[cocci] Deleting duplicate condition checks (with SmPL)?

[cocci] Deleting duplicate condition checks (with SmPL)?

[Markus Elfring]

Hello,

I noticed another contribution.

I got into the mood to construct another script for the semantic patch language.


// See also:
// [PATCH net-next] ipv6: mcast: Remove unnecessary null check in mld_del_delrec()
// https://lore.kernel.org/lkml/20250712092811.2992283-1-yuehaibing@huawei.com/
//
// https://elixir.bootlin.com/linux/v6.16-rc5/source/net/ipv6/mcast.c#L786-L813
@deletions@
expression x;
@@
 if (x)
 {
 <+...
-}
-if (x)
-{
 ...+>
 }


Questionable test result (by the software combination “Coccinelle 1.3.0” for example):
Markus_Elfring@Sonne:…/Projekte/Coccinelle/janitor> /usr/bin/spatch --parse-cocci delete_duplicate_condition_checks2.cocci
…
minus: parse error: 
  File "delete_duplicate_condition_checks2.cocci", line 12, column 1, charpos = 307
  around = '}',
  whole content = -}


The transformation approach is working as expected if the specification “... when any”
would be applied instead of the SmPL nest construct.
Which software improvements will become more desirable for such an use case?

Regards,
Markus

Re: [cocci] Deleting duplicate condition checks (with SmPL)?

[Markus Elfring]

> Which software improvements will become more desirable for such an use case?

Will another script variant for the semantic patch language become helpful?


// See also:
// [PATCH net-next] ipv6: mcast: Remove unnecessary null check in mld_del_delrec()
// https://lore.kernel.org/netdev/CAAVpQUC-mV=SuNNhKbpy_1Mbh_sOs856+oNqDVJ4KcLjhDh2kw@mail.gmail.com/
// https://lkml.org/lkml/2025/7/12/336
//
// https://elixir.bootlin.com/linux/v6.16-rc5/source/net/ipv6/mcast.c#L786-L813
@adjustment disable braces0@
expression e;
identifier i;
@@
 void i(...)
 {
 ... when any
 if (
+    !
     e
    )
+   return;
-   {
    ... when any
-   }
-if (e)
-   {
    ... when any
-   }
 ... when any
 }


Would you be looking for further source code layout refinements then?

Regards,
Markus

Re: [cocci] Deleting duplicate condition checks (with SmPL)?

[Julia Lawall]

> On 14 Jul 2025, at 03:27, Markus Elfring <Markus.Elfring@web.de> wrote:
> 
> 
>> 
>> Which software improvements will become more desirable for such an use case?
> 
> Will another script variant for the semantic patch language become helpful?
> 
> 
> // See also:
> // [PATCH net-next] ipv6: mcast: Remove unnecessary null check in mld_del_delrec()
> // https://lore.kernel.org/netdev/CAAVpQUC-mV=SuNNhKbpy_1Mbh_sOs856+oNqDVJ4KcLjhDh2kw@mail.gmail.com/
> // https://lkml.org/lkml/2025/7/12/336
> //
> // https://elixir.bootlin.com/linux/v6.16-rc5/source/net/ipv6/mcast.c#L786-L813
> @adjustment disable braces0@
> expression e;
> identifier i;
> @@
> void i(...)
> {
> ... when any
> if (
> +    !
>     e
>    )
> +   return;
> -   {
>    ... when any

I have no idea what is the question here, but the above when any is useless. Braces always match up properly.

> -   }
> -if (e)
> -   {
>    ... when any
> -   }
> ... when any
> }
> 
> 
> Would you be looking for further source code layout refinements then?
> 
> Regards,
> Markus

Re: [cocci] Deleting duplicate condition checks (with SmPL)?

[Markus Elfring]

>> // https://elixir.bootlin.com/linux/v6.16-rc5/source/net/ipv6/mcast.c#L786-L813
>> @adjustment disable braces0@
>> expression e;
>> identifier i;
>> @@
>> void i(...)
>> {
>> ... when any
>> if (
>> +    !
>>     e
>>    )
>> +   return;
>> -   {
>>    ... when any
> 
> I have no idea what is the question here,

Would you find any details improvable in the generated diff output?

Markus_Elfring@Sonne:…/Projekte/Linux/next-analyses> time /usr/bin/spatch …/Projekte/Coccinelle/janitor/delete_duplicate_condition_checks4.cocci net/ipv6/mcast.c
…
@@ -789,34 +789,31 @@ static void mld_del_delrec(struct inet6_
…


Examples:
* Coding style preferences?
…
+       if (!pmc)
+               return;
+       if (pmc_prev)
…

* Indentation?
…
+               for_each_psf_mclock(im, psf)
+                       psf->sf_crcount = idev->mc_qrv;
+} else {
+               im->mca_crcount = idev->mc_qrv;
        }
…


> but the above when any is useless. Braces always match up properly.

Will such information become relevant for a better software documentation?

Regards,
Markus

Re: [cocci] Deleting duplicate condition checks (with SmPL)?

[Julia Lawall]

[-- Attachment #1: Type: text/plain, Size: 1533 bytes --]



On Mon, 14 Jul 2025, Markus Elfring wrote:

> >> // https://elixir.bootlin.com/linux/v6.16-rc5/source/net/ipv6/mcast.c#L786-L813
> >> @adjustment disable braces0@
> >> expression e;
> >> identifier i;
> >> @@
> >> void i(...)
> >> {
> >> ... when any
> >> if (
> >> +    !
> >>     e
> >>    )
> >> +   return;
> >> -   {
> >>    ... when any
> >
> > I have no idea what is the question here,
>
> Would you find any details improvable in the generated diff output?
>
> Markus_Elfring@Sonne:…/Projekte/Linux/next-analyses> time /usr/bin/spatch …/Projekte/Coccinelle/janitor/delete_duplicate_condition_checks4.cocci net/ipv6/mcast.c
> …
> @@ -789,34 +789,31 @@ static void mld_del_delrec(struct inet6_
> …
>
>
> Examples:
> * Coding style preferences?
> …
> +       if (!pmc)
> +               return;
> +       if (pmc_prev)
> …
>
> * Indentation?
> …
> +               for_each_psf_mclock(im, psf)
> +                       psf->sf_crcount = idev->mc_qrv;
> +} else {
> +               im->mca_crcount = idev->mc_qrv;
>         }
> …
>

I still have no idea what the question is.  Please provide a single email
with both the semantic patch and a small C program thta illustrates
whatever you are concerned about.  This is the rule for all issues that
you want to raise.  Everything should be in one email and easily
reproducible,

julia


>
> > but the above when any is useless. Braces always match up properly.
>
> Will such information become relevant for a better software documentation?
>
> Regards,
> Markus
>

Re: [cocci] Deleting duplicate condition checks (with SmPL)?

[Markus Elfring]

> I still have no idea what the question is.

Can you reproduce my observations in your software test environment?

Regards,
Markus

Re: [cocci] Deleting duplicate condition checks (with SmPL)?

[Julia Lawall]

Sent from my iPhone

> On 13 Jul 2025, at 03:21, Markus Elfring <Markus.Elfring@web.de> wrote:
> 
> Hello,
> 
> I noticed another contribution.
> 
> I got into the mood to construct another script for the semantic patch language.
> 
> 
> // See also:
> // [PATCH net-next] ipv6: mcast: Remove unnecessary null check in mld_del_delrec()
> // https://lore.kernel.org/lkml/20250712092811.2992283-1-yuehaibing@huawei.com/
> //
> // https://elixir.bootlin.com/linux/v6.16-rc5/source/net/ipv6/mcast.c#L786-L813
> @deletions@
> expression x;
> @@
> if (x)
> {
> <+...
> -}
> -if (x)
> -{
> ...+>
> }
> 
> 

It expects a real statement inside <+… …+>

> Questionable test result (by the software combination “Coccinelle 1.3.0” for example):
> Markus_Elfring@Sonne:…/Projekte/Coccinelle/janitor> /usr/bin/spatch --parse-cocci delete_duplicate_condition_checks2.cocci
> …
> minus: parse error:
>  File "delete_duplicate_condition_checks2.cocci", line 12, column 1, charpos = 307
>  around = '}',
>  whole content = -}
> 
> 
> The transformation approach is working as expected if the specification “... when any”
> would be applied instead of the SmPL nest construct.
> Which software improvements will become more desirable for such an use case?
> 
> Regards,
> Markus