[PATCH 1/2] platform/x86: think-lmi: Fix memory leaks when parsing ThinkStation WMI strings

[PATCH 1/2] platform/x86: think-lmi: Fix memory leaks when parsing ThinkStation WMI strings

[Mark Pearson]

My previous commit introduced a memory leak where the item allocated
from tlmi_setting was not freed.
This commit also renames it to avoid confusion with the similarly name
variable in the same function.

Fixes: 8a02d70679fc ("platform/x86: think-lmi: Add possible_values for ThinkStation")
Reported-by: Mirsad Todorovac <mirsad.todorovac@alu.unizg.hr>
Link: https://lore.kernel.org/lkml/df26ff45-8933-f2b3-25f4-6ee51ccda7d8@gmx.de/T/
Signed-off-by: Mark Pearson <mpearson-lenovo@squebb.ca>
---
This patch series builds on top of the proposed patch from Armin Wolf
"platform/x86: think-lmi: Fix memory leak when showing current settings"

 drivers/platform/x86/think-lmi.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/platform/x86/think-lmi.c b/drivers/platform/x86/think-lmi.c
index 88da7bcb6ce9..ad952b49617b 100644
--- a/drivers/platform/x86/think-lmi.c
+++ b/drivers/platform/x86/think-lmi.c
@@ -1459,10 +1459,10 @@ static int tlmi_analyze(void)
 			 * name string.
 			 * Try and pull that out if it's available.
 			 */
-			char *item, *optstart, *optend;
+			char *optitem, *optstart, *optend;
 
-			if (!tlmi_setting(setting->index, &item, LENOVO_BIOS_SETTING_GUID)) {
-				optstart = strstr(item, "[Optional:");
+			if (!tlmi_setting(setting->index, &optitem, LENOVO_BIOS_SETTING_GUID)) {
+				optstart = strstr(optitem, "[Optional:");
 				if (optstart) {
 					optstart += strlen("[Optional:");
 					optend = strstr(optstart, "]");
@@ -1471,6 +1471,7 @@ static int tlmi_analyze(void)
 							kstrndup(optstart, optend - optstart,
 									GFP_KERNEL);
 				}
+				kfree(optitem);
 			}
 		}
 		/*
-- 
2.39.2

[PATCH 2/2] platform/x86: think-lmi: Clean up display of current_value on Thinkstation

[Mark Pearson]

On ThinkStations on retrieving the attribute value the BIOS appends the
possible values to the string.
Clean up the display in the current_value_show function so the options
part is not displayed.

Fixes: a40cd7ef22fb ("platform/x86: think-lmi: Add WMI interface support on Lenovo platforms")
Reported by Mario Limoncello <Mario.Limonciello@amd.com>
Link: https://github.com/fwupd/fwupd/issues/5077#issuecomment-1488730526
Signed-off-by: Mark Pearson <mpearson-lenovo@squebb.ca>
---
 drivers/platform/x86/think-lmi.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/drivers/platform/x86/think-lmi.c b/drivers/platform/x86/think-lmi.c
index ad952b49617b..29e94db03238 100644
--- a/drivers/platform/x86/think-lmi.c
+++ b/drivers/platform/x86/think-lmi.c
@@ -920,7 +920,7 @@ static ssize_t display_name_show(struct kobject *kobj, struct kobj_attribute *at
 static ssize_t current_value_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
 {
 	struct tlmi_attr_setting *setting = to_tlmi_attr_setting(kobj);
-	char *item, *value;
+	char *item, *value, *p;
 	int ret;
 
 	ret = tlmi_setting(setting->index, &item, LENOVO_BIOS_SETTING_GUID);
@@ -931,11 +931,13 @@ static ssize_t current_value_show(struct kobject *kobj, struct kobj_attribute *a
 	value = strpbrk(item, ",");
 	if (!value || value == item || !strlen(value + 1))
 		ret = -EINVAL;
-	else
-		ret = sysfs_emit(buf, "%s\n", value + 1);		
-
+	else {
+		/* On Workstations remove the Options part after the value */
+		p = strchrnul(value, ';');
+		*p = '\0';
+		ret = sysfs_emit(buf, "%s\n", value + 1);
+	}
 	kfree(item);
-
 	return ret;
 }
 
-- 
2.39.2

Re: [PATCH 1/2] platform/x86: think-lmi: Fix memory leaks when parsing ThinkStation WMI strings

[Mario Limonciello]

On 4/1/23 13:43, Mark Pearson wrote:

> My previous commit introduced a memory leak where the item allocated
> from tlmi_setting was not freed.
> This commit also renames it to avoid confusion with the similarly name
> variable in the same function.
>
> Fixes: 8a02d70679fc ("platform/x86: think-lmi: Add possible_values for ThinkStation")
> Reported-by: Mirsad Todorovac <mirsad.todorovac@alu.unizg.hr>
> Link: https://lore.kernel.org/lkml/df26ff45-8933-f2b3-25f4-6ee51ccda7d8@gmx.de/T/
> Signed-off-by: Mark Pearson <mpearson-lenovo@squebb.ca>
> ---
> This patch series builds on top of the proposed patch from Armin Wolf
> "platform/x86: think-lmi: Fix memory leak when showing current settings"

Which version of the patch did you do it from?  I tried to apply v3 of 
Armin's patch
to Linus' tree and then apply these two.

The first applied, but the second had conflicts.  Here was my base (with 
patch 1 in
this series applied).

50c77e3211d7 (HEAD -> master) platform/x86: think-lmi: Fix memory leaks 
when parsing ThinkStation WMI strings
ede629af1e2c platform/x86: think-lmi: Fix memory leak when showing 
current settings
00c7b5f4ddc5 (origin/master, origin/HEAD) Merge tag 'input-for-v6.3-rc4' 
of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input

......

Applying: platform/x86: think-lmi: Clean up display of current_value on 
Thinkstation
error: patch failed: drivers/platform/x86/think-lmi.c:931
error: drivers/platform/x86/think-lmi.c: patch does not apply

>   drivers/platform/x86/think-lmi.c | 7 ++++---
>   1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/platform/x86/think-lmi.c b/drivers/platform/x86/think-lmi.c
> index 88da7bcb6ce9..ad952b49617b 100644
> --- a/drivers/platform/x86/think-lmi.c
> +++ b/drivers/platform/x86/think-lmi.c
> @@ -1459,10 +1459,10 @@ static int tlmi_analyze(void)
>   			 * name string.
>   			 * Try and pull that out if it's available.
>   			 */
> -			char *item, *optstart, *optend;
> +			char *optitem, *optstart, *optend;
>   
> -			if (!tlmi_setting(setting->index, &item, LENOVO_BIOS_SETTING_GUID)) {
> -				optstart = strstr(item, "[Optional:");
> +			if (!tlmi_setting(setting->index, &optitem, LENOVO_BIOS_SETTING_GUID)) {
> +				optstart = strstr(optitem, "[Optional:");
>   				if (optstart) {
>   					optstart += strlen("[Optional:");
>   					optend = strstr(optstart, "]");
> @@ -1471,6 +1471,7 @@ static int tlmi_analyze(void)
>   							kstrndup(optstart, optend - optstart,
>   									GFP_KERNEL);
>   				}
> +				kfree(optitem);
>   			}
>   		}
>   		/*

Re: [PATCH 1/2] platform/x86: think-lmi: Fix memory leaks when parsing ThinkStation WMI strings

[Mark Pearson]

Hi Mario

On Sun, Apr 2, 2023, at 10:04 AM, Mario Limonciello wrote:
> On 4/1/23 13:43, Mark Pearson wrote:
>
>> My previous commit introduced a memory leak where the item allocated
>> from tlmi_setting was not freed.
>> This commit also renames it to avoid confusion with the similarly name
>> variable in the same function.
>>
>> Fixes: 8a02d70679fc ("platform/x86: think-lmi: Add possible_values for ThinkStation")
>> Reported-by: Mirsad Todorovac <mirsad.todorovac@alu.unizg.hr>
>> Link: https://lore.kernel.org/lkml/df26ff45-8933-f2b3-25f4-6ee51ccda7d8@gmx.de/T/
>> Signed-off-by: Mark Pearson <mpearson-lenovo@squebb.ca>
>> ---
>> This patch series builds on top of the proposed patch from Armin Wolf
>> "platform/x86: think-lmi: Fix memory leak when showing current settings"
>
> Which version of the patch did you do it from?  I tried to apply v3 of 
> Armin's patch
> to Linus' tree and then apply these two.
>
> The first applied, but the second had conflicts.  Here was my base (with 
> patch 1 in
> this series applied).

I was using v2 - but there shouldn't be any code changes between v3 and v2....so I hadn't re-based.

I'm working in the fixes branch from pdx86.git (with the aim of making it easier for Hans to pull in the patches). It's possible I've goofed somehow...though not sure how :(

I'll go double check - but likely won't be until tomorrow I'm afraid. I should probably just wait for the patch to be accepted and then work from there....

Mark

Re: [PATCH 1/2] platform/x86: think-lmi: Fix memory leaks when parsing ThinkStation WMI strings

[Limonciello, Mario]

On 4/2/2023 14:23, Mark Pearson wrote:
> Hi Mario
> 
> On Sun, Apr 2, 2023, at 10:04 AM, Mario Limonciello wrote:
>> On 4/1/23 13:43, Mark Pearson wrote:
>>
>>> My previous commit introduced a memory leak where the item allocated
>>> from tlmi_setting was not freed.
>>> This commit also renames it to avoid confusion with the similarly name
>>> variable in the same function.
>>>
>>> Fixes: 8a02d70679fc ("platform/x86: think-lmi: Add possible_values for ThinkStation")
>>> Reported-by: Mirsad Todorovac <mirsad.todorovac@alu.unizg.hr>
>>> Link: https://lore.kernel.org/lkml/df26ff45-8933-f2b3-25f4-6ee51ccda7d8@gmx.de/T/
>>> Signed-off-by: Mark Pearson <mpearson-lenovo@squebb.ca>
>>> ---
>>> This patch series builds on top of the proposed patch from Armin Wolf
>>> "platform/x86: think-lmi: Fix memory leak when showing current settings"
>>
>> Which version of the patch did you do it from?  I tried to apply v3 of
>> Armin's patch
>> to Linus' tree and then apply these two.
>>
>> The first applied, but the second had conflicts.  Here was my base (with
>> patch 1 in
>> this series applied).
> 
> I was using v2 - but there shouldn't be any code changes between v3 and v2....so I hadn't re-based.
> 
> I'm working in the fixes branch from pdx86.git (with the aim of making it easier for Hans to pull in the patches). It's possible I've goofed somehow...though not sure how :(
> 
> I'll go double check - but likely won't be until tomorrow I'm afraid. I should probably just wait for the patch to be accepted and then work from there....
> 
> Mark

Hmm, I just tried on platform-x86/for-next.
commit e3271a5917d1501089b1a224d702aa053e2877f4 (tag: 
platform-drivers-x86-v6.3-4, platform-x86/fixes)

I get the same problem there.  I just applied Armin's single v3 patch 
and then tried your series.

Re: [PATCH 1/2] platform/x86: think-lmi: Fix memory leaks when parsing ThinkStation WMI strings

[Mark Pearson]

On Sun, Apr 2, 2023, at 6:17 PM, Limonciello, Mario wrote:
> On 4/2/2023 14:23, Mark Pearson wrote:
>> Hi Mario
>> 
>> On Sun, Apr 2, 2023, at 10:04 AM, Mario Limonciello wrote:
>>> On 4/1/23 13:43, Mark Pearson wrote:
>>>
>>>> My previous commit introduced a memory leak where the item allocated
>>>> from tlmi_setting was not freed.
>>>> This commit also renames it to avoid confusion with the similarly name
>>>> variable in the same function.
>>>>
>>>> Fixes: 8a02d70679fc ("platform/x86: think-lmi: Add possible_values for ThinkStation")
>>>> Reported-by: Mirsad Todorovac <mirsad.todorovac@alu.unizg.hr>
>>>> Link: https://lore.kernel.org/lkml/df26ff45-8933-f2b3-25f4-6ee51ccda7d8@gmx.de/T/
>>>> Signed-off-by: Mark Pearson <mpearson-lenovo@squebb.ca>
>>>> ---
>>>> This patch series builds on top of the proposed patch from Armin Wolf
>>>> "platform/x86: think-lmi: Fix memory leak when showing current settings"
>>>
>>> Which version of the patch did you do it from?  I tried to apply v3 of
>>> Armin's patch
>>> to Linus' tree and then apply these two.
>>>
>>> The first applied, but the second had conflicts.  Here was my base (with
>>> patch 1 in
>>> this series applied).
>> 
>> I was using v2 - but there shouldn't be any code changes between v3 and v2....so I hadn't re-based.
>> 
>> I'm working in the fixes branch from pdx86.git (with the aim of making it easier for Hans to pull in the patches). It's possible I've goofed somehow...though not sure how :(
>> 
>> I'll go double check - but likely won't be until tomorrow I'm afraid. I should probably just wait for the patch to be accepted and then work from there....
>> 
>> Mark
>
> Hmm, I just tried on platform-x86/for-next.
> commit e3271a5917d1501089b1a224d702aa053e2877f4 (tag: 
> platform-drivers-x86-v6.3-4, platform-x86/fixes)
>
> I get the same problem there.  I just applied Armin's single v3 patch 
> and then tried your series.

You're absolutely right...it doesn't apply if I start again from clean. 
I have no idea why - it must have been something I did when applying the patch in the first place but I'm at a loss as to what. Urgh :(

I'm regenerating the patch and will resend as a v2. Just going to run some tests as a sanity check first. 

Mark