Hi Michal,
I have built a Xilinx Zynq based design which doesn't use the SDHCI Write-protect line.
However, we use the MIO_0 pin for configuring a multiplexer and consequently the
SDHCI driver thinks that the card is read-only (RO).
In the patch below I have proposed a new quirk, similar to the QUIRK_BROKEN_CARD_DETECTION
flag, which causes the sdhci driver to assume a card with RW access.
This patch was made against the latest tree from Xilinx (github.com/Xilinx/linux-xlnx), Kernel 3.9.0
and I have attached it to the LKML to propose the additional quirk which may be useful on other platforms.
Kind regards,
James Walmsley
---
drivers/mmc/host/sdhci-of-xilinxps.c | 4 ++++
drivers/mmc/host/sdhci.c | 4 ++++
include/linux/mmc/sdhci.h | 1 +
3 files changed, 9 insertions(+)
diff --git a/drivers/mmc/host/sdhci-of-xilinxps.c b/drivers/mmc/host/sdhci-of-xilinxps.c
index f79e0db..5553bc0 100644
--- a/drivers/mmc/host/sdhci-of-xilinxps.c
+++ b/drivers/mmc/host/sdhci-of-xilinxps.c
@@ -211,6 +211,10 @@ static int sdhci_zynq_probe(struct platform_device *pdev)
if (prop == NULL || (!(u32) be32_to_cpup(prop)))
host->quirks |= SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+ prop = of_get_property(np, "xlnx,has-wp", NULL);
+ if (prop == NULL || (!(u32) be32_to_cpup(prop)))
+ host->quirks2 |= SDHCI_QUIRK2_BROKEN_WRITE_PROTECT;
+
return 0;
clk_notif_unreg:
diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
index 51bbba4..c40365f 100644
--- a/drivers/mmc/host/sdhci.c
+++ b/drivers/mmc/host/sdhci.c
@@ -1609,6 +1609,10 @@ static int sdhci_do_get_ro(struct sdhci_host *host)
{
int i, ro_count;
+ if(host->quirks2 & SDHCI_QUIRK2_BROKEN_WRITE_PROTECT) {
+ return 0; // Return zero to assume RW.
+ }
+
if (!(host->quirks & SDHCI_QUIRK_UNSTABLE_RO_DETECT))
return sdhci_check_ro(host);
diff --git a/include/linux/mmc/sdhci.h b/include/linux/mmc/sdhci.h
index b838ffc..3eccf12 100644
--- a/include/linux/mmc/sdhci.h
+++ b/include/linux/mmc/sdhci.h
@@ -95,6 +95,7 @@ struct sdhci_host {
/* The system physically doesn't support 1.8v, even if the host does */
#define SDHCI_QUIRK2_NO_1_8_V (1<<2)
#define SDHCI_QUIRK2_PRESET_VALUE_BROKEN (1<<3)
+#define SDHCI_QUIRK2_BROKEN_WRITE_PROTECT (1<<4)
int irq; /* Device IRQ */
void __iomem *ioaddr; /* Mapped address */
--
1.8.1.2
Hi James,
isn't it better to implement host->ops->get_ro() function for it?
It seems me better choice than introducing new quirk option.
Chris: What do you think?
Thanks,
Michal
On 08/16/2013 02:14 PM, James Walmsley wrote:
> Hi Michal,
>
> I have built a Xilinx Zynq based design which doesn't use the SDHCI Write-protect line.
> However, we use the MIO_0 pin for configuring a multiplexer and consequently the
> SDHCI driver thinks that the card is read-only (RO).
>
> In the patch below I have proposed a new quirk, similar to the QUIRK_BROKEN_CARD_DETECTION
> flag, which causes the sdhci driver to assume a card with RW access.
>
> This patch was made against the latest tree from Xilinx (github.com/Xilinx/linux-xlnx), Kernel 3.9.0
> and I have attached it to the LKML to propose the additional quirk which may be useful on other platforms.
>
> Kind regards,
>
> James Walmsley
>
> ---
> drivers/mmc/host/sdhci-of-xilinxps.c | 4 ++++
> drivers/mmc/host/sdhci.c | 4 ++++
> include/linux/mmc/sdhci.h | 1 +
> 3 files changed, 9 insertions(+)
>
> diff --git a/drivers/mmc/host/sdhci-of-xilinxps.c b/drivers/mmc/host/sdhci-of-xilinxps.c
> index f79e0db..5553bc0 100644
> --- a/drivers/mmc/host/sdhci-of-xilinxps.c
> +++ b/drivers/mmc/host/sdhci-of-xilinxps.c
> @@ -211,6 +211,10 @@ static int sdhci_zynq_probe(struct platform_device *pdev)
> if (prop == NULL || (!(u32) be32_to_cpup(prop)))
> host->quirks |= SDHCI_QUIRK_BROKEN_CARD_DETECTION;
>
> + prop = of_get_property(np, "xlnx,has-wp", NULL);
> + if (prop == NULL || (!(u32) be32_to_cpup(prop)))
> + host->quirks2 |= SDHCI_QUIRK2_BROKEN_WRITE_PROTECT;
> +
> return 0;
>
> clk_notif_unreg:
> diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
> index 51bbba4..c40365f 100644
> --- a/drivers/mmc/host/sdhci.c
> +++ b/drivers/mmc/host/sdhci.c
> @@ -1609,6 +1609,10 @@ static int sdhci_do_get_ro(struct sdhci_host *host)
> {
> int i, ro_count;
>
> + if(host->quirks2 & SDHCI_QUIRK2_BROKEN_WRITE_PROTECT) {
> + return 0; // Return zero to assume RW.
> + }
> +
> if (!(host->quirks & SDHCI_QUIRK_UNSTABLE_RO_DETECT))
> return sdhci_check_ro(host);
>
> diff --git a/include/linux/mmc/sdhci.h b/include/linux/mmc/sdhci.h
> index b838ffc..3eccf12 100644
> --- a/include/linux/mmc/sdhci.h
> +++ b/include/linux/mmc/sdhci.h
> @@ -95,6 +95,7 @@ struct sdhci_host {
> /* The system physically doesn't support 1.8v, even if the host does */
> #define SDHCI_QUIRK2_NO_1_8_V (1<<2)
> #define SDHCI_QUIRK2_PRESET_VALUE_BROKEN (1<<3)
> +#define SDHCI_QUIRK2_BROKEN_WRITE_PROTECT (1<<4)
>
> int irq; /* Device IRQ */
> void __iomem *ioaddr; /* Mapped address */
>
--
Michal Simek, Ing. (M.Eng), OpenPGP -> KeyID: FE3D1F91
w: http://www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel - Microblaze cpu - http://www.monstr.eu/fdt/
Maintainer of Linux kernel - Xilinx Zynq ARM architecture
Microblaze U-BOOT custodian and responsible for u-boot arm zynq platform
Hi,
On Tue, Aug 20 2013, Michal Simek wrote:
> Hi James,
>
> isn't it better to implement host->ops->get_ro() function for it?
> It seems me better choice than introducing new quirk option.
Yes, I agree. Thanks,
- Chris.
--
Chris Ball <[email protected]> <http://printf.net/>