[PATCH] PNPACPI: proper handling ACPI IO/Memory resources

[PATCH] PNPACPI: proper handling ACPI IO/Memory resources

[Zhang Rui]

Before commit b355cee88e3b1a193f0e9a81db810f6f83ad728b,
if acpi_dev_resource_memory()/acpi_dev_resource_io() returns false,
it means the the resource is not a memeory/IO resource.

But after commit b355cee88e3b1a193f0e9a81db810f6f83ad728b,
if the memory/IO resource is invalid (the length of the resource is zero),
acpi_dev_resource_memory()/acpi_dev_resource_io() returns false as well.

This breaks pnpacpi_allocated_resource(), because the current code would
recognize the invalid memory/io resource as unknown resource type.
Thus users will get warning messages on machines with zero length
ACPI memeory/IO resources.

This patch fixes the problem by invoking acpi_dev_resource_memory()/
acpi_dev_resource_io() for ACPI memory/IO resources only.

Signed-off-by: Zhang Rui <rui.zhang@intel.com>
---
 drivers/pnp/pnpacpi/rsparser.c |   15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/drivers/pnp/pnpacpi/rsparser.c b/drivers/pnp/pnpacpi/rsparser.c
index 167f3d0..66977eb 100644
--- a/drivers/pnp/pnpacpi/rsparser.c
+++ b/drivers/pnp/pnpacpi/rsparser.c
@@ -183,9 +183,7 @@ static acpi_status pnpacpi_allocated_resource(struct acpi_resource *res,
 	struct resource r = {0};
 	int i, flags;
 
-	if (acpi_dev_resource_memory(res, &r)
-	    || acpi_dev_resource_io(res, &r)
-	    || acpi_dev_resource_address_space(res, &r)
+	if (acpi_dev_resource_address_space(res, &r)
 	    || acpi_dev_resource_ext_address_space(res, &r)) {
 		pnp_add_resource(dev, &r);
 		return AE_OK;
@@ -217,6 +215,17 @@ static acpi_status pnpacpi_allocated_resource(struct acpi_resource *res,
 	}
 
 	switch (res->type) {
+	case ACPI_RESOURCE_TYPE_MEMORY24:
+	case ACPI_RESOURCE_TYPE_MEMORY32:
+	case ACPI_RESOURCE_TYPE_FIXED_MEMORY32:
+		if (acpi_dev_resource_memory(res, &r))
+			pnp_add_resource(dev, &r);
+		break;
+	case ACPI_RESOURCE_TYPE_IO:
+	case ACPI_RESOURCE_TYPE_FIXED_IO:
+		if (acpi_dev_resource_io(res, &r))
+			pnp_add_resource(dev, &r);
+		break;
 	case ACPI_RESOURCE_TYPE_DMA:
 		dma = &res->data.dma;
 		if (dma->channel_count > 0 && dma->channels[0] != (u8) -1)
-- 
1.7.9.5

RE: [PATCH] PNPACPI: proper handling ACPI IO/Memory resources

[Zhang, Rui]

Julian, Markus and Paul,

Can you please check if this patch fixes the problem for you or not?

Thanks,
rui

> -----Original Message-----
> From: Zhang, Rui
> Sent: Tuesday, March 11, 2014 10:40 PM
> To: linux-acpi@vger.kernel.org; linux-kernel@vger.kernel.org
> Cc: Wysocki, Rafael J; jwollrath@web.de; markus@trippelsdorf.de;
> pebolle@tiscali.nl; Zhang, Rui
> Subject: [PATCH] PNPACPI: proper handling ACPI IO/Memory resources
> Importance: High
> 
> Before commit b355cee88e3b1a193f0e9a81db810f6f83ad728b,
> if acpi_dev_resource_memory()/acpi_dev_resource_io() returns false, it
> means the the resource is not a memeory/IO resource.
> 
> But after commit b355cee88e3b1a193f0e9a81db810f6f83ad728b,
> if the memory/IO resource is invalid (the length of the resource is
> zero),
> acpi_dev_resource_memory()/acpi_dev_resource_io() returns false as well.
> 
> This breaks pnpacpi_allocated_resource(), because the current code
> would recognize the invalid memory/io resource as unknown resource type.
> Thus users will get warning messages on machines with zero length ACPI
> memeory/IO resources.
> 
> This patch fixes the problem by invoking acpi_dev_resource_memory()/
> acpi_dev_resource_io() for ACPI memory/IO resources only.
> 
> Signed-off-by: Zhang Rui <rui.zhang@intel.com>
> ---
>  drivers/pnp/pnpacpi/rsparser.c |   15 ++++++++++++---
>  1 file changed, 12 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/pnp/pnpacpi/rsparser.c
> b/drivers/pnp/pnpacpi/rsparser.c index 167f3d0..66977eb 100644
> --- a/drivers/pnp/pnpacpi/rsparser.c
> +++ b/drivers/pnp/pnpacpi/rsparser.c
> @@ -183,9 +183,7 @@ static acpi_status
> pnpacpi_allocated_resource(struct acpi_resource *res,
>  	struct resource r = {0};
>  	int i, flags;
> 
> -	if (acpi_dev_resource_memory(res, &r)
> -	    || acpi_dev_resource_io(res, &r)
> -	    || acpi_dev_resource_address_space(res, &r)
> +	if (acpi_dev_resource_address_space(res, &r)
>  	    || acpi_dev_resource_ext_address_space(res, &r)) {
>  		pnp_add_resource(dev, &r);
>  		return AE_OK;
> @@ -217,6 +215,17 @@ static acpi_status
> pnpacpi_allocated_resource(struct acpi_resource *res,
>  	}
> 
>  	switch (res->type) {
> +	case ACPI_RESOURCE_TYPE_MEMORY24:
> +	case ACPI_RESOURCE_TYPE_MEMORY32:
> +	case ACPI_RESOURCE_TYPE_FIXED_MEMORY32:
> +		if (acpi_dev_resource_memory(res, &r))
> +			pnp_add_resource(dev, &r);
> +		break;
> +	case ACPI_RESOURCE_TYPE_IO:
> +	case ACPI_RESOURCE_TYPE_FIXED_IO:
> +		if (acpi_dev_resource_io(res, &r))
> +			pnp_add_resource(dev, &r);
> +		break;
>  	case ACPI_RESOURCE_TYPE_DMA:
>  		dma = &res->data.dma;
>  		if (dma->channel_count > 0 && dma->channels[0] != (u8) -1)
> --
> 1.7.9.5

Re: [PATCH] PNPACPI: proper handling ACPI IO/Memory resources

[markus]

On 2014.03.11 at 14:50 +0000, Zhang, Rui wrote:
> Julian, Markus and Paul,
> 
> Can you please check if this patch fixes the problem for you or not?

Your patch fixes the issue for me.

Thanks.

-- 
Markus

Re: [PATCH] PNPACPI: proper handling ACPI IO/Memory resources

[Julian Wollrath]

Am Tue, 11 Mar 2014 14:50:42 +0000
schrieb "Zhang, Rui" <rui.zhang@intel.com>:
> Can you please check if this patch fixes the problem for you or not?
That patch fixes the problem, thank you. Feel free to add a 
Tested-by: Julian Wollrath <jwollrath@web.de>


Cheers,
Julian Wollrath

Re: [PATCH] PNPACPI: proper handling ACPI IO/Memory resources

[Paul Bolle]

Rui,

Zhang, Rui schreef op di 11-03-2014 om 14:50 [+0000]:
> Can you please check if this patch fixes the problem for you or not?

Yes, the two messages (one error and one warning) about "_CRS" are gone
with this patch applied on top of v3.14-rc6.

> > Before commit b355cee88e3b1a193f0e9a81db810f6f83ad728b,
> > if acpi_dev_resource_memory()/acpi_dev_resource_io() returns false, it
> > means the the resource is not a memeory/IO resource.

memeory?

> > But after commit b355cee88e3b1a193f0e9a81db810f6f83ad728b,
> > if the memory/IO resource is invalid (the length of the resource is
> > zero),
> > acpi_dev_resource_memory()/acpi_dev_resource_io() returns false as well.
> > 
> > This breaks pnpacpi_allocated_resource(), because the current code
> > would recognize the invalid memory/io resource as unknown resource type.
> > Thus users will get warning messages on machines with zero length ACPI
> > memeory/IO resources.

Ditto.
 
> > This patch fixes the problem by invoking acpi_dev_resource_memory()/
> > acpi_dev_resource_io() for ACPI memory/IO resources only.
> > 
> > Signed-off-by: Zhang Rui <rui.zhang@intel.com>

Thanks,


Paul Bolle

Re: [PATCH] PNPACPI: proper handling ACPI IO/Memory resources

[Rafael J. Wysocki]

On Tuesday, March 11, 2014 10:40:27 PM Zhang Rui wrote:
> Before commit b355cee88e3b1a193f0e9a81db810f6f83ad728b,
> if acpi_dev_resource_memory()/acpi_dev_resource_io() returns false,
> it means the the resource is not a memeory/IO resource.
> 
> But after commit b355cee88e3b1a193f0e9a81db810f6f83ad728b,
> if the memory/IO resource is invalid (the length of the resource is zero),
> acpi_dev_resource_memory()/acpi_dev_resource_io() returns false as well.
> 
> This breaks pnpacpi_allocated_resource(), because the current code would
> recognize the invalid memory/io resource as unknown resource type.
> Thus users will get warning messages on machines with zero length
> ACPI memeory/IO resources.
> 
> This patch fixes the problem by invoking acpi_dev_resource_memory()/
> acpi_dev_resource_io() for ACPI memory/IO resources only.
> 
> Signed-off-by: Zhang Rui <rui.zhang@intel.com>

Applied, thanks Rui!

> ---
>  drivers/pnp/pnpacpi/rsparser.c |   15 ++++++++++++---
>  1 file changed, 12 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/pnp/pnpacpi/rsparser.c b/drivers/pnp/pnpacpi/rsparser.c
> index 167f3d0..66977eb 100644
> --- a/drivers/pnp/pnpacpi/rsparser.c
> +++ b/drivers/pnp/pnpacpi/rsparser.c
> @@ -183,9 +183,7 @@ static acpi_status pnpacpi_allocated_resource(struct acpi_resource *res,
>  	struct resource r = {0};
>  	int i, flags;
>  
> -	if (acpi_dev_resource_memory(res, &r)
> -	    || acpi_dev_resource_io(res, &r)
> -	    || acpi_dev_resource_address_space(res, &r)
> +	if (acpi_dev_resource_address_space(res, &r)
>  	    || acpi_dev_resource_ext_address_space(res, &r)) {
>  		pnp_add_resource(dev, &r);
>  		return AE_OK;
> @@ -217,6 +215,17 @@ static acpi_status pnpacpi_allocated_resource(struct acpi_resource *res,
>  	}
>  
>  	switch (res->type) {
> +	case ACPI_RESOURCE_TYPE_MEMORY24:
> +	case ACPI_RESOURCE_TYPE_MEMORY32:
> +	case ACPI_RESOURCE_TYPE_FIXED_MEMORY32:
> +		if (acpi_dev_resource_memory(res, &r))
> +			pnp_add_resource(dev, &r);
> +		break;
> +	case ACPI_RESOURCE_TYPE_IO:
> +	case ACPI_RESOURCE_TYPE_FIXED_IO:
> +		if (acpi_dev_resource_io(res, &r))
> +			pnp_add_resource(dev, &r);
> +		break;
>  	case ACPI_RESOURCE_TYPE_DMA:
>  		dma = &res->data.dma;
>  		if (dma->channel_count > 0 && dma->channels[0] != (u8) -1)
> 

-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.