[lm-sensors] [PATCH 3/3] hwmon: (abituguru3) match partial DMI

[lm-sensors] [PATCH 3/3] hwmon: (abituguru3) match partial DMI

[Alistair John Strachan]

Abit DMI product name strings are composed of the product branding
and the north/south bridge chipset names, respectively.

Unfortunately, these components are inconsistently separated by a
variable number of spaces. On some boards, there is no space, and on
others, there are many spaces.

Since we don't care about matching bridge chipset names, update the
DMI probe routine to only match the first N characters of the DMI
product name, and update the motherboard table to include only the
product branding.

This change preemptively works around potential variances in the DMI
product string between different BIOS revisions for the same board.

Signed-off-by: Alistair John Strachan <alistair@devzero.co.uk>
Tested-by: Justin Piszcz <jpiszcz@lucidpixels.com>
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Jean Delvare <khali@linux-fr.org>
---
 drivers/hwmon/abituguru3.c |   18 ++++++++++++++----
 1 file changed, 14 insertions(+), 4 deletions(-)

diff --git a/drivers/hwmon/abituguru3.c b/drivers/hwmon/abituguru3.c
index ff6aad8..59b3d02 100644
--- a/drivers/hwmon/abituguru3.c
+++ b/drivers/hwmon/abituguru3.c
@@ -279,7 +279,7 @@ static const struct abituguru3_motherboard_info abituguru3_motherboards[] = {
 		{ "OTES1 Fan",		36, 2, 60, 1, 0 },
 		{ NULL, 0, 0, 0, 0, 0 } }
 	},
-	{ 0x0011, "AT8 32X(ATI RD580-ULI M1575)", {
+	{ 0x0011, "AT8 32X", {
 		{ "CPU Core",		 0, 0, 10, 1, 0 },
 		{ "DDR",		 1, 0, 20, 1, 0 },
 		{ "DDR VTT",		 2, 0, 10, 1, 0 },
@@ -402,7 +402,7 @@ static const struct abituguru3_motherboard_info abituguru3_motherboards[] = {
 		{ "AUX3 Fan",		36, 2, 60, 1, 0 },
 		{ NULL, 0, 0, 0, 0, 0 } }
 	},
-	{ 0x0016, "AW9D-MAX       (Intel i975-ICH7)", {
+	{ 0x0016, "AW9D-MAX", {
 		{ "CPU Core",		 0, 0, 10, 1, 0 },
 		{ "DDR2",		 1, 0, 20, 1, 0 },
 		{ "DDR2 VTT",		 2, 0, 10, 1, 0 },
@@ -509,7 +509,7 @@ static const struct abituguru3_motherboard_info abituguru3_motherboards[] = {
 		{ "AUX3 FAN",		36, 2, 60, 1, 0 },
 		{ NULL, 0, 0, 0, 0, 0 } }
 	},
-	{ 0x001A, "IP35 Pro(Intel P35-ICH9R)", {
+	{ 0x001A, "IP35 Pro", {
 		{ "CPU Core",		 0, 0, 10, 1, 0 },
 		{ "DDR2",		 1, 0, 20, 1, 0 },
 		{ "DDR2 VTT",		 2, 0, 10, 1, 0 },
@@ -1139,7 +1139,17 @@ static int __init abituguru3_dmi_detect(void)
 
 	for (i = 0; abituguru3_motherboards[i].id; i++) {
 		const char *dmi_name = abituguru3_motherboards[i].dmi_name;
-		if (dmi_name && !strcmp(dmi_name, board_name))
+
+		/*
+		 * Some of the vendor DMI strings are formatted inconsistently
+		 * e.g. "Board Name(Chipset)" vs "Board Name      (Chipset)"
+		 *
+		 * Work around this ambiguity by only matching partial DMI
+		 * board name strings, in an effort to avoid potential
+		 * variances between vendor BIOS revisions.
+		 */
+		if (dmi_name &&
+		    !strncmp(dmi_name, board_name, strlen(dmi_name)))
 			break;
 	}
 

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH 3/3] hwmon: (abituguru3) match partial DMI

[Hans de Goede]

Alistair John Strachan wrote:
> Abit DMI product name strings are composed of the product branding
> and the north/south bridge chipset names, respectively.
> 
> Unfortunately, these components are inconsistently separated by a
> variable number of spaces. On some boards, there is no space, and on
> others, there are many spaces.
> 
> Since we don't care about matching bridge chipset names, update the
> DMI probe routine to only match the first N characters of the DMI
> product name, and update the motherboard table to include only the
> product branding.
> 
> This change preemptively works around potential variances in the DMI
> product string between different BIOS revisions for the same board.
> 
> Signed-off-by: Alistair John Strachan <alistair@devzero.co.uk>
> Tested-by: Justin Piszcz <jpiszcz@lucidpixels.com>
> Cc: Hans de Goede <hdegoede@redhat.com>
> Cc: Jean Delvare <khali@linux-fr.org>

Looks good,

Acked-by: Hans de Goede <hdegoede@redhat.com>

Regards,

Hans


> ---
>  drivers/hwmon/abituguru3.c |   18 ++++++++++++++----
>  1 file changed, 14 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/hwmon/abituguru3.c b/drivers/hwmon/abituguru3.c
> index ff6aad8..59b3d02 100644
> --- a/drivers/hwmon/abituguru3.c
> +++ b/drivers/hwmon/abituguru3.c
> @@ -279,7 +279,7 @@ static const struct abituguru3_motherboard_info abituguru3_motherboards[] = {
>  		{ "OTES1 Fan",		36, 2, 60, 1, 0 },
>  		{ NULL, 0, 0, 0, 0, 0 } }
>  	},
> -	{ 0x0011, "AT8 32X(ATI RD580-ULI M1575)", {
> +	{ 0x0011, "AT8 32X", {
>  		{ "CPU Core",		 0, 0, 10, 1, 0 },
>  		{ "DDR",		 1, 0, 20, 1, 0 },
>  		{ "DDR VTT",		 2, 0, 10, 1, 0 },
> @@ -402,7 +402,7 @@ static const struct abituguru3_motherboard_info abituguru3_motherboards[] = {
>  		{ "AUX3 Fan",		36, 2, 60, 1, 0 },
>  		{ NULL, 0, 0, 0, 0, 0 } }
>  	},
> -	{ 0x0016, "AW9D-MAX       (Intel i975-ICH7)", {
> +	{ 0x0016, "AW9D-MAX", {
>  		{ "CPU Core",		 0, 0, 10, 1, 0 },
>  		{ "DDR2",		 1, 0, 20, 1, 0 },
>  		{ "DDR2 VTT",		 2, 0, 10, 1, 0 },
> @@ -509,7 +509,7 @@ static const struct abituguru3_motherboard_info abituguru3_motherboards[] = {
>  		{ "AUX3 FAN",		36, 2, 60, 1, 0 },
>  		{ NULL, 0, 0, 0, 0, 0 } }
>  	},
> -	{ 0x001A, "IP35 Pro(Intel P35-ICH9R)", {
> +	{ 0x001A, "IP35 Pro", {
>  		{ "CPU Core",		 0, 0, 10, 1, 0 },
>  		{ "DDR2",		 1, 0, 20, 1, 0 },
>  		{ "DDR2 VTT",		 2, 0, 10, 1, 0 },
> @@ -1139,7 +1139,17 @@ static int __init abituguru3_dmi_detect(void)
>  
>  	for (i = 0; abituguru3_motherboards[i].id; i++) {
>  		const char *dmi_name = abituguru3_motherboards[i].dmi_name;
> -		if (dmi_name && !strcmp(dmi_name, board_name))
> +
> +		/*
> +		 * Some of the vendor DMI strings are formatted inconsistently
> +		 * e.g. "Board Name(Chipset)" vs "Board Name      (Chipset)"
> +		 *
> +		 * Work around this ambiguity by only matching partial DMI
> +		 * board name strings, in an effort to avoid potential
> +		 * variances between vendor BIOS revisions.
> +		 */
> +		if (dmi_name &&
> +		    !strncmp(dmi_name, board_name, strlen(dmi_name)))
>  			break;
>  	}
>  

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH 3/3] hwmon: (abituguru3) match partial DMI

[Jean Delvare]

Hi Alistair, Hans,

On Tue, 21 Oct 2008 17:59:33 +0100, Alistair John Strachan wrote:
> Abit DMI product name strings are composed of the product branding
> and the north/south bridge chipset names, respectively.
> 
> Unfortunately, these components are inconsistently separated by a
> variable number of spaces. On some boards, there is no space, and on
> others, there are many spaces.
> 
> Since we don't care about matching bridge chipset names, update the
> DMI probe routine to only match the first N characters of the DMI
> product name, and update the motherboard table to include only the
> product branding.
> 
> This change preemptively works around potential variances in the DMI
> product string between different BIOS revisions for the same board.
> 
> Signed-off-by: Alistair John Strachan <alistair@devzero.co.uk>
> Tested-by: Justin Piszcz <jpiszcz@lucidpixels.com>
> Cc: Hans de Goede <hdegoede@redhat.com>
> Cc: Jean Delvare <khali@linux-fr.org>
> ---
>  drivers/hwmon/abituguru3.c |   18 ++++++++++++++----
>  1 file changed, 14 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/hwmon/abituguru3.c b/drivers/hwmon/abituguru3.c
> index ff6aad8..59b3d02 100644
> --- a/drivers/hwmon/abituguru3.c
> +++ b/drivers/hwmon/abituguru3.c
> @@ -279,7 +279,7 @@ static const struct abituguru3_motherboard_info abituguru3_motherboards[] = {
>  		{ "OTES1 Fan",		36, 2, 60, 1, 0 },
>  		{ NULL, 0, 0, 0, 0, 0 } }
>  	},
> -	{ 0x0011, "AT8 32X(ATI RD580-ULI M1575)", {
> +	{ 0x0011, "AT8 32X", {
>  		{ "CPU Core",		 0, 0, 10, 1, 0 },
>  		{ "DDR",		 1, 0, 20, 1, 0 },
>  		{ "DDR VTT",		 2, 0, 10, 1, 0 },
> @@ -402,7 +402,7 @@ static const struct abituguru3_motherboard_info abituguru3_motherboards[] = {
>  		{ "AUX3 Fan",		36, 2, 60, 1, 0 },
>  		{ NULL, 0, 0, 0, 0, 0 } }
>  	},
> -	{ 0x0016, "AW9D-MAX       (Intel i975-ICH7)", {
> +	{ 0x0016, "AW9D-MAX", {
>  		{ "CPU Core",		 0, 0, 10, 1, 0 },
>  		{ "DDR2",		 1, 0, 20, 1, 0 },
>  		{ "DDR2 VTT",		 2, 0, 10, 1, 0 },
> @@ -509,7 +509,7 @@ static const struct abituguru3_motherboard_info abituguru3_motherboards[] = {
>  		{ "AUX3 FAN",		36, 2, 60, 1, 0 },
>  		{ NULL, 0, 0, 0, 0, 0 } }
>  	},
> -	{ 0x001A, "IP35 Pro(Intel P35-ICH9R)", {
> +	{ 0x001A, "IP35 Pro", {
>  		{ "CPU Core",		 0, 0, 10, 1, 0 },
>  		{ "DDR2",		 1, 0, 20, 1, 0 },
>  		{ "DDR2 VTT",		 2, 0, 10, 1, 0 },
> @@ -1139,7 +1139,17 @@ static int __init abituguru3_dmi_detect(void)
>  
>  	for (i = 0; abituguru3_motherboards[i].id; i++) {
>  		const char *dmi_name = abituguru3_motherboards[i].dmi_name;
> -		if (dmi_name && !strcmp(dmi_name, board_name))
> +
> +		/*
> +		 * Some of the vendor DMI strings are formatted inconsistently
> +		 * e.g. "Board Name(Chipset)" vs "Board Name      (Chipset)"
> +		 *
> +		 * Work around this ambiguity by only matching partial DMI
> +		 * board name strings, in an effort to avoid potential
> +		 * variances between vendor BIOS revisions.
> +		 */
> +		if (dmi_name &&
> +		    !strncmp(dmi_name, board_name, strlen(dmi_name)))
>  			break;
>  	}
>  

I know that I'm the one who suggested this change, but... Are there no
known Abit boards those name is another board's name with an added
suffix? Looking at this list for example:
http://www.abit.com.tw/page/en/motherboard/motherboard_type.php?fMTYPE=Socket+AM2
I see "AN52V", "AN52 V2.0", "AN52" and "AN52S". An entry "AN52" in the
abituguru3_motherboards table would match all 4 boards. And there are
other examples such as "AW9D" and "AW9D-MAX" (which is even more
relevant as the latter is in the abituguru3_motherboards table already.)

So I don't think that the implementation above is safe, unless the
entries in abituguru3_dmi_detect are sorted specifically to make the
short string comparison always correct. But that's easy to screw this
up later.

An alternative would be to make a slightly more customized comparison
function: instead of passing strlen(dmi_name) as the length parameter
of strncmp(), you would compute the length by looking for the last
non-space character before the opening parenthesis in board_name. This
would require some more code, but would be more robust. Whether it's
worth it, I'll leave up to you: either do that or leave the code as it
is now.

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH 3/3] hwmon: (abituguru3) match partial DMI

[Alistair John Strachan]

Hi Jean,

On Tuesday 21 October 2008 20:53:00 Jean Delvare wrote:
[snip]
> So I don't think that the implementation above is safe, unless the
> entries in abituguru3_dmi_detect are sorted specifically to make the
> short string comparison always correct. But that's easy to screw this
> up later.
>
> An alternative would be to make a slightly more customized comparison
> function: instead of passing strlen(dmi_name) as the length parameter
> of strncmp(), you would compute the length by looking for the last
> non-space character before the opening parenthesis in board_name. This
> would require some more code, but would be more robust. Whether it's
> worth it, I'll leave up to you: either do that or leave the code as it
> is now.

Or drop the patch. Dropping patch 3/3 also works (at least on the boards it 
was tested on) and the patch only plugs a theoretical problem, and ultimately 
might pose more problems than it solves, as you outlined.

I don't think it's worth adding code which massages the length any further. 
For example, there's nothing stopping Abit from putting out a future board 
which uses parens in the model name too.

I'd currently vote to drop the patch, and, if it is discovered that the 
problem we identified is more than theoretical, we can apply this patch as-is, 
or rework it.

-- 
Cheers,
Alistair.

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH 3/3] hwmon: (abituguru3) match partial DMI

[Jean Delvare]

Hi Alistair,

On Tue, 21 Oct 2008 21:15:13 +0100, Alistair John Strachan wrote:
> Hi Jean,
> 
> On Tuesday 21 October 2008 20:53:00 Jean Delvare wrote:
> [snip]
> > So I don't think that the implementation above is safe, unless the
> > entries in abituguru3_dmi_detect are sorted specifically to make the
> > short string comparison always correct. But that's easy to screw this
> > up later.
> >
> > An alternative would be to make a slightly more customized comparison
> > function: instead of passing strlen(dmi_name) as the length parameter
> > of strncmp(), you would compute the length by looking for the last
> > non-space character before the opening parenthesis in board_name. This
> > would require some more code, but would be more robust. Whether it's
> > worth it, I'll leave up to you: either do that or leave the code as it
> > is now.
> 
> Or drop the patch. Dropping patch 3/3 also works (at least on the boards it 
> was tested on) and the patch only plugs a theoretical problem, and ultimately 
> might pose more problems than it solves, as you outlined.
> 
> I don't think it's worth adding code which massages the length any further. 
> For example, there's nothing stopping Abit from putting out a future board 
> which uses parens in the model name too.
> 
> I'd currently vote to drop the patch, and, if it is discovered that the 
> problem we identified is more than theoretical, we can apply this patch as-is, 
> or rework it.

Fine with me (this is actually what I meant by "leave the code as it is
now".)

BTW, thanks a lot for sending perfectly formatted patches as you've
been doing tonight, that's really a pleasure to work with you!

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH 3/3] hwmon: (abituguru3) match partial DMI

[Hans de Goede]

Alistair John Strachan wrote:
> Hi Jean,
> 
> On Tuesday 21 October 2008 20:53:00 Jean Delvare wrote:
> [snip]
>> So I don't think that the implementation above is safe, unless the
>> entries in abituguru3_dmi_detect are sorted specifically to make the
>> short string comparison always correct. But that's easy to screw this
>> up later.
>>
>> An alternative would be to make a slightly more customized comparison
>> function: instead of passing strlen(dmi_name) as the length parameter
>> of strncmp(), you would compute the length by looking for the last
>> non-space character before the opening parenthesis in board_name. This
>> would require some more code, but would be more robust. Whether it's
>> worth it, I'll leave up to you: either do that or leave the code as it
>> is now.
> 
> Or drop the patch. Dropping patch 3/3 also works (at least on the boards it 
> was tested on) and the patch only plugs a theoretical problem, and ultimately 
> might pose more problems than it solves, as you outlined.
> 
> I don't think it's worth adding code which massages the length any further. 
> For example, there's nothing stopping Abit from putting out a future board 
> which uses parens in the model name too.
> 
> I'd currently vote to drop the patch, and, if it is discovered that the 
> problem we identified is more than theoretical, we can apply this patch as-is, 
> or rework it.
> 

I agree, drop the patch.

Regards,

Hans

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors