[PATCH] PCI: imx6: fix occasional link failure

[PATCH] PCI: imx6: fix occasional link failure

[Tim Harvey]

According to the IMX6 reference manuals, REF_SSP_EN (Reference clock enable
for SS function) must remain deasserted until the reference clock is running
at the appropriate frequency.

Without this patch we find a high link failure rate (>5%) on certain
IMX6 boards at various temperatures.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
---
 drivers/pci/host/pci-imx6.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/pci/host/pci-imx6.c b/drivers/pci/host/pci-imx6.c
index 1be6073..9b6bab9 100644
--- a/drivers/pci/host/pci-imx6.c
+++ b/drivers/pci/host/pci-imx6.c
@@ -256,11 +256,6 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port *pp)
 	struct imx6_pcie *imx6_pcie = to_imx6_pcie(pp);
 	int ret;
 
-	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
-			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
-	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
-			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
-
 	ret = clk_prepare_enable(imx6_pcie->pcie_phy);
 	if (ret) {
 		dev_err(pp->dev, "unable to enable pcie_phy clock\n");
@@ -282,6 +277,12 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port *pp)
 	/* allow the clocks to stabilize */
 	usleep_range(200, 500);
 
+	/* power up core phy and enable ref clock */
+	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
+			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
+	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
+			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
+
 	/* Some boards don't have PCIe reset GPIO. */
 	if (gpio_is_valid(imx6_pcie->reset_gpio)) {
 		gpio_set_value(imx6_pcie->reset_gpio, 0);
-- 
1.8.3.2

RE: [PATCH] PCI: imx6: fix occasional link failure

[Hong-Xing.Zhu]

Hi Tim:

> -----Original Message-----
> From: linux-pci-owner@vger.kernel.org [mailto:linux-pci-owner@vger.kernel.org]
> On Behalf Of Tim Harvey
> Sent: Friday, August 08, 2014 2:37 PM
> To: l.stach@pengutronix.de
> Cc: Fabio Estevam; Guo Shawn-R65073; stable@vger.kernel.org; marex@denx.de;
> bhelgaas@google.com; linux-pci@vger.kernel.org
> Subject: [PATCH] PCI: imx6: fix occasional link failure
> 
> According to the IMX6 reference manuals, REF_SSP_EN (Reference clock enable
> for SS function) must remain deasserted until the reference clock is running
> at the appropriate frequency.
> 
> Without this patch we find a high link failure rate (>5%) on certain
> IMX6 boards at various temperatures.
> 
> Signed-off-by: Tim Harvey <tharvey@gateworks.com>
> ---
>  drivers/pci/host/pci-imx6.c | 11 ++++++-----
>  1 file changed, 6 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/pci/host/pci-imx6.c b/drivers/pci/host/pci-imx6.c index
> 1be6073..9b6bab9 100644
> --- a/drivers/pci/host/pci-imx6.c
> +++ b/drivers/pci/host/pci-imx6.c
> @@ -256,11 +256,6 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port
> *pp)
>  	struct imx6_pcie *imx6_pcie = to_imx6_pcie(pp);
>  	int ret;
> 
> -	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> -			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
> -	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> -			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
> -
>  	ret = clk_prepare_enable(imx6_pcie->pcie_phy);
>  	if (ret) {
>  		dev_err(pp->dev, "unable to enable pcie_phy clock\n"); @@ -282,6
> +277,12 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port *pp)
>  	/* allow the clocks to stabilize */
>  	usleep_range(200, 500);
> 
> +	/* power up core phy and enable ref clock */
> +	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> +			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
> +	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> +			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
> +

[Richard] It's better add some comments into the codes, why
Enable_pcie_clock should be placed before ssp_en is set.
For example:
    - enable_pcie_clock should be call before ssp_en is set,
      since that ssp_en control the phy_ref clk gate, turn on
      it after the source of the pcie clks are stable.

Best Regards
Richard Zhu

>  	/* Some boards don't have PCIe reset GPIO. */
>  	if (gpio_is_valid(imx6_pcie->reset_gpio)) {
>  		gpio_set_value(imx6_pcie->reset_gpio, 0);
> --
> 1.8.3.2
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pci" in the
> body of a message to majordomo@vger.kernel.org More majordomo info at
> http://vger.kernel.org/majordomo-info.html

Re: [PATCH] PCI: imx6: fix occasional link failure

[Greg KH]

On Thu, Aug 07, 2014 at 11:36:40PM -0700, Tim Harvey wrote:
> According to the IMX6 reference manuals, REF_SSP_EN (Reference clock enable
> for SS function) must remain deasserted until the reference clock is running
> at the appropriate frequency.
> 
> Without this patch we find a high link failure rate (>5%) on certain
> IMX6 boards at various temperatures.
> 
> Signed-off-by: Tim Harvey <tharvey@gateworks.com>
> ---
>  drivers/pci/host/pci-imx6.c | 11 ++++++-----
>  1 file changed, 6 insertions(+), 5 deletions(-)

<formletter>

This is not the correct way to submit patches for inclusion in the
stable kernel tree.  Please read Documentation/stable_kernel_rules.txt
for how to do this properly.

</formletter>

Re: [PATCH] PCI: imx6: fix occasional link failure

[Fabio Estevam]

On Fri, Aug 8, 2014 at 3:36 AM, Tim Harvey <tharvey@gateworks.com> wrote:
> According to the IMX6 reference manuals, REF_SSP_EN (Reference clock enable
> for SS function) must remain deasserted until the reference clock is running
> at the appropriate frequency.
>
> Without this patch we find a high link failure rate (>5%) on certain
> IMX6 boards at various temperatures.
>
> Signed-off-by: Tim Harvey <tharvey@gateworks.com>

Thanks, Tim.

Marek,

Does this fix the PCI linkup failures you reported earlier?

Regards,

Fabio Estevam

Re: [PATCH] PCI: imx6: fix occasional link failure

[Fabio Estevam]

On Fri, Aug 8, 2014 at 3:36 AM, Tim Harvey <tharvey@gateworks.com> wrote:
> According to the IMX6 reference manuals, REF_SSP_EN (Reference clock enable
> for SS function) must remain deasserted until the reference clock is running
> at the appropriate frequency.
>
> Without this patch we find a high link failure rate (>5%) on certain
> IMX6 boards at various temperatures.
>
> Signed-off-by: Tim Harvey <tharvey@gateworks.com>

Thanks for the patch, Tim:

Tested-by: Fabio Estevam <fabio.estevam@freescale.com>

Re: [PATCH] PCI: imx6: fix occasional link failure

[Marek Vasut]

On Friday, August 08, 2014 at 05:59:23 PM, Fabio Estevam wrote:
> On Fri, Aug 8, 2014 at 3:36 AM, Tim Harvey <tharvey@gateworks.com> wrote:
> > According to the IMX6 reference manuals, REF_SSP_EN (Reference clock
> > enable for SS function) must remain deasserted until the reference clock
> > is running at the appropriate frequency.
> > 
> > Without this patch we find a high link failure rate (>5%) on certain
> > IMX6 boards at various temperatures.
> > 
> > Signed-off-by: Tim Harvey <tharvey@gateworks.com>
> 
> Thanks, Tim.
> 
> Marek,
> 
> Does this fix the PCI linkup failures you reported earlier?

I currently don't have the board available, so I will moan later if this is 
still broken. Please go ahead and apply, the patch looks good to me:

Acked-by: Marek Vasut <marex@denx.de>

Best regards,
Marek Vasut

[PATCH] PCI: imx6: fix occasional link failure

[Tim Harvey]

According to the IMX6 reference manuals, REF_SSP_EN (Reference clock enable
for SS function) must remain deasserted until the reference clock is running
at the appropriate frequency.

Without this patch we find a high link failure rate (>5%) on certain
IMX6 boards at various temperatures.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Tested-by: Fabio Estevam <festevam@gmail.com>
Acked-by: Marek Vasut <marex@denx.de>
---

v2:
 - added Tested-by Fabio Estevam
 - added Acked-by Marek Vasut

---
 drivers/pci/host/pci-imx6.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/pci/host/pci-imx6.c b/drivers/pci/host/pci-imx6.c
index 1be6073..9b6bab9 100644
--- a/drivers/pci/host/pci-imx6.c
+++ b/drivers/pci/host/pci-imx6.c
@@ -256,11 +256,6 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port *pp)
 	struct imx6_pcie *imx6_pcie = to_imx6_pcie(pp);
 	int ret;
 
-	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
-			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
-	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
-			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
-
 	ret = clk_prepare_enable(imx6_pcie->pcie_phy);
 	if (ret) {
 		dev_err(pp->dev, "unable to enable pcie_phy clock\n");
@@ -282,6 +277,12 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port *pp)
 	/* allow the clocks to stabilize */
 	usleep_range(200, 500);
 
+	/* power up core phy and enable ref clock */
+	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
+			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
+	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
+			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
+
 	/* Some boards don't have PCIe reset GPIO. */
 	if (gpio_is_valid(imx6_pcie->reset_gpio)) {
 		gpio_set_value(imx6_pcie->reset_gpio, 0);
-- 
1.8.3.2

Re: [PATCH] PCI: imx6: fix occasional link failure

[Bjorn Helgaas]

On Tue, Sep 02, 2014 at 09:04:37PM -0700, Tim Harvey wrote:
> According to the IMX6 reference manuals, REF_SSP_EN (Reference clock enable
> for SS function) must remain deasserted until the reference clock is running
> at the appropriate frequency.
> 
> Without this patch we find a high link failure rate (>5%) on certain
> IMX6 boards at various temperatures.
> 
> Signed-off-by: Tim Harvey <tharvey@gateworks.com>
> Tested-by: Fabio Estevam <festevam@gmail.com>
> Acked-by: Marek Vasut <marex@denx.de>

I'm looking for the comment update requested by Richard, and an ack from
Richard and/or Lucas.  Also, please add the stable tag if you want it
(see Documentation/stable_kernel_rules.txt for details).

Bjorn

> ---
> 
> v2:
>  - added Tested-by Fabio Estevam
>  - added Acked-by Marek Vasut
> 
> ---
>  drivers/pci/host/pci-imx6.c | 11 ++++++-----
>  1 file changed, 6 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/pci/host/pci-imx6.c b/drivers/pci/host/pci-imx6.c
> index 1be6073..9b6bab9 100644
> --- a/drivers/pci/host/pci-imx6.c
> +++ b/drivers/pci/host/pci-imx6.c
> @@ -256,11 +256,6 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port *pp)
>  	struct imx6_pcie *imx6_pcie = to_imx6_pcie(pp);
>  	int ret;
>  
> -	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> -			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
> -	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> -			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
> -
>  	ret = clk_prepare_enable(imx6_pcie->pcie_phy);
>  	if (ret) {
>  		dev_err(pp->dev, "unable to enable pcie_phy clock\n");
> @@ -282,6 +277,12 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port *pp)
>  	/* allow the clocks to stabilize */
>  	usleep_range(200, 500);
>  
> +	/* power up core phy and enable ref clock */
> +	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> +			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
> +	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> +			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
> +
>  	/* Some boards don't have PCIe reset GPIO. */
>  	if (gpio_is_valid(imx6_pcie->reset_gpio)) {
>  		gpio_set_value(imx6_pcie->reset_gpio, 0);
> -- 
> 1.8.3.2
> 

Re: [PATCH] PCI: imx6: fix occasional link failure

[Lucas Stach]

Am Freitag, den 05.09.2014, 11:32 -0600 schrieb Bjorn Helgaas:
> On Tue, Sep 02, 2014 at 09:04:37PM -0700, Tim Harvey wrote:
> > According to the IMX6 reference manuals, REF_SSP_EN (Reference clock enable
> > for SS function) must remain deasserted until the reference clock is running
> > at the appropriate frequency.
> > 
> > Without this patch we find a high link failure rate (>5%) on certain
> > IMX6 boards at various temperatures.
> > 
> > Signed-off-by: Tim Harvey <tharvey@gateworks.com>
> > Tested-by: Fabio Estevam <festevam@gmail.com>
> > Acked-by: Marek Vasut <marex@denx.de>
> 
> I'm looking for the comment update requested by Richard, and an ack from
> Richard and/or Lucas.  Also, please add the stable tag if you want it
> (see Documentation/stable_kernel_rules.txt for details).
> 
We already have a comment some line above regarding clock stabilization,
so I don't think another one is strictly needed. Especially as this
initialization order is clearly defined by the reference manual.

So with or without the comment added:
Acked-by: Lucas Stach <l.stach@pengutronix.de>

> Bjorn
> 
> > ---
> > 
> > v2:
> >  - added Tested-by Fabio Estevam
> >  - added Acked-by Marek Vasut
> > 
> > ---
> >  drivers/pci/host/pci-imx6.c | 11 ++++++-----
> >  1 file changed, 6 insertions(+), 5 deletions(-)
> > 
> > diff --git a/drivers/pci/host/pci-imx6.c b/drivers/pci/host/pci-imx6.c
> > index 1be6073..9b6bab9 100644
> > --- a/drivers/pci/host/pci-imx6.c
> > +++ b/drivers/pci/host/pci-imx6.c
> > @@ -256,11 +256,6 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port *pp)
> >  	struct imx6_pcie *imx6_pcie = to_imx6_pcie(pp);
> >  	int ret;
> >  
> > -	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> > -			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
> > -	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> > -			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
> > -
> >  	ret = clk_prepare_enable(imx6_pcie->pcie_phy);
> >  	if (ret) {
> >  		dev_err(pp->dev, "unable to enable pcie_phy clock\n");
> > @@ -282,6 +277,12 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port *pp)
> >  	/* allow the clocks to stabilize */
> >  	usleep_range(200, 500);
> >  
> > +	/* power up core phy and enable ref clock */
> > +	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> > +			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
> > +	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> > +			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
> > +
> >  	/* Some boards don't have PCIe reset GPIO. */
> >  	if (gpio_is_valid(imx6_pcie->reset_gpio)) {
> >  		gpio_set_value(imx6_pcie->reset_gpio, 0);
> > -- 
> > 1.8.3.2
> > 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

-- 
Pengutronix e.K.             | Lucas Stach                 |
Industrial Linux Solutions   | http://www.pengutronix.de/  |

Re: [PATCH] PCI: imx6: fix occasional link failure

[Bjorn Helgaas]

On Thu, Aug 07, 2014 at 11:36:40PM -0700, Tim Harvey wrote:
> According to the IMX6 reference manuals, REF_SSP_EN (Reference clock enable
> for SS function) must remain deasserted until the reference clock is running
> at the appropriate frequency.
> 
> Without this patch we find a high link failure rate (>5%) on certain
> IMX6 boards at various temperatures.
> 
> Signed-off-by: Tim Harvey <tharvey@gateworks.com>

Applied to pci/host-imx6 for v3.18 with acks from Marek and Lucas, thanks!

> ---
>  drivers/pci/host/pci-imx6.c | 11 ++++++-----
>  1 file changed, 6 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/pci/host/pci-imx6.c b/drivers/pci/host/pci-imx6.c
> index 1be6073..9b6bab9 100644
> --- a/drivers/pci/host/pci-imx6.c
> +++ b/drivers/pci/host/pci-imx6.c
> @@ -256,11 +256,6 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port *pp)
>  	struct imx6_pcie *imx6_pcie = to_imx6_pcie(pp);
>  	int ret;
>  
> -	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> -			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
> -	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> -			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
> -
>  	ret = clk_prepare_enable(imx6_pcie->pcie_phy);
>  	if (ret) {
>  		dev_err(pp->dev, "unable to enable pcie_phy clock\n");
> @@ -282,6 +277,12 @@ static int imx6_pcie_deassert_core_reset(struct pcie_port *pp)
>  	/* allow the clocks to stabilize */
>  	usleep_range(200, 500);
>  
> +	/* power up core phy and enable ref clock */
> +	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> +			IMX6Q_GPR1_PCIE_TEST_PD, 0 << 18);
> +	regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR1,
> +			IMX6Q_GPR1_PCIE_REF_CLK_EN, 1 << 16);
> +
>  	/* Some boards don't have PCIe reset GPIO. */
>  	if (gpio_is_valid(imx6_pcie->reset_gpio)) {
>  		gpio_set_value(imx6_pcie->reset_gpio, 0);
> -- 
> 1.8.3.2
> 

Re: [PATCH] PCI: imx6: fix occasional link failure

[Marek Vasut]

On Wednesday, September 17, 2014 at 01:28:59 AM, Bjorn Helgaas wrote:
> On Thu, Aug 07, 2014 at 11:36:40PM -0700, Tim Harvey wrote:
> > According to the IMX6 reference manuals, REF_SSP_EN (Reference clock
> > enable for SS function) must remain deasserted until the reference clock
> > is running at the appropriate frequency.
> > 
> > Without this patch we find a high link failure rate (>5%) on certain
> > IMX6 boards at various temperatures.
> > 
> > Signed-off-by: Tim Harvey <tharvey@gateworks.com>
> 
> Applied to pci/host-imx6 for v3.18 with acks from Marek and Lucas, thanks!

Thanks!

Best regards,
Marek Vasut