best Regards,
Anderson Briglia
Anderson Briglia wrote:
>
> This patch is needed only for lock/unlock commands. So, it's necessary to
> make MMC omap works when using that feature. It's not a generic patch.
> But I can take off this one from the series and send after (if) the
> series
> is integrated.
>
The patches are marked "[RFC]" which I interpret as that I shouldn't
merge it. Is this incorrect?
>
> frame depends on data->blksz. When we were using data->blksz_bits
> everything was
> ok because we always had a multiple of 16 bits (2 bytes). Once a pwd
> can has a size
> not multiple of 2, the value must be rounded.
> According to MMC OMAP Technical Reference Manual, because of each DMA
> transfer is of
> equal size, it is necessary to have the block size of the transfer be
> a multiple of
> the DMA write access size (which is 2 bytes).
>
This sounds very generic and not something that is specific to the
password command.
Rgds
--
-- Pierre Ossman
Linux kernel, MMC maintainer http://www.kernel.org
PulseAudio, core developer http://pulseaudio.org
rdesktop, core developer http://www.rdesktop.org
ext Pierre Ossman wrote:
> Anderson Briglia wrote:
>> This patch is needed only for lock/unlock commands. So, it's necessary to
>> make MMC omap works when using that feature. It's not a generic patch.
>> But I can take off this one from the series and send after (if) the
>> series
>> is integrated.
>>
>
> The patches are marked "[RFC]" which I interpret as that I shouldn't
> merge it. Is this incorrect?
Yes, you're right. But I believe this code is almost "ready" to be applied, do
you agree?
The next series I'll send without the "[RFC]" mark, what do you think?
>
>> frame depends on data->blksz. When we were using data->blksz_bits
>> everything was
>> ok because we always had a multiple of 16 bits (2 bytes). Once a pwd
>> can has a size
>> not multiple of 2, the value must be rounded.
>> According to MMC OMAP Technical Reference Manual, because of each DMA
>> transfer is of
>> equal size, it is necessary to have the block size of the transfer be
>> a multiple of
>> the DMA write access size (which is 2 bytes).
>>
>
> This sounds very generic and not something that is specific to the
> password command.
I'm still investigating where is the problem.
Actually this patch will not be included on the next series, as you
suggested. But it is still needed to make MMC lock/unlock works for
OMAP.
Best regards,
Anderson Briglia