On 01/09/23 11:10 pm, Christophe Leroy wrote:
>
> Le 01/09/2023 à 19:19, [email protected] a écrit :
>> [Vous ne recevez pas souvent de courriers de [email protected]. Découvrez pourquoi ceci est important à https://aka.ms/LearnAboutSenderIdentification ]
>>
>> From: Immad Mir <[email protected]>
>>
>> -1 is not a valid error code. This patch replaces it with -EPERM.
> Can you explain how it will work ?
> In scom_debug_init() rc is built by oring the value returned by
> scom_debug_init_one().
> What will be the result when oring some valid values with -EPERM ?
> It was working well with -1 because when you or -1 with anything you get
> -1 as result. But with your change I don't think it will work.
if EPERM is not always necessarily equal to 1, we can put a check in
scom_debug_init before returning rc. If it is less than 1 (because AFAIK
or-ring with negative number results back into the same negative number)
we set rc equal to -1.
Immad.
>
> Christophe
>
>> Signed-off-by: Immad Mir <[email protected]>
>> ---
>> arch/powerpc/platforms/powernv/opal-xscom.c | 4 ++--
>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/powerpc/platforms/powernv/opal-xscom.c b/arch/powerpc/platforms/powernv/opal-xscom.c
>> index 262cd6fac..ce4b089dd 100644
>> --- a/arch/powerpc/platforms/powernv/opal-xscom.c
>> +++ b/arch/powerpc/platforms/powernv/opal-xscom.c
>> @@ -171,7 +171,7 @@ static int scom_debug_init_one(struct dentry *root, struct device_node *dn,
>> if (IS_ERR(dir)) {
>> kfree(ent->path.data);
>> kfree(ent);
>> - return -1;
>> + return -EPERM;
>> }
>>
>> debugfs_create_blob("devspec", 0400, dir, &ent->path);
>> @@ -191,7 +191,7 @@ static int scom_debug_init(void)
>>
>> root = debugfs_create_dir("scom", arch_debugfs_dir);
>> if (IS_ERR(root))
>> - return -1;
>> + return -EPERM;
>>
>> rc = 0;
>> for_each_node_with_property(dn, "scom-controller") {
>> --
>> 2.40.0
>>
Le 01/09/2023 à 20:03, Immad Mir a écrit :
>
> On 01/09/23 11:10 pm, Christophe Leroy wrote:
>>
>> Le 01/09/2023 à 19:19, [email protected] a écrit :
>>> [Vous ne recevez pas souvent de courriers de [email protected].
>>> Découvrez pourquoi ceci est important à
>>> https://aka.ms/LearnAboutSenderIdentification ]
>>>
>>> From: Immad Mir <[email protected]>
>>>
>>> -1 is not a valid error code. This patch replaces it with -EPERM.
>> Can you explain how it will work ?
>> In scom_debug_init() rc is built by oring the value returned by
>> scom_debug_init_one().
>> What will be the result when oring some valid values with -EPERM ?
>> It was working well with -1 because when you or -1 with anything you get
>> -1 as result. But with your change I don't think it will work.
>
>
> if EPERM is not always necessarily equal to 1, we can put a check in
> scom_debug_init before returning rc. If it is less than 1 (because AFAIK
> or-ring with negative number results back into the same negative number)
> we set rc equal to -1.
>
The point is that EPERM is 1 by coincidence, the intention here is not
to return -EPERM but really -1, so by changing this you just make the
code harded to understand and maintain.
Christophe