[PATCH v02 0/3] arm: introduce basic DRA7 platform support

[PATCH v02 0/3] arm: introduce basic DRA7 platform support

[Andrii Tseglytskyi]

The following patch series adds basic support needed for DRA7
evm board.

v02 changes:
- DRA7 data is introduced in existing OMAP5 data file
- Useless header removed
- Comments are addressed

Andrii Anisov (1):
  xen/dt: add match for non-available nodes

Andrii Tseglytskyi (2):
  xen/arm: dra7: Add UART base address for early logging
  xen/arm: add DRA7 platform definition

 xen/arch/arm/Rules.mk          |    5 +++++
 xen/arch/arm/platforms/omap5.c |   25 ++++++++++++++++++++++++-
 xen/common/device_tree.c       |    6 +++++-
 xen/include/xen/device_tree.h  |    2 ++
 4 files changed, 36 insertions(+), 2 deletions(-)

-- 
1.7.9.5

[PATCH v02 1/3] xen/dt: add match for non-available nodes

[Andrii Tseglytskyi]

From: Andrii Anisov <andrii.anisov@globallogic.com>

Signed-off-by: Andrii Anisov <andrii.anisov@globallogic.com>
Acked-by: Julien Grall <julien.grall@linaro.org>
---
 xen/common/device_tree.c      |    6 +++++-
 xen/include/xen/device_tree.h |    2 ++
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/xen/common/device_tree.c b/xen/common/device_tree.c
index 55716a8..b14d060 100644
--- a/xen/common/device_tree.c
+++ b/xen/common/device_tree.c
@@ -692,7 +692,8 @@ bool_t dt_match_node(const struct dt_device_match *matches,
     if ( !matches )
         return 0;
 
-    while ( matches->path || matches->type || matches->compatible )
+    while ( matches->path || matches->type ||
+            matches->compatible || matches->not_available )
     {
         bool_t match = 1;
 
@@ -705,6 +706,9 @@ bool_t dt_match_node(const struct dt_device_match *matches,
         if ( matches->compatible )
             match &= dt_device_is_compatible(node, matches->compatible);
 
+        if ( matches->not_available )
+            match &= !dt_device_is_available(node);
+
         if ( match )
             return match;
 
diff --git a/xen/include/xen/device_tree.h b/xen/include/xen/device_tree.h
index 9a8c3de..a2e1f2a 100644
--- a/xen/include/xen/device_tree.h
+++ b/xen/include/xen/device_tree.h
@@ -62,11 +62,13 @@ struct dt_device_match {
     const char *path;
     const char *type;
     const char *compatible;
+    const bool_t not_available;
 };
 
 #define DT_MATCH_PATH(p)                { .path = p }
 #define DT_MATCH_TYPE(typ)              { .type = typ }
 #define DT_MATCH_COMPATIBLE(compat)     { .compatible = compat }
+#define DT_MATCH_NOT_AVAILABLE()        { .not_available = 1 }
 
 typedef u32 dt_phandle;
 
-- 
1.7.9.5

Re: [PATCH v02 1/3] xen/dt: add match for non-available nodes

[Ian Campbell]

On Wed, 2014-07-02 at 16:49 +0300, Andrii Tseglytskyi wrote:
> From: Andrii Anisov <andrii.anisov@globallogic.com>
> 
> Signed-off-by: Andrii Anisov <andrii.anisov@globallogic.com>
> Acked-by: Julien Grall <julien.grall@linaro.org>

Acked-by: Ian Campbell <ian.campbell@citrix.com>

One nit:
> +#define DT_MATCH_NOT_AVAILABLE()        { .not_available = 1 }

Since its a boot_t s/1/true/?

Ian.

Re: [PATCH v02 1/3] xen/dt: add match for non-available nodes

[Andrii Tseglytskyi]

Hi Ian

On Thu, Jul 3, 2014 at 12:53 PM, Ian Campbell <Ian.Campbell@citrix.com> wrote:
> On Wed, 2014-07-02 at 16:49 +0300, Andrii Tseglytskyi wrote:
>> From: Andrii Anisov <andrii.anisov@globallogic.com>
>>
>> Signed-off-by: Andrii Anisov <andrii.anisov@globallogic.com>
>> Acked-by: Julien Grall <julien.grall@linaro.org>
>
> Acked-by: Ian Campbell <ian.campbell@citrix.com>
>
> One nit:
>> +#define DT_MATCH_NOT_AVAILABLE()        { .not_available = 1 }
>
> Since its a boot_t s/1/true/?

I tried this first, but one more include is needed to put "true"
instead of "1". After I found this code in devtree file:

xen/common/device_tree.c

 689 bool_t dt_match_node(const struct dt_device_match *matches,
 690                      const struct dt_device_node *node)
 691 {
 692     if ( !matches )
 693         return 0;
 694
 695     while ( matches->path || matches->type ||
 696             matches->compatible || matches->not_available )
 697     {
 698         *bool_t match = 1;*


So, I decided to do the same.

Regards,
Andrii
>
> Ian.
>



-- 

Andrii Tseglytskyi | Embedded Dev
GlobalLogic
www.globallogic.com

Re: [PATCH v02 1/3] xen/dt: add match for non-available nodes

[Ian Campbell]

On Thu, 2014-07-03 at 13:32 +0300, Andrii Tseglytskyi wrote:
> Hi Ian
> 
> On Thu, Jul 3, 2014 at 12:53 PM, Ian Campbell <Ian.Campbell@citrix.com> wrote:
> > On Wed, 2014-07-02 at 16:49 +0300, Andrii Tseglytskyi wrote:
> >> From: Andrii Anisov <andrii.anisov@globallogic.com>
> >>
> >> Signed-off-by: Andrii Anisov <andrii.anisov@globallogic.com>
> >> Acked-by: Julien Grall <julien.grall@linaro.org>
> >
> > Acked-by: Ian Campbell <ian.campbell@citrix.com>
> >
> > One nit:
> >> +#define DT_MATCH_NOT_AVAILABLE()        { .not_available = 1 }
> >
> > Since its a boot_t s/1/true/?
> 
> I tried this first, but one more include is needed to put "true"
> instead of "1". After I found this code in devtree file:
> 
> xen/common/device_tree.c
> 
>  689 bool_t dt_match_node(const struct dt_device_match *matches,
>  690                      const struct dt_device_node *node)
>  691 {
>  692     if ( !matches )
>  693         return 0;
>  694
>  695     while ( matches->path || matches->type ||
>  696             matches->compatible || matches->not_available )
>  697     {
>  698         *bool_t match = 1;*
> 
> 
> So, I decided to do the same.

OK. I htink the existing code is also wrong but I won't object to you
follow suit.

Ian.

[PATCH v02 2/3] xen/arm: dra7: Add UART base address for early logging

[Andrii Tseglytskyi]

DRA7 platform uses UART1 for console logging. Patch adds
its address to make early printk working.

Signed-off-by: Andrii Tseglytskyi <andrii.tseglytskyi@globallogic.com>
Acked-by: Julien Grall <julien.grall@linaro.org>
---
 xen/arch/arm/Rules.mk |    5 +++++
 1 file changed, 5 insertions(+)

diff --git a/xen/arch/arm/Rules.mk b/xen/arch/arm/Rules.mk
index aa2e79f..7552d40 100644
--- a/xen/arch/arm/Rules.mk
+++ b/xen/arch/arm/Rules.mk
@@ -67,6 +67,11 @@ EARLY_PRINTK_INC := 8250
 EARLY_UART_BASE_ADDRESS := 0x48020000
 EARLY_UART_REG_SHIFT := 2
 endif
+ifeq ($(CONFIG_EARLY_PRINTK), dra7)
+EARLY_PRINTK_INC := 8250
+EARLY_UART_BASE_ADDRESS := 0x4806A000
+EARLY_UART_REG_SHIFT := 2
+endif
 ifeq ($(CONFIG_EARLY_PRINTK), sun6i)
 EARLY_PRINTK_INC := 8250
 EARLY_UART_BASE_ADDRESS := 0x01c28000
-- 
1.7.9.5

Re: [PATCH v02 2/3] xen/arm: dra7: Add UART base address for early logging

[Ian Campbell]

On Wed, 2014-07-02 at 16:49 +0300, Andrii Tseglytskyi wrote:
> DRA7 platform uses UART1 for console logging. Patch adds
> its address to make early printk working.
> 
> Signed-off-by: Andrii Tseglytskyi <andrii.tseglytskyi@globallogic.com>
> Acked-by: Julien Grall <julien.grall@linaro.org>

Acked-by: Ian Campbell <ian.campbell@citrix.com>

Ian.

[PATCH v02 3/3] xen/arm: add DRA7 platform definition

[Andrii Tseglytskyi]

DRA7 platform definition is added to already existing
OMAP5 data. Data definitions, needed for DRA7 are
almost common for OMAP5 family.

Signed-off-by: Andrii Tseglytskyi <andrii.tseglytskyi@globallogic.com>
---
 xen/arch/arm/platforms/omap5.c |   25 ++++++++++++++++++++++++-
 1 file changed, 24 insertions(+), 1 deletion(-)

diff --git a/xen/arch/arm/platforms/omap5.c b/xen/arch/arm/platforms/omap5.c
index 76d4d9b..837e143 100644
--- a/xen/arch/arm/platforms/omap5.c
+++ b/xen/arch/arm/platforms/omap5.c
@@ -144,12 +144,24 @@ static int __init omap5_smp_init(void)
     return 0;
 }
 
-static const char const *omap5_dt_compat[] __initconst =
+static const char * const omap5_dt_compat[] __initconst =
 {
     "ti,omap5",
+    "ti,dra7",
     NULL
 };
 
+static const struct dt_device_match omap5_blacklist_dev[] __initconst =
+{
+    /* OMAP Linux kernel handles devices with status "disabled" in a
+     * weird manner - tries to reset them. While their memory ranges
+     * are not mapped, this leads to data aborts, so skip these devices
+     * from DT for dom0.
+     */
+    DT_MATCH_NOT_AVAILABLE(),
+    { /* sentinel */ },
+};
+
 PLATFORM_START(omap5, "TI OMAP5")
     .compatible = omap5_dt_compat,
     .init_time = omap5_init_time,
@@ -161,6 +173,17 @@ PLATFORM_START(omap5, "TI OMAP5")
     .dom0_gnttab_size = 0x20000,
 PLATFORM_END
 
+PLATFORM_START(dra7, "TI DRA7")
+    .compatible = omap5_dt_compat,
+    .init_time = omap5_init_time,
+    .cpu_up = cpu_up_send_sgi,
+    .smp_init = omap5_smp_init,
+
+    .dom0_gnttab_start = 0x4b000000,
+    .dom0_gnttab_size = 0x20000,
+    .blacklist_dev = omap5_blacklist_dev,
+PLATFORM_END
+
 /*
  * Local variables:
  * mode: C
-- 
1.7.9.5

Re: [PATCH v02 3/3] xen/arm: add DRA7 platform definition

[Julien Grall]

Hi Andrii,

(Adding Chen Baozi)

On 07/02/2014 02:50 PM, Andrii Tseglytskyi wrote:
> DRA7 platform definition is added to already existing
> OMAP5 data. Data definitions, needed for DRA7 are
> almost common for OMAP5 family.
> 
> Signed-off-by: Andrii Tseglytskyi <andrii.tseglytskyi@globallogic.com>
> ---
>  xen/arch/arm/platforms/omap5.c |   25 ++++++++++++++++++++++++-
>  1 file changed, 24 insertions(+), 1 deletion(-)
> 
> diff --git a/xen/arch/arm/platforms/omap5.c b/xen/arch/arm/platforms/omap5.c
> index 76d4d9b..837e143 100644
> --- a/xen/arch/arm/platforms/omap5.c
> +++ b/xen/arch/arm/platforms/omap5.c
> @@ -144,12 +144,24 @@ static int __init omap5_smp_init(void)
>      return 0;
>  }
>  
> -static const char const *omap5_dt_compat[] __initconst =
> +static const char * const omap5_dt_compat[] __initconst =
>  {
>      "ti,omap5",
> +    "ti,dra7",
>      NULL
>  };
>  
> +static const struct dt_device_match omap5_blacklist_dev[] __initconst =
> +{
> +    /* OMAP Linux kernel handles devices with status "disabled" in a
> +     * weird manner - tries to reset them. While their memory ranges
> +     * are not mapped, this leads to data aborts, so skip these devices
> +     * from DT for dom0.
> +     */
> +    DT_MATCH_NOT_AVAILABLE(),
> +    { /* sentinel */ },
> +};
> +
>  PLATFORM_START(omap5, "TI OMAP5")
>      .compatible = omap5_dt_compat,
>      .init_time = omap5_init_time,
> @@ -161,6 +173,17 @@ PLATFORM_START(omap5, "TI OMAP5")
>      .dom0_gnttab_size = 0x20000,
>  PLATFORM_END
>  
> +PLATFORM_START(dra7, "TI DRA7")
> +    .compatible = omap5_dt_compat,
> +    .init_time = omap5_init_time,
> +    .cpu_up = cpu_up_send_sgi,
> +    .smp_init = omap5_smp_init,
> +
> +    .dom0_gnttab_start = 0x4b000000,
> +    .dom0_gnttab_size = 0x20000,
> +    .blacklist_dev = omap5_blacklist_dev,
> +PLATFORM_END
> +

Hmmm... you are sharing the same compatible list with the "TI OMAP5"
platform. You are very lucky that "TI DRA7" is always use otherwise the
devices won't be blacklisted and, therefore, DOM0 will crash.

I've looked to the wiki page created by Chen [1]. And I believe we also
need to blacklist device for generic omap5. This will avoid the
different changes in the device tree [2].

I would merge both platform definition in one.

Regards,

[1]
http://wiki.xen.org/wiki/Xen_ARM_with_Virtualization_Extensions/OMAP5432_uEVM
[2]
http://wiki.xen.org/wiki/Xen_ARM_with_Virtualization_Extensions/OMAP5432_uEVM#2.3_Build_FDT

-- 
Julien Grall

Re: [PATCH v02 3/3] xen/arm: add DRA7 platform definition

[Andrii Tseglytskyi]

Hi Julien,

Do you mean this code?

+static const struct dt_device_match omap5_blacklist_dev[] __initconst =
+{
+    /* OMAP Linux kernel handles devices with status "disabled" in a
+     * weird manner - tries to reset them. While their memory ranges
+     * are not mapped, this leads to data aborts, so skip these devices
+     * from DT for dom0.
+     */
+    DT_MATCH_NOT_AVAILABLE(),
+    { /* sentinel */ },
+};
+
+PLATFORM_START(omap5, "TI OMAP5")
+    .compatible = omap5_dt_compat,
+    .init_time = omap5_init_time,
+    .cpu_up = cpu_up_send_sgi,
+    .smp_init = omap5_smp_init,
+
+    .dom0_gnttab_start = 0x4b000000,
+    .dom0_gnttab_size = 0x20000,
+    *.blacklist_dev = omap5_blacklist_dev,*
+PLATFORM_END

And do not use standalone DRA7 ?

Regards,
Andrii


On Wed, Jul 2, 2014 at 5:05 PM, Julien Grall <julien.grall@linaro.org> wrote:
> Hi Andrii,
>
> (Adding Chen Baozi)
>
> On 07/02/2014 02:50 PM, Andrii Tseglytskyi wrote:
>> DRA7 platform definition is added to already existing
>> OMAP5 data. Data definitions, needed for DRA7 are
>> almost common for OMAP5 family.
>>
>> Signed-off-by: Andrii Tseglytskyi <andrii.tseglytskyi@globallogic.com>
>> ---
>>  xen/arch/arm/platforms/omap5.c |   25 ++++++++++++++++++++++++-
>>  1 file changed, 24 insertions(+), 1 deletion(-)
>>
>> diff --git a/xen/arch/arm/platforms/omap5.c b/xen/arch/arm/platforms/omap5.c
>> index 76d4d9b..837e143 100644
>> --- a/xen/arch/arm/platforms/omap5.c
>> +++ b/xen/arch/arm/platforms/omap5.c
>> @@ -144,12 +144,24 @@ static int __init omap5_smp_init(void)
>>      return 0;
>>  }
>>
>> -static const char const *omap5_dt_compat[] __initconst =
>> +static const char * const omap5_dt_compat[] __initconst =
>>  {
>>      "ti,omap5",
>> +    "ti,dra7",
>>      NULL
>>  };
>>
>> +static const struct dt_device_match omap5_blacklist_dev[] __initconst =
>> +{
>> +    /* OMAP Linux kernel handles devices with status "disabled" in a
>> +     * weird manner - tries to reset them. While their memory ranges
>> +     * are not mapped, this leads to data aborts, so skip these devices
>> +     * from DT for dom0.
>> +     */
>> +    DT_MATCH_NOT_AVAILABLE(),
>> +    { /* sentinel */ },
>> +};
>> +
>>  PLATFORM_START(omap5, "TI OMAP5")
>>      .compatible = omap5_dt_compat,
>>      .init_time = omap5_init_time,
>> @@ -161,6 +173,17 @@ PLATFORM_START(omap5, "TI OMAP5")
>>      .dom0_gnttab_size = 0x20000,
>>  PLATFORM_END
>>
>> +PLATFORM_START(dra7, "TI DRA7")
>> +    .compatible = omap5_dt_compat,
>> +    .init_time = omap5_init_time,
>> +    .cpu_up = cpu_up_send_sgi,
>> +    .smp_init = omap5_smp_init,
>> +
>> +    .dom0_gnttab_start = 0x4b000000,
>> +    .dom0_gnttab_size = 0x20000,
>> +    .blacklist_dev = omap5_blacklist_dev,
>> +PLATFORM_END
>> +
>
> Hmmm... you are sharing the same compatible list with the "TI OMAP5"
> platform. You are very lucky that "TI DRA7" is always use otherwise the
> devices won't be blacklisted and, therefore, DOM0 will crash.
>
> I've looked to the wiki page created by Chen [1]. And I believe we also
> need to blacklist device for generic omap5. This will avoid the
> different changes in the device tree [2].
>
> I would merge both platform definition in one.
>
> Regards,
>
> [1]
> http://wiki.xen.org/wiki/Xen_ARM_with_Virtualization_Extensions/OMAP5432_uEVM
> [2]
> http://wiki.xen.org/wiki/Xen_ARM_with_Virtualization_Extensions/OMAP5432_uEVM#2.3_Build_FDT
>
> --
> Julien Grall



-- 

Andrii Tseglytskyi | Embedded Dev
GlobalLogic
www.globallogic.com

Re: [PATCH v02 3/3] xen/arm: add DRA7 platform definition

[Julien Grall]

On 07/02/2014 04:02 PM, Andrii Tseglytskyi wrote:
> Hi Julien,

Hi Andrii,

> Do you mean this code?

Yes. Sorry I wasn't very clear on my previous email.

> +static const struct dt_device_match omap5_blacklist_dev[] __initconst =
> +{
> +    /* OMAP Linux kernel handles devices with status "disabled" in a
> +     * weird manner - tries to reset them. While their memory ranges
> +     * are not mapped, this leads to data aborts, so skip these devices
> +     * from DT for dom0.
> +     */
> +    DT_MATCH_NOT_AVAILABLE(),
> +    { /* sentinel */ },
> +};
> +
> +PLATFORM_START(omap5, "TI OMAP5")
> +    .compatible = omap5_dt_compat,
> +    .init_time = omap5_init_time,
> +    .cpu_up = cpu_up_send_sgi,
> +    .smp_init = omap5_smp_init,
> +
> +    .dom0_gnttab_start = 0x4b000000,
> +    .dom0_gnttab_size = 0x20000,
> +    *.blacklist_dev = omap5_blacklist_dev,*
> +PLATFORM_END
> 
> And do not use standalone DRA7 ?

Right. If case of someone want to keep DRA7 standalone, you will have to
create 2 distinct compatible list.

Regards,

-- 
Julien Grall

Re: [PATCH v02 3/3] xen/arm: add DRA7 platform definition

[Andrii Tseglytskyi]

Hi Julien,

On Wed, Jul 2, 2014 at 6:18 PM, Julien Grall <julien.grall@linaro.org> wrote:
> On 07/02/2014 04:02 PM, Andrii Tseglytskyi wrote:
>> Hi Julien,
>
> Hi Andrii,
>
>> Do you mean this code?
>
> Yes. Sorry I wasn't very clear on my previous email.
>
>> +static const struct dt_device_match omap5_blacklist_dev[] __initconst =
>> +{
>> +    /* OMAP Linux kernel handles devices with status "disabled" in a
>> +     * weird manner - tries to reset them. While their memory ranges
>> +     * are not mapped, this leads to data aborts, so skip these devices
>> +     * from DT for dom0.
>> +     */
>> +    DT_MATCH_NOT_AVAILABLE(),
>> +    { /* sentinel */ },
>> +};
>> +
>> +PLATFORM_START(omap5, "TI OMAP5")
>> +    .compatible = omap5_dt_compat,
>> +    .init_time = omap5_init_time,
>> +    .cpu_up = cpu_up_send_sgi,
>> +    .smp_init = omap5_smp_init,
>> +
>> +    .dom0_gnttab_start = 0x4b000000,
>> +    .dom0_gnttab_size = 0x20000,
>> +    *.blacklist_dev = omap5_blacklist_dev,*
>> +PLATFORM_END
>>
>> And do not use standalone DRA7 ?
>
> Right. If case of someone want to keep DRA7 standalone, you will have to
> create 2 distinct compatible list.
>

I would like to create two standalone platform definitions and two
compatible lists. The reason is - common OMAP5 uses specific_mapping
callback, which is not needed by DRA7. In case if we decide to merge
platforms - this should be resolved somehow.
What do you think?

Regards,
Andrii

> Regards,
>
> --
> Julien Grall



-- 

Andrii Tseglytskyi | Embedded Dev
GlobalLogic
www.globallogic.com

Re: [PATCH v02 3/3] xen/arm: add DRA7 platform definition

[Julien Grall]

On 07/02/2014 05:34 PM, Andrii Tseglytskyi wrote:
> I would like to create two standalone platform definitions and two
> compatible lists. The reason is - common OMAP5 uses specific_mapping
> callback, which is not needed by DRA7. In case if we decide to merge
> platforms - this should be resolved somehow.
> What do you think?

Oops, I didn't notice that you didn't use the specific_mapping callback.
In this case, you will have to create 2 compatible lists.

Regards,

-- 
Julien Grall

Re: [PATCH v02 3/3] xen/arm: add DRA7 platform definition

[Andrii Tseglytskyi]

Hi,

On Wed, Jul 2, 2014 at 7:51 PM, Julien Grall <julien.grall@linaro.org> wrote:
> On 07/02/2014 05:34 PM, Andrii Tseglytskyi wrote:
>> I would like to create two standalone platform definitions and two
>> compatible lists. The reason is - common OMAP5 uses specific_mapping
>> callback, which is not needed by DRA7. In case if we decide to merge
>> platforms - this should be resolved somehow.
>> What do you think?
>
> Oops, I didn't notice that you didn't use the specific_mapping callback.
> In this case, you will have to create 2 compatible lists.
>

OK.

Regards,
Andrii

> Regards,
>
> --
> Julien Grall



-- 

Andrii Tseglytskyi | Embedded Dev
GlobalLogic
www.globallogic.com