On Wed, 10 Nov 2010, Simon Horman wrote:
> On Wed, Nov 10, 2010 at 08:00:49PM +0900, Paul Mundt wrote:
> > On Wed, Nov 10, 2010 at 07:27:44PM +0900, Simon Horman wrote:
> > > Use pr_fmt to set the prefix for pr_ messages.
> > > I believe this method is common in other
> > > source files in the ker
> > Removing the dev.parent-branch will break some PCI-based solutions.
> Hmm, I suspected this :-(. Unfortunately I need to not pass the parent
> for the problem described in the patch.
> How to proceed? Do I have to re-introduce the sdhci-stm driver?
The current mainline code does not work for y
Have to add DMA slave configuration to enable DMA for the sh7372 MMCIF
controller.
Signed-off-by: Guennadi Liakhovetski
---
arch/arm/mach-shmobile/board-ap4evb.c | 10 ++
1 files changed, 10 insertions(+), 0 deletions(-)
diff --git a/arch/arm/mach-shmobile/board-ap4evb.c
b/arch/arm/m
The MMCIF controller on sh-mobile platforms can use the DMA controller for data
transfers. Interface to the SH dmaengine driver to enable DMA.
Signed-off-by: Guennadi Liakhovetski
---
...This has been a nice exception: just one missing piece of the datasheet
puzzle had to be guessed and it work
Currently shdma only works, if it manages to allocate DMA descriptors for all
sg elements during the prepare stage. Allocation is carried out from a
static pool of 32 DMA descriptors. This is not enough for, e.g., MMC, where up
to 61 sg elements can be sent at once. Ideally we should be able to reu
Replace pr_* with respective dev_*, sort headers alphabetically, remove an
unused struct member, superfluous variable initialisations and type-casts.
Signed-off-by: Guennadi Liakhovetski
---
drivers/mmc/host/sh_mmcif.c | 63 --
1 files changed, 30 insert
Add DMA slave IDs and slave definitions for MMCIF on sh7372.
Signed-off-by: Guennadi Liakhovetski
---
arch/arm/mach-shmobile/include/mach/sh7372.h |2 ++
arch/arm/mach-shmobile/setup-sh7372.c| 10 ++
2 files changed, 12 insertions(+), 0 deletions(-)
diff --git a/arch/arm/m
In sh_mmcif.c an event is used as a completion, switch over. When a
wait_for_completion*_timeout() returns, it suffices to check the remaining
time, setting an additional flag before waking up the waiting task only reduces
the race window, but does not eliminate it. This patch switches the driver t
This patch series slightly cleans up the sh_mmcif MMC driver and adds a
DMA capability to it. One of the patches is not directly related, but is
required for a proper DMA functionality.
Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-tech
Hi Ohad,
On 11/10/2010 9:28 PM, Ohad Ben-Cohen wrote:
>
> Hi Giuseppe,
>
> On Wed, Nov 10, 2010 at 5:28 PM, Giuseppe CAVALLARO
> wrote:
> > HC driver will be able to use the pm_flags to
> > undestand if the system can be woken-up by the driver.
> > So the mmc_suspend_host hasn't to reset this fie
Hi Wolfram
On 11/10/2010 4:43 PM, Wolfram Sang wrote:
> Hi,
>
> thanks for the work in general, just...
>
>> -if (pdev->dev.parent)
>> -host = sdhci_alloc_host(pdev->dev.parent, 0);
>> -else
>> -host = sdhci_alloc_host(&pdev->dev, 0);
>> -
>> +host = sdhci_alloc
Hi Park,
Though I tried to patch yours to "sdhci.c" of linux-2.6.36,
but it seems your patch has already patched in linux-2.6.36.
Using this linux, MMC card is not recognized.
Inserting the MMC card, I saw the following erro message.
[ 11.035441] sdhci: Secure Digital Host Controller Interfac
We should rename sdhci-pxa.c to sdhci-mmp2.c and adjust Kconfig and the
makefile accordingly since it is clear code is clearly processor dependent
or
The additional changes to support the pxa family of chips should be posted to
the web.
On Nov 10, 2010, at 6:08 PM, zhangfei gao wrote:
> On
On Wed, Nov 10, 2010 at 4:58 PM, Philip Rakity wrote:
>
>
>
>> +#define SD_CLOCK_AND_BURST_SIZE_SETUP 0x10A
>
> This offset is only for MMP2. On other pxa processors use a different
> offset. sdhci-pxa.c needs changing to support other pxa soc chips.
The first stage is only consider
Hi Giuseppe,
On Wed, Nov 10, 2010 at 5:28 PM, Giuseppe CAVALLARO
wrote:
> HC driver will be able to use the pm_flags to
> undestand if the system can be woken-up by the driver.
> So the mmc_suspend_host hasn't to reset this field
> in the host structure.
>
> Signed-off-by: Giuseppe Cavallaro
> -
The following patch is on top of v8 to support hardware clock gating. I do not
have hardware to test
this code until next week.
Comments welcome. Will revise once it is tested.
Philip
diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
index 8bf542c..5339fa8 100644
--- a/driver
On Wed, Nov 10, 2010 at 6:25 PM, Linus Walleij
wrote:
> Not that one, because the only place where ungate is called
> is immediately before the request or set_ios(). So the request
> or set_ios() will complete
How can you assume that ? Nothing prevents a context switch after
ungate, and before th
Hi Linus,
On Wed, Nov 10, 2010 at 05:25:43PM +0100, Linus Walleij wrote:
> Chris, do you want an incremental patch or shall I spin an all-new
> v10 patch?
I think incremental patches make sense from here on out, now that it's
in -next. We can decide whether it's worth performing any rebasing
onc
Ohad Ben-Cohen wrote:
> Hi Linus,
> On Wed, Nov 3, 2010 at 11:22 AM, Linus Walleij
> wrote:
> ...
>> +static void mmc_host_clk_gate_delayed(struct mmc_host *host)
>> +{
>> + unsigned long tick_ns;
>> + unsigned long freq = host->ios.clock;
>> + unsigned long flags;
>> + in
Hi,
thanks for the work in general, just...
> - if (pdev->dev.parent)
> - host = sdhci_alloc_host(pdev->dev.parent, 0);
> - else
> - host = sdhci_alloc_host(&pdev->dev, 0);
> -
> + host = sdhci_alloc_host(&pdev->dev, 0);
> if (IS_ERR(host)) {
>
Signed-off-by: Giuseppe Cavallaro
---
drivers/mmc/host/sdhci-pci.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/mmc/host/sdhci-pci.c b/drivers/mmc/host/sdhci-pci.c
index 59f3de5..8634fb1 100644
--- a/drivers/mmc/host/sdhci-pci.c
+++ b/drivers/mmc/host/sdhci-pc
Signed-off-by: Giuseppe Cavallaro
---
drivers/mmc/host/sdhci-pltfm.c |2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/drivers/mmc/host/sdhci-pltfm.c b/drivers/mmc/host/sdhci-pltfm.c
index 4e881a5..6ea32e1 100644
--- a/drivers/mmc/host/sdhci-pltfm.c
+++ b/drivers/mmc/host/
This patch reworks the wake-up support in the sdhci driver.
It is now able to wake-up the system on: Card interrupts,
Card insertion and Card removal events.
This has been tested on ST STB where can make sense to
wake-up the box as soon as a card is inserted.
Signed-off-by: Giuseppe Cavallaro
---
Hello.
The following patches, currently built against an ST Kernel
"2.6.32", show how to improve the wakeup support in the SDHCI
device driver.
Note:
I'm going to rework them, for example, against the mmc-next
after performing the review process and, obviously, if you
think that they can be actua
ST targets use the sdhci-pltfm driver but there are some
problems when re-insert the driver on our platforms.
Within sdhci-pltfm d.d. the pdata->init invokes own
platform function to claim some resource (based on devres):
see the example code below:
static int mmc_pad_resources(struct sdhci_host *
HC driver will be able to use the pm_flags to
undestand if the system can be woken-up by the driver.
So the mmc_suspend_host hasn't to reset this field
in the host structure.
Signed-off-by: Giuseppe Cavallaro
---
drivers/mmc/core/core.c |1 -
1 files changed, 0 insertions(+), 1 deletions(-)
Hello.
The following patches, currently built against an ST Kernel
"2.6.32", show how to improve the wakeup support in the SDHCI
device driver.
Note:
I'm going to rework them, for example, against the mmc-next
after performing the review process and, obviously, if you
think that they can be actua
On Wed, Nov 10, 2010 at 08:00:49PM +0900, Paul Mundt wrote:
> On Wed, Nov 10, 2010 at 07:27:44PM +0900, Simon Horman wrote:
> > Use pr_fmt to set the prefix for pr_ messages.
> > I believe this method is common in other
> > source files in the kernel tree.
> >
> pr_fmt() is ok when you have no bet
On Wed, Nov 10, 2010 at 06:32:23PM +0900, Paul Mundt wrote:
> On Tue, Nov 09, 2010 at 05:47:02PM +0900, Simon Horman wrote:
> > This moves some constants from sh_mmcif.c to sh_mmcif.h
> > so that they can be used in sh_mmcif_boot_init().
> >
> > It also alters the definition of SOFT_RST_OFF from (
Use pr_fmt to set the prefix for pr_ messages.
I believe this method is common in other
source files in the kernel tree.
Cc: Yusuke Goda
Signed-off-by: Simon Horman
---
drivers/mmc/host/sh_mmcif.c | 52 +-
1 files changed, 26 insertions(+), 26 deletions
On Tue, Nov 09, 2010 at 05:47:02PM +0900, Simon Horman wrote:
> This moves some constants from sh_mmcif.c to sh_mmcif.h
> so that they can be used in sh_mmcif_boot_init().
>
> It also alters the definition of SOFT_RST_OFF from (0 << 31) to
> ~SOFT_RST_ON (= ~(1 << 31)). The former seems bogus. Th
Hi Linus,
On Wed, Nov 3, 2010 at 11:22 AM, Linus Walleij
wrote:
...
> +static void mmc_host_clk_gate_delayed(struct mmc_host *host)
> +{
> + unsigned long tick_ns;
> + unsigned long freq = host->ios.clock;
> + unsigned long flags;
> + int users;
> +
> + if (!freq) {
> +#define SD_CLOCK_AND_BURST_SIZE_SETUP 0x10A
This offset is only for MMP2. On other pxa processors use a different offset.
sdhci-pxa.c needs changing to support other pxa soc chips.
The value only needs setting at initialization time or sdhci reset (RESET_ALL)
--
To unsubscrib
On Wed, Nov 10, 2010 at 07:27:44PM +0900, Simon Horman wrote:
> Use pr_fmt to set the prefix for pr_ messages.
> I believe this method is common in other
> source files in the kernel tree.
>
pr_fmt() is ok when you have no better options, but in this case you have
the struct device working for you
34 matches
Mail list logo