[PATCH 1/2 v3] mtd/uclinux: support ROM and allow passing the base address

[PATCH 1/2 v3] mtd/uclinux: support ROM and allow passing the base address

[Uwe Kleine-König]

This allows to put the filesystem at a defined address in ROM allowing
to save more precious RAM.

I think it's safe to default to ROM because the intention of using the
uclinux map is to use a romfs and so mtd-ram doesn't give you anything
that mtd-rom doesn't.

Just in case I missed something a warning is added if the driver has to
fall back to the RAM mapping.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
---
Changes since v2, id:1350027693-19528-1-git-send-email-u.kleine-koenig@pengutronix.de:

 - drop a few "ram"s from printks instead of making them "ram/rom"
 - fix a typo in the commit log and add rational for the introduced warning.

Who is responsible for taking (or not) these two patches? David? Artem?

Best regards
Uwe

---
 drivers/mtd/maps/Kconfig   |    2 +-
 drivers/mtd/maps/uclinux.c |   26 +++++++++++++++++++++-----
 2 files changed, 22 insertions(+), 6 deletions(-)

diff --git a/drivers/mtd/maps/Kconfig b/drivers/mtd/maps/Kconfig
index 2e47c2e..0dc86fc 100644
--- a/drivers/mtd/maps/Kconfig
+++ b/drivers/mtd/maps/Kconfig
@@ -443,7 +443,7 @@ config MTD_GPIO_ADDR
 
 config MTD_UCLINUX
 	bool "Generic uClinux RAM/ROM filesystem support"
-	depends on MTD_RAM=y && (!MMU || COLDFIRE)
+	depends on (MTD_RAM=y || MTD_ROM=y) && (!MMU || COLDFIRE)
 	help
 	  Map driver to support image based filesystems for uClinux.
 
diff --git a/drivers/mtd/maps/uclinux.c b/drivers/mtd/maps/uclinux.c
index 299bf88..b3a9c54 100644
--- a/drivers/mtd/maps/uclinux.c
+++ b/drivers/mtd/maps/uclinux.c
@@ -24,11 +24,12 @@
 /****************************************************************************/
 
 struct map_info uclinux_ram_map = {
-	.name = "RAM",
-	.phys = (unsigned long)__bss_stop,
 	.size = 0,
 };
 
+static unsigned long physaddr = -1;
+module_param(physaddr, ulong, S_IRUGO);
+
 static struct mtd_info *uclinux_ram_mtdinfo;
 
 /****************************************************************************/
@@ -60,11 +61,17 @@ static int __init uclinux_mtd_init(void)
 	struct map_info *mapp;
 
 	mapp = &uclinux_ram_map;
+
+	if (physaddr == -1)
+		mapp->phys = (resource_size_t)__bss_stop;
+	else
+		mapp->phys = physaddr;
+
 	if (!mapp->size)
 		mapp->size = PAGE_ALIGN(ntohl(*((unsigned long *)(mapp->phys + 8))));
 	mapp->bankwidth = 4;
 
-	printk("uclinux[mtd]: RAM probe address=0x%x size=0x%x\n",
+	printk("uclinux[mtd]: probe address=0x%x size=0x%x\n",
 	       	(int) mapp->phys, (int) mapp->size);
 
 	/*
@@ -82,7 +89,16 @@ static int __init uclinux_mtd_init(void)
 
 	simple_map_init(mapp);
 
-	mtd = do_map_probe("map_ram", mapp);
+	mapp->name = "ROM";
+	mtd = do_map_probe("map_rom", mapp);
+	if (!mtd) {
+		/* fall back to ram probing for compatibility reasons */
+		mapp->name = "RAM";
+		mtd = do_map_probe("map_ram", mapp);
+		if (mtd && IS_ENABLED(CONFIG_MTD_ROM))
+			pr_err("Failed to map rom, but ram succeeded. Please report this issue!\n");
+	}
+
 	if (!mtd) {
 		printk("uclinux[mtd]: failed to find a mapping?\n");
 		return(-ENXIO);
@@ -118,6 +134,6 @@ module_exit(uclinux_mtd_cleanup);
 
 MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Greg Ungerer <gerg@snapgear.com>");
-MODULE_DESCRIPTION("Generic RAM based MTD for uClinux");
+MODULE_DESCRIPTION("Generic MTD for uClinux");
 
 /****************************************************************************/
-- 
1.7.10.4

[PATCH 2/2 v3] mtd/uclinux: add a comment about why uclinux_ram_map must not be static

[Uwe Kleine-König]

I was (at least) the second person trying to fix a warning by sparse, so
document in the code why this is a bad idea and add an extern declaration to
make sparse happy.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Greg Ungerer <gerg@uclinux.org>
---
Changes since v2, sent with
Message-Id: 1350027693-19528-2-git-send-email-u.kleine-koenig@pengutronix.de:

- note in the commit log that an extern declaration was added for sparse

 drivers/mtd/maps/uclinux.c |    6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/mtd/maps/uclinux.c b/drivers/mtd/maps/uclinux.c
index b3a9c54..fe1c83c 100644
--- a/drivers/mtd/maps/uclinux.c
+++ b/drivers/mtd/maps/uclinux.c
@@ -23,6 +23,12 @@
 
 /****************************************************************************/
 
+/*
+ * Blackfin uses uclinux_ram_map during startup, so it must not be static.
+ * Provide a dummy declaration to make sparse happy.
+ */
+extern struct map_info uclinux_ram_map;
+
 struct map_info uclinux_ram_map = {
 	.size = 0,
 };
-- 
1.7.10.4

Re: [PATCH 1/2 v3] mtd/uclinux: support ROM and allow passing the base address

[Greg Ungerer]

On 01/11/2013 02:31 AM, Uwe Kleine-König wrote:
> This allows to put the filesystem at a defined address in ROM allowing
> to save more precious RAM.
>
> I think it's safe to default to ROM because the intention of using the
> uclinux map is to use a romfs and so mtd-ram doesn't give you anything
> that mtd-rom doesn't.
>
> Just in case I missed something a warning is added if the driver has to
> fall back to the RAM mapping.
>
> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> ---
> Changes since v2, id:1350027693-19528-1-git-send-email-u.kleine-koenig@pengutronix.de:
>
>   - drop a few "ram"s from printks instead of making them "ram/rom"
>   - fix a typo in the commit log and add rational for the introduced warning.
>
> Who is responsible for taking (or not) these two patches? David? Artem?
>
> Best regards
> Uwe
>
> ---
>   drivers/mtd/maps/Kconfig   |    2 +-
>   drivers/mtd/maps/uclinux.c |   26 +++++++++++++++++++++-----
>   2 files changed, 22 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/mtd/maps/Kconfig b/drivers/mtd/maps/Kconfig
> index 2e47c2e..0dc86fc 100644
> --- a/drivers/mtd/maps/Kconfig
> +++ b/drivers/mtd/maps/Kconfig
> @@ -443,7 +443,7 @@ config MTD_GPIO_ADDR
>
>   config MTD_UCLINUX
>   	bool "Generic uClinux RAM/ROM filesystem support"
> -	depends on MTD_RAM=y && (!MMU || COLDFIRE)
> +	depends on (MTD_RAM=y || MTD_ROM=y) && (!MMU || COLDFIRE)
>   	help
>   	  Map driver to support image based filesystems for uClinux.
>
> diff --git a/drivers/mtd/maps/uclinux.c b/drivers/mtd/maps/uclinux.c
> index 299bf88..b3a9c54 100644
> --- a/drivers/mtd/maps/uclinux.c
> +++ b/drivers/mtd/maps/uclinux.c
> @@ -24,11 +24,12 @@
>   /****************************************************************************/
>
>   struct map_info uclinux_ram_map = {
> -	.name = "RAM",
> -	.phys = (unsigned long)__bss_stop,
>   	.size = 0,
>   };
>
> +static unsigned long physaddr = -1;
> +module_param(physaddr, ulong, S_IRUGO);
> +
>   static struct mtd_info *uclinux_ram_mtdinfo;
>
>   /****************************************************************************/
> @@ -60,11 +61,17 @@ static int __init uclinux_mtd_init(void)
>   	struct map_info *mapp;
>
>   	mapp = &uclinux_ram_map;
> +
> +	if (physaddr == -1)
> +		mapp->phys = (resource_size_t)__bss_stop;
> +	else
> +		mapp->phys = physaddr;
> +
>   	if (!mapp->size)
>   		mapp->size = PAGE_ALIGN(ntohl(*((unsigned long *)(mapp->phys + 8))));
>   	mapp->bankwidth = 4;
>
> -	printk("uclinux[mtd]: RAM probe address=0x%x size=0x%x\n",
> +	printk("uclinux[mtd]: probe address=0x%x size=0x%x\n",
>   	       	(int) mapp->phys, (int) mapp->size);
>
>   	/*
> @@ -82,7 +89,16 @@ static int __init uclinux_mtd_init(void)
>
>   	simple_map_init(mapp);
>
> -	mtd = do_map_probe("map_ram", mapp);
> +	mapp->name = "ROM";
> +	mtd = do_map_probe("map_rom", mapp);
> +	if (!mtd) {
> +		/* fall back to ram probing for compatibility reasons */
> +		mapp->name = "RAM";
> +		mtd = do_map_probe("map_ram", mapp);
> +		if (mtd && IS_ENABLED(CONFIG_MTD_ROM))
> +			pr_err("Failed to map rom, but ram succeeded. Please report this issue!\n");

I still don't like this, as per:

http://lists.infradead.org/pipermail/linux-mtd/2012-October/044612.html

Regards
Greg


> +	}
> +
>   	if (!mtd) {
>   		printk("uclinux[mtd]: failed to find a mapping?\n");
>   		return(-ENXIO);
> @@ -118,6 +134,6 @@ module_exit(uclinux_mtd_cleanup);
>
>   MODULE_LICENSE("GPL");
>   MODULE_AUTHOR("Greg Ungerer <gerg@snapgear.com>");
> -MODULE_DESCRIPTION("Generic RAM based MTD for uClinux");
> +MODULE_DESCRIPTION("Generic MTD for uClinux");
>
>   /****************************************************************************/
>

Re: [PATCH 1/2 v3] mtd/uclinux: support ROM and allow passing the base address

[Uwe Kleine-König]

Hello Greg,

On Fri, Jan 11, 2013 at 11:37:29PM +1000, Greg Ungerer wrote:
> On 01/11/2013 02:31 AM, Uwe Kleine-König wrote:
> >This allows to put the filesystem at a defined address in ROM allowing
> >to save more precious RAM.
> >
> >I think it's safe to default to ROM because the intention of using the
> >uclinux map is to use a romfs and so mtd-ram doesn't give you anything
> >that mtd-rom doesn't.
> >
> >Just in case I missed something a warning is added if the driver has to
> >fall back to the RAM mapping.
> >
> >Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> >---
> >Changes since v2, id:1350027693-19528-1-git-send-email-u.kleine-koenig@pengutronix.de:
> >
> >  - drop a few "ram"s from printks instead of making them "ram/rom"
> >  - fix a typo in the commit log and add rational for the introduced warning.
> >
> >Who is responsible for taking (or not) these two patches? David? Artem?
> >
> >Best regards
> >Uwe
> >
> >---
> >  drivers/mtd/maps/Kconfig   |    2 +-
> >  drivers/mtd/maps/uclinux.c |   26 +++++++++++++++++++++-----
> >  2 files changed, 22 insertions(+), 6 deletions(-)
> >
> >diff --git a/drivers/mtd/maps/Kconfig b/drivers/mtd/maps/Kconfig
> >index 2e47c2e..0dc86fc 100644
> >--- a/drivers/mtd/maps/Kconfig
> >+++ b/drivers/mtd/maps/Kconfig
> >@@ -443,7 +443,7 @@ config MTD_GPIO_ADDR
> >
> >  config MTD_UCLINUX
> >  	bool "Generic uClinux RAM/ROM filesystem support"
> >-	depends on MTD_RAM=y && (!MMU || COLDFIRE)
> >+	depends on (MTD_RAM=y || MTD_ROM=y) && (!MMU || COLDFIRE)
> >  	help
> >  	  Map driver to support image based filesystems for uClinux.
> >
> >diff --git a/drivers/mtd/maps/uclinux.c b/drivers/mtd/maps/uclinux.c
> >index 299bf88..b3a9c54 100644
> >--- a/drivers/mtd/maps/uclinux.c
> >+++ b/drivers/mtd/maps/uclinux.c
> >@@ -24,11 +24,12 @@
> >  /****************************************************************************/
> >
> >  struct map_info uclinux_ram_map = {
> >-	.name = "RAM",
> >-	.phys = (unsigned long)__bss_stop,
> >  	.size = 0,
> >  };
> >
> >+static unsigned long physaddr = -1;
> >+module_param(physaddr, ulong, S_IRUGO);
> >+
> >  static struct mtd_info *uclinux_ram_mtdinfo;
> >
> >  /****************************************************************************/
> >@@ -60,11 +61,17 @@ static int __init uclinux_mtd_init(void)
> >  	struct map_info *mapp;
> >
> >  	mapp = &uclinux_ram_map;
> >+
> >+	if (physaddr == -1)
> >+		mapp->phys = (resource_size_t)__bss_stop;
> >+	else
> >+		mapp->phys = physaddr;
> >+
> >  	if (!mapp->size)
> >  		mapp->size = PAGE_ALIGN(ntohl(*((unsigned long *)(mapp->phys + 8))));
> >  	mapp->bankwidth = 4;
> >
> >-	printk("uclinux[mtd]: RAM probe address=0x%x size=0x%x\n",
> >+	printk("uclinux[mtd]: probe address=0x%x size=0x%x\n",
> >  	       	(int) mapp->phys, (int) mapp->size);
> >
> >  	/*
> >@@ -82,7 +89,16 @@ static int __init uclinux_mtd_init(void)
> >
> >  	simple_map_init(mapp);
> >
> >-	mtd = do_map_probe("map_ram", mapp);
> >+	mapp->name = "ROM";
> >+	mtd = do_map_probe("map_rom", mapp);
> >+	if (!mtd) {
> >+		/* fall back to ram probing for compatibility reasons */
> >+		mapp->name = "RAM";
> >+		mtd = do_map_probe("map_ram", mapp);
> >+		if (mtd && IS_ENABLED(CONFIG_MTD_ROM))
> >+			pr_err("Failed to map rom, but ram succeeded. Please report this issue!\n");
> 
> I still don't like this, as per:
> 
> http://lists.infradead.org/pipermail/linux-mtd/2012-October/044612.html
I remember your concerns, but as you didn't write back on my last try to
explain why I consider the warning useful, I resent.

You wrote:

 The message is only printed if both ROM and RAM mappings are enabled.
 Many of the configs I use only have RAM mappings enabled.

Yeah, so the warning will not trigger for you. And this is correct,
because the only thing I want to catch is map_rom not being able to
provide everything needed when map_ram does.

I don't know what you would prefer. Just drop the warning? Force the
usage of map_rom by depending on (or selecting) MTD_ROM?

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

Re: [PATCH 1/2 v3] mtd/uclinux: support ROM and allow passing the base address

[Greg Ungerer]

On 01/12/2013 12:02 AM, Uwe Kleine-König wrote:
> Hello Greg,
>
> On Fri, Jan 11, 2013 at 11:37:29PM +1000, Greg Ungerer wrote:
>> On 01/11/2013 02:31 AM, Uwe Kleine-König wrote:
>>> This allows to put the filesystem at a defined address in ROM allowing
>>> to save more precious RAM.
>>>
>>> I think it's safe to default to ROM because the intention of using the
>>> uclinux map is to use a romfs and so mtd-ram doesn't give you anything
>>> that mtd-rom doesn't.
>>>
>>> Just in case I missed something a warning is added if the driver has to
>>> fall back to the RAM mapping.
>>>
>>> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
>>> ---
>>> Changes since v2, id:1350027693-19528-1-git-send-email-u.kleine-koenig@pengutronix.de:
>>>
>>>   - drop a few "ram"s from printks instead of making them "ram/rom"
>>>   - fix a typo in the commit log and add rational for the introduced warning.
>>>
>>> Who is responsible for taking (or not) these two patches? David? Artem?
>>>
>>> Best regards
>>> Uwe
>>>
>>> ---
>>>   drivers/mtd/maps/Kconfig   |    2 +-
>>>   drivers/mtd/maps/uclinux.c |   26 +++++++++++++++++++++-----
>>>   2 files changed, 22 insertions(+), 6 deletions(-)
>>>
>>> diff --git a/drivers/mtd/maps/Kconfig b/drivers/mtd/maps/Kconfig
>>> index 2e47c2e..0dc86fc 100644
>>> --- a/drivers/mtd/maps/Kconfig
>>> +++ b/drivers/mtd/maps/Kconfig
>>> @@ -443,7 +443,7 @@ config MTD_GPIO_ADDR
>>>
>>>   config MTD_UCLINUX
>>>   	bool "Generic uClinux RAM/ROM filesystem support"
>>> -	depends on MTD_RAM=y && (!MMU || COLDFIRE)
>>> +	depends on (MTD_RAM=y || MTD_ROM=y) && (!MMU || COLDFIRE)
>>>   	help
>>>   	  Map driver to support image based filesystems for uClinux.
>>>
>>> diff --git a/drivers/mtd/maps/uclinux.c b/drivers/mtd/maps/uclinux.c
>>> index 299bf88..b3a9c54 100644
>>> --- a/drivers/mtd/maps/uclinux.c
>>> +++ b/drivers/mtd/maps/uclinux.c
>>> @@ -24,11 +24,12 @@
>>>   /****************************************************************************/
>>>
>>>   struct map_info uclinux_ram_map = {
>>> -	.name = "RAM",
>>> -	.phys = (unsigned long)__bss_stop,
>>>   	.size = 0,
>>>   };
>>>
>>> +static unsigned long physaddr = -1;
>>> +module_param(physaddr, ulong, S_IRUGO);
>>> +
>>>   static struct mtd_info *uclinux_ram_mtdinfo;
>>>
>>>   /****************************************************************************/
>>> @@ -60,11 +61,17 @@ static int __init uclinux_mtd_init(void)
>>>   	struct map_info *mapp;
>>>
>>>   	mapp = &uclinux_ram_map;
>>> +
>>> +	if (physaddr == -1)
>>> +		mapp->phys = (resource_size_t)__bss_stop;
>>> +	else
>>> +		mapp->phys = physaddr;
>>> +
>>>   	if (!mapp->size)
>>>   		mapp->size = PAGE_ALIGN(ntohl(*((unsigned long *)(mapp->phys + 8))));
>>>   	mapp->bankwidth = 4;
>>>
>>> -	printk("uclinux[mtd]: RAM probe address=0x%x size=0x%x\n",
>>> +	printk("uclinux[mtd]: probe address=0x%x size=0x%x\n",
>>>   	       	(int) mapp->phys, (int) mapp->size);
>>>
>>>   	/*
>>> @@ -82,7 +89,16 @@ static int __init uclinux_mtd_init(void)
>>>
>>>   	simple_map_init(mapp);
>>>
>>> -	mtd = do_map_probe("map_ram", mapp);
>>> +	mapp->name = "ROM";
>>> +	mtd = do_map_probe("map_rom", mapp);
>>> +	if (!mtd) {
>>> +		/* fall back to ram probing for compatibility reasons */
>>> +		mapp->name = "RAM";
>>> +		mtd = do_map_probe("map_ram", mapp);
>>> +		if (mtd && IS_ENABLED(CONFIG_MTD_ROM))
>>> +			pr_err("Failed to map rom, but ram succeeded. Please report this issue!\n");
>>
>> I still don't like this, as per:
>>
>> http://lists.infradead.org/pipermail/linux-mtd/2012-October/044612.html
> I remember your concerns, but as you didn't write back on my last try to
> explain why I consider the warning useful, I resent.
 >
> You wrote:
>
>   The message is only printed if both ROM and RAM mappings are enabled.
>   Many of the configs I use only have RAM mappings enabled.
>
> Yeah, so the warning will not trigger for you. And this is correct,
> because the only thing I want to catch is map_rom not being able to
> provide everything needed when map_ram does.

But why?
I still don't see how this is a useful runtime message.


> I don't know what you would prefer. Just drop the warning? Force the
> usage of map_rom by depending on (or selecting) MTD_ROM?

I would suggest dropping the message. And yes making the use map_rom
depend on CONFIG_MAP_ROM being enabled would seem to make sense to me.

Regards
Greg

[PATCH v4 1/2] mtd/uclinux: support ROM and allow passing the base address

[Uwe Kleine-König]

This allows to put the filesystem at a defined address in ROM allowing
to save more precious RAM.

I think it's safe to default to ROM because the intention of using the
uclinux map is to use a romfs and so mtd-ram doesn't give you anything
that mtd-rom doesn't.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
---
Changes since v3,
id:1357835498-23904-1-git-send-email-u.kleine-koenig@pengutronix.de:

 - drop runtime warning
 - only probe one of map_rom or map_ram as both should succeed in the
   same cases

Changes since v2,
id:1350027693-19528-1-git-send-email-u.kleine-koenig@pengutronix.de:

 - drop a few "ram"s from printks instead of making them "ram/rom".
 - fix a typo in the commit log and add rational for the introduced
   warning.

Changes since v1,
id:1349709952-4332-1-git-send-email-u.kleine-koenig@pengutronix.de:

 - don't make uclinux_ram_map static as pointed out by Mike and Greg.
---
 drivers/mtd/maps/Kconfig   |    2 +-
 drivers/mtd/maps/uclinux.c |   24 +++++++++++++++++++-----
 2 files changed, 20 insertions(+), 6 deletions(-)

diff --git a/drivers/mtd/maps/Kconfig b/drivers/mtd/maps/Kconfig
index 62ba82c..3ed17c4 100644
--- a/drivers/mtd/maps/Kconfig
+++ b/drivers/mtd/maps/Kconfig
@@ -429,7 +429,7 @@ config MTD_GPIO_ADDR
 
 config MTD_UCLINUX
 	bool "Generic uClinux RAM/ROM filesystem support"
-	depends on MTD_RAM=y && (!MMU || COLDFIRE)
+	depends on (MTD_RAM=y || MTD_ROM=y) && (!MMU || COLDFIRE)
 	help
 	  Map driver to support image based filesystems for uClinux.
 
diff --git a/drivers/mtd/maps/uclinux.c b/drivers/mtd/maps/uclinux.c
index 299bf88..f56d0aa 100644
--- a/drivers/mtd/maps/uclinux.c
+++ b/drivers/mtd/maps/uclinux.c
@@ -23,12 +23,20 @@
 
 /****************************************************************************/
 
+#ifdef CONFIG_MTD_ROM
+#define MAP_NAME "rom"
+#else
+#define MAP_NAME "ram"
+#endif
+
 struct map_info uclinux_ram_map = {
-	.name = "RAM",
-	.phys = (unsigned long)__bss_stop,
+	.name = MAP_NAME,
 	.size = 0,
 };
 
+static unsigned long physaddr = -1;
+module_param(physaddr, ulong, S_IRUGO);
+
 static struct mtd_info *uclinux_ram_mtdinfo;
 
 /****************************************************************************/
@@ -60,11 +68,17 @@ static int __init uclinux_mtd_init(void)
 	struct map_info *mapp;
 
 	mapp = &uclinux_ram_map;
+
+	if (physaddr == -1)
+		mapp->phys = (resource_size_t)__bss_stop;
+	else
+		mapp->phys = physaddr;
+
 	if (!mapp->size)
 		mapp->size = PAGE_ALIGN(ntohl(*((unsigned long *)(mapp->phys + 8))));
 	mapp->bankwidth = 4;
 
-	printk("uclinux[mtd]: RAM probe address=0x%x size=0x%x\n",
+	printk("uclinux[mtd]: probe address=0x%x size=0x%x\n",
 	       	(int) mapp->phys, (int) mapp->size);
 
 	/*
@@ -82,7 +96,7 @@ static int __init uclinux_mtd_init(void)
 
 	simple_map_init(mapp);
 
-	mtd = do_map_probe("map_ram", mapp);
+	mtd = do_map_probe("map_" MAP_NAME, mapp);
 	if (!mtd) {
 		printk("uclinux[mtd]: failed to find a mapping?\n");
 		return(-ENXIO);
@@ -118,6 +132,6 @@ module_exit(uclinux_mtd_cleanup);
 
 MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Greg Ungerer <gerg@snapgear.com>");
-MODULE_DESCRIPTION("Generic RAM based MTD for uClinux");
+MODULE_DESCRIPTION("Generic MTD for uClinux");
 
 /****************************************************************************/
-- 
1.7.10.4

[PATCH v4 2/2] mtd/uclinux: add a comment about why uclinux_ram_map must not be static

[Uwe Kleine-König]

I was (at least) the second person trying to fix a warning by sparse, so
document in the code why this is a bad idea and add an extern declaration to
make sparse happy.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Greg Ungerer <gerg@uclinux.org>
---
No changes between v3 and v4.

Changes since v2,
id:1350027693-19528-2-git-send-email-u.kleine-koenig@pengutronix.de:

 - note in the commit log that an extern declaration was added for sparse

Changes since v1:

 - new patch
---
 drivers/mtd/maps/uclinux.c |    6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/mtd/maps/uclinux.c b/drivers/mtd/maps/uclinux.c
index f56d0aa..c1af83d 100644
--- a/drivers/mtd/maps/uclinux.c
+++ b/drivers/mtd/maps/uclinux.c
@@ -29,6 +29,12 @@
 #define MAP_NAME "ram"
 #endif
 
+/*
+ * Blackfin uses uclinux_ram_map during startup, so it must not be static.
+ * Provide a dummy declaration to make sparse happy.
+ */
+extern struct map_info uclinux_ram_map;
+
 struct map_info uclinux_ram_map = {
 	.name = MAP_NAME,
 	.size = 0,
-- 
1.7.10.4

Re: [PATCH v4 1/2] mtd/uclinux: support ROM and allow passing the base address

[Greg Ungerer]

On 17/01/13 00:36, Uwe Kleine-König wrote:
> This allows to put the filesystem at a defined address in ROM allowing
> to save more precious RAM.
>
> I think it's safe to default to ROM because the intention of using the
> uclinux map is to use a romfs and so mtd-ram doesn't give you anything
> that mtd-rom doesn't.
>
> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>

Acked-by: Greg Ungerer <gerg@uclinux.org>


> ---
> Changes since v3,
> id:1357835498-23904-1-git-send-email-u.kleine-koenig@pengutronix.de:
>
>   - drop runtime warning
>   - only probe one of map_rom or map_ram as both should succeed in the
>     same cases
>
> Changes since v2,
> id:1350027693-19528-1-git-send-email-u.kleine-koenig@pengutronix.de:
>
>   - drop a few "ram"s from printks instead of making them "ram/rom".
>   - fix a typo in the commit log and add rational for the introduced
>     warning.
>
> Changes since v1,
> id:1349709952-4332-1-git-send-email-u.kleine-koenig@pengutronix.de:
>
>   - don't make uclinux_ram_map static as pointed out by Mike and Greg.
> ---
>   drivers/mtd/maps/Kconfig   |    2 +-
>   drivers/mtd/maps/uclinux.c |   24 +++++++++++++++++++-----
>   2 files changed, 20 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/mtd/maps/Kconfig b/drivers/mtd/maps/Kconfig
> index 62ba82c..3ed17c4 100644
> --- a/drivers/mtd/maps/Kconfig
> +++ b/drivers/mtd/maps/Kconfig
> @@ -429,7 +429,7 @@ config MTD_GPIO_ADDR
>
>   config MTD_UCLINUX
>   	bool "Generic uClinux RAM/ROM filesystem support"
> -	depends on MTD_RAM=y && (!MMU || COLDFIRE)
> +	depends on (MTD_RAM=y || MTD_ROM=y) && (!MMU || COLDFIRE)
>   	help
>   	  Map driver to support image based filesystems for uClinux.
>
> diff --git a/drivers/mtd/maps/uclinux.c b/drivers/mtd/maps/uclinux.c
> index 299bf88..f56d0aa 100644
> --- a/drivers/mtd/maps/uclinux.c
> +++ b/drivers/mtd/maps/uclinux.c
> @@ -23,12 +23,20 @@
>
>   /****************************************************************************/
>
> +#ifdef CONFIG_MTD_ROM
> +#define MAP_NAME "rom"
> +#else
> +#define MAP_NAME "ram"
> +#endif
> +
>   struct map_info uclinux_ram_map = {
> -	.name = "RAM",
> -	.phys = (unsigned long)__bss_stop,
> +	.name = MAP_NAME,
>   	.size = 0,
>   };
>
> +static unsigned long physaddr = -1;
> +module_param(physaddr, ulong, S_IRUGO);
> +
>   static struct mtd_info *uclinux_ram_mtdinfo;
>
>   /****************************************************************************/
> @@ -60,11 +68,17 @@ static int __init uclinux_mtd_init(void)
>   	struct map_info *mapp;
>
>   	mapp = &uclinux_ram_map;
> +
> +	if (physaddr == -1)
> +		mapp->phys = (resource_size_t)__bss_stop;
> +	else
> +		mapp->phys = physaddr;
> +
>   	if (!mapp->size)
>   		mapp->size = PAGE_ALIGN(ntohl(*((unsigned long *)(mapp->phys + 8))));
>   	mapp->bankwidth = 4;
>
> -	printk("uclinux[mtd]: RAM probe address=0x%x size=0x%x\n",
> +	printk("uclinux[mtd]: probe address=0x%x size=0x%x\n",
>   	       	(int) mapp->phys, (int) mapp->size);
>
>   	/*
> @@ -82,7 +96,7 @@ static int __init uclinux_mtd_init(void)
>
>   	simple_map_init(mapp);
>
> -	mtd = do_map_probe("map_ram", mapp);
> +	mtd = do_map_probe("map_" MAP_NAME, mapp);
>   	if (!mtd) {
>   		printk("uclinux[mtd]: failed to find a mapping?\n");
>   		return(-ENXIO);
> @@ -118,6 +132,6 @@ module_exit(uclinux_mtd_cleanup);
>
>   MODULE_LICENSE("GPL");
>   MODULE_AUTHOR("Greg Ungerer <gerg@snapgear.com>");
> -MODULE_DESCRIPTION("Generic RAM based MTD for uClinux");
> +MODULE_DESCRIPTION("Generic MTD for uClinux");
>
>   /****************************************************************************/
>

Re: [PATCH v4 1/2] mtd/uclinux: support ROM and allow passing the base address

[Artem Bityutskiy]

[-- Attachment #1: Type: text/plain, Size: 420 bytes --]

On Wed, 2013-01-16 at 15:36 +0100, Uwe Kleine-König wrote:
> This allows to put the filesystem at a defined address in ROM allowing
> to save more precious RAM.
> 
> I think it's safe to default to ROM because the intention of using the
> uclinux map is to use a romfs and so mtd-ram doesn't give you anything
> that mtd-rom doesn't.

Pushed both to l2-mtd.git, thanks!

-- 
Best Regards,
Artem Bityutskiy

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]