Skip to contentSkip to conversation

coccinelle: misc: remove "complex return code" warnings

PATCH In review Johan Hovold 3 participants 11 messages Sep 30, 2015 – Oct 28, 2015 Coccinelle/Semantic Patches (SmPL)

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 coccinelle: misc: remove "complex return code" warnings

This effectively reverts 932058a5d5f9 ("coccinelle: misc: semantic patch to delete overly complex return code processing").

There can be both symmetry and readability reasons for not wanting to do the final function call as part of the return statement and to maintain a clear separation of success and error paths.

Since this is in no way mandated by the coding standard, let's just remove this semantic patch to avoid having "clean up" patches being posted over and over in response to these Coccinelle warnings.

---
 scripts/coccinelle/misc/simple_return.cocci | 180 ----------------------------
 1 file changed, 180 deletions(-)
 delete mode 100644 scripts/coccinelle/misc/simple_return.cocci
scripts/coccinelle/misc/simple_return.cocci deleted +0 -180
deleted file mode 100644
index e8b6313b116f..000000000000
@@ -1,180 +0,0 @@
-/// Simplify a trivial if-return sequence.  Possibly combine with a
-/// preceding function call.
-///
-// Confidence: High
-// Copyright: (C) 2014 Julia Lawall, INRIA/LIP6.  GPLv2.
-// Copyright: (C) 2014 Gilles Muller, INRIA/LiP6.  GPLv2.
-// URL: http://coccinelle.lip6.fr/
-// Comments:
-// Options: --no-includes --include-headers
-
-virtual patch
-virtual context
-virtual org
-virtual report
-
-@r depends on patch@
-local idexpression e;
-identifier i,f,fn;
-@@
-
-fn(...) { <...
-- e@i =
-+ return
-    f(...);
--if (i != 0) return i;
--return 0;
-...> }
-
-@depends on patch@
-identifier r.i;
-type t;
-@@
-
--t i;
- ... when != i
-
-@depends on patch@
-expression e;
-@@
-
--if (e != 0)
-   return e;
--return 0;
-
-// -----------------------------------------------------------------------
-
-@s1 depends on context || org || report@
-local idexpression e;
-identifier i,f,fn;
-position p,p1,p2;
-@@
-
-fn(...) { <...
-* e@i@p = f(...);
-  if (\(i@p1 != 0\|i@p2 < 0\))
-     return i;
-  return 0;
-...> }
-
-@s2 depends on context || org || report forall@
-identifier s1.i;
-type t;
-position q,s1.p;
-expression e,f;
-@@
-
-* t i@q;
-  ... when != i
-  e@p = f(...);
-
-@s3 depends on context || org || report@
-expression e;
-position p1!=s1.p1;
-position p2!=s1.p2;
-@@
-
-*if (\(e@p1 != 0\|e@p2 < 0\))
-   return e;
- return 0;
-
-// -----------------------------------------------------------------------
-
-@script:python depends on org@
-p << s1.p;
-p1 << s1.p1;
-q << s2.q;
-@@
-
-cocci.print_main("decl",q)
-cocci.print_secs("use",p)
-cocci.include_match(False)
-
-@script:python depends on org@
-p << s1.p;
-p2 << s1.p2;
-q << s2.q;
-@@
-
-cocci.print_main("decl",q)
-cocci.print_secs("use with questionable test",p)
-cocci.include_match(False)
-
-@script:python depends on org@
-p << s1.p;
-p1 << s1.p1;
-@@
-
-cocci.print_main("use",p)
-
-@script:python depends on org@
-p << s1.p;
-p2 << s1.p2;
-@@
-
-cocci.print_main("use with questionable test",p)
-
-@script:python depends on org@
-p << s3.p1;
-@@
-
-cocci.print_main("test",p)
-
-@script:python depends on org@
-p << s3.p2;
-@@
-
-cocci.print_main("questionable test",p)
-
-// -----------------------------------------------------------------------
-
-@script:python depends on report@
-p << s1.p;
-p1 << s1.p1;
-q << s2.q;
-@@
-
-msg = "WARNING: end returns can be simpified and declaration on line %s can be dropped" % (q[0].line)
-coccilib.report.print_report(p[0],msg)
-cocci.include_match(False)
-
-@script:python depends on report@
-p << s1.p;
-p1 << s1.p1;
-q << s2.q
-;
-@@
-
-msg = "WARNING: end returns may be simpified if negative or 0 value and declaration on line %s can be dropped" % (q[0].line)
-coccilib.report.print_report(p[0],msg)
-cocci.include_match(False)
-
-@script:python depends on report@
-p << s1.p;
-p1 << s1.p1;
-@@
-
-msg = "WARNING: end returns can be simpified"
-coccilib.report.print_report(p[0],msg)
-
-@script:python depends on report@
-p << s1.p;
-p2 << s1.p2;
-@@
-
-msg = "WARNING: end returns can be simpified if negative or 0 value"
-coccilib.report.print_report(p[0],msg)
-
-@script:python depends on report@
-p << s3.p1;
-@@
-
-msg = "WARNING: end returns can be simpified"
-coccilib.report.print_report(p[0],msg)
-
-@script:python depends on report@
-p << s3.p2;
-@@
-
-msg = "WARNING: end returns can be simpified if tested value is negative or 0"
-coccilib.report.print_report(p[0],msg)
Signature · 1 line
2.4.9
#1 lore raw
Quoted 10 lines · Johan Hovold

This effectively reverts 932058a5d5f9 ("coccinelle: misc: semantic patch to delete overly complex return code processing").

There can be both symmetry and readability reasons for not wanting to do the final function call as part of the return statement and to maintain a clear separation of success and error paths.

Since this is in no way mandated by the coding standard, let's just remove this semantic patch to avoid having "clean up" patches being posted over and over in response to these Coccinelle warnings.

What do you mean by "posted"? Are you referring to 0-day build testing or individual usage of make coccicheck? Maybe it would make sense to remove the semantic patch from 0-day build testing but leave it in the kernel, perhaps removing the < 0 case because that one in practice doesn't seem to turn up much that is useful?

Perhaps it could also be improved to detect a previous != 0 case and then not return a warning. On some functions, this change can make some nice simplifications.

julia

Quoted 196 lines
---
 scripts/coccinelle/misc/simple_return.cocci | 180 ----------------------------
 1 file changed, 180 deletions(-)
 delete mode 100644 scripts/coccinelle/misc/simple_return.cocci
scripts/coccinelle/misc/simple_return.cocci deleted +0 -180
deleted file mode 100644
index e8b6313b116f..000000000000
@@ -1,180 +0,0 @@
-/// Simplify a trivial if-return sequence.  Possibly combine with a
-/// preceding function call.
-///
-// Confidence: High
-// Copyright: (C) 2014 Julia Lawall, INRIA/LIP6.  GPLv2.
-// Copyright: (C) 2014 Gilles Muller, INRIA/LiP6.  GPLv2.
-// URL: http://coccinelle.lip6.fr/
-// Comments:
-// Options: --no-includes --include-headers
-
-virtual patch
-virtual context
-virtual org
-virtual report
-
- at r depends on patch@
-local idexpression e;
-identifier i,f,fn;
-@@
-
-fn(...) { <...
-- e at i =
-+ return
-    f(...);
--if (i != 0) return i;
--return 0;
-...> }
-
- at depends on patch@
-identifier r.i;
-type t;
-@@
-
--t i;
- ... when != i
-
- at depends on patch@
-expression e;
-@@
-
--if (e != 0)
-   return e;
--return 0;
-
-// -----------------------------------------------------------------------
-
- at s1 depends on context || org || report@
-local idexpression e;
-identifier i,f,fn;
-position p,p1,p2;
-@@
-
-fn(...) { <...
-* e at i@p = f(...);
-  if (\(i at p1 != 0\|i at p2 < 0\))
-     return i;
-  return 0;
-...> }
-
- at s2 depends on context || org || report forall@
-identifier s1.i;
-type t;
-position q,s1.p;
-expression e,f;
-@@
-
-* t i at q;
-  ... when != i
-  e at p = f(...);
-
- at s3 depends on context || org || report@
-expression e;
-position p1!=s1.p1;
-position p2!=s1.p2;
-@@
-
-*if (\(e at p1 != 0\|e at p2 < 0\))
-   return e;
- return 0;
-
-// -----------------------------------------------------------------------
-
- at script:python depends on org@
-p << s1.p;
-p1 << s1.p1;
-q << s2.q;
-@@
-
-cocci.print_main("decl",q)
-cocci.print_secs("use",p)
-cocci.include_match(False)
-
- at script:python depends on org@
-p << s1.p;
-p2 << s1.p2;
-q << s2.q;
-@@
-
-cocci.print_main("decl",q)
-cocci.print_secs("use with questionable test",p)
-cocci.include_match(False)
-
- at script:python depends on org@
-p << s1.p;
-p1 << s1.p1;
-@@
-
-cocci.print_main("use",p)
-
- at script:python depends on org@
-p << s1.p;
-p2 << s1.p2;
-@@
-
-cocci.print_main("use with questionable test",p)
-
- at script:python depends on org@
-p << s3.p1;
-@@
-
-cocci.print_main("test",p)
-
- at script:python depends on org@
-p << s3.p2;
-@@
-
-cocci.print_main("questionable test",p)
-
-// -----------------------------------------------------------------------
-
- at script:python depends on report@
-p << s1.p;
-p1 << s1.p1;
-q << s2.q;
-@@
-
-msg = "WARNING: end returns can be simpified and declaration on line %s can be dropped" % (q[0].line)
-coccilib.report.print_report(p[0],msg)
-cocci.include_match(False)
-
- at script:python depends on report@
-p << s1.p;
-p1 << s1.p1;
-q << s2.q
-;
-@@
-
-msg = "WARNING: end returns may be simpified if negative or 0 value and declaration on line %s can be dropped" % (q[0].line)
-coccilib.report.print_report(p[0],msg)
-cocci.include_match(False)
-
- at script:python depends on report@
-p << s1.p;
-p1 << s1.p1;
-@@
-
-msg = "WARNING: end returns can be simpified"
-coccilib.report.print_report(p[0],msg)
-
- at script:python depends on report@
-p << s1.p;
-p2 << s1.p2;
-@@
-
-msg = "WARNING: end returns can be simpified if negative or 0 value"
-coccilib.report.print_report(p[0],msg)
-
- at script:python depends on report@
-p << s3.p1;
-@@
-
-msg = "WARNING: end returns can be simpified"
-coccilib.report.print_report(p[0],msg)
-
- at script:python depends on report@
-p << s3.p2;
-@@
-
-msg = "WARNING: end returns can be simpified if tested value is negative or 0"
-coccilib.report.print_report(p[0],msg)
Signature · 1 line
2.4.9
#2 lore raw
Quoted 13 lines · Julia Lawall
Quoted 10 lines · Johan Hovold

This effectively reverts 932058a5d5f9 ("coccinelle: misc: semantic patch to delete overly complex return code processing").

There can be both symmetry and readability reasons for not wanting to do the final function call as part of the return statement and to maintain a clear separation of success and error paths.

Since this is in no way mandated by the coding standard, let's just remove this semantic patch to avoid having "clean up" patches being posted over and over in response to these Coccinelle warnings.

Quoted 5 lines

What do you mean by "posted"? Are you referring to 0-day build testing or individual usage of make coccicheck? Maybe it would make sense to remove the semantic patch from 0-day build testing but leave it in the kernel, perhaps removing the < 0 case because that one in practice doesn't seem to turn up much that is useful?

Individuals running coccicheck on in-kernel code and posting patches to "fix warnings", where the end result is not necessarily an improvement.

But I don't think these warnings should be enabled for 0-day build testing either as it is should be up to the author to decide what style to prefer in each case.

Quoted 3 lines

Perhaps it could also be improved to detect a previous != 0 case and then not return a warning. On some functions, this change can make some nice simplifications.

Yes, that would at least improve things.

I don't think warnings should be generated at all for the following code:

{
	int ret;

	ret = init_a(...);
	if (ret)
		return ret;

	ret = init_b(...);
	if (ret)
		return ret;

	return 0;
}

as it is (at least to me) preferred over:

{
	int ret;

	ret = init_a(...);
	if (ret)
		return ret;

	return init_b(...);
}

for symmetry and readability reasons (e.g. I don't have to look at init_b to figure out what the functions returns). And with a long parameter list to init_b with line breaks, this would look even worse.

But either way, it should be up to the author of the code to decide what style to use.

Thanks,
Johan

#3 lore raw

Do you consider that this function would be better off in two lines?

static int mxt_acquire_irq(struct mxt_data *data)

{
        int error;

        enable_irq(data->irq);

        error = mxt_process_messages_until_invalid(data);
        if (error)
                return error;

        return 0;
}

Would simplifying the code at the end of the following function be helpful or not?

static int adnp_gpio_setup(struct adnp *adnp, unsigned int num_gpios)

{
        struct gpio_chip *chip = &adnp->gpio;
        int err;

        adnp->reg_shift = get_count_order(num_gpios) - 3;

        chip->direction_input = adnp_gpio_direction_input;
        chip->direction_output = adnp_gpio_direction_output;
        chip->get = adnp_gpio_get;
        chip->set = adnp_gpio_set;
        chip->can_sleep = true;

        if (IS_ENABLED(CONFIG_DEBUG_FS))
                chip->dbg_show = adnp_gpio_dbg_show;

        chip->base = -1;
        chip->ngpio = num_gpios;
        chip->label = adnp->client->name;
        chip->dev = &adnp->client->dev;
        chip->of_node = chip->dev->of_node;
        chip->owner = THIS_MODULE;

        err = gpiochip_add(chip);
        if (err)
                return err;

        return 0;
}

thanks,
julia

Acked-by
#5 lore raw
In reply to Johan Hovold

Perhaps there is a more restricted version that can be acceptable, but I'm OK with dropping the current version.

julia

Quoted 72 lines · Johan Hovold
Quoted 13 lines · Julia Lawall
Quoted 10 lines · Johan Hovold

This effectively reverts 932058a5d5f9 ("coccinelle: misc: semantic patch to delete overly complex return code processing").

There can be both symmetry and readability reasons for not wanting to do the final function call as part of the return statement and to maintain a clear separation of success and error paths.

Since this is in no way mandated by the coding standard, let's just remove this semantic patch to avoid having "clean up" patches being posted over and over in response to these Coccinelle warnings.

Quoted 5 lines

What do you mean by "posted"? Are you referring to 0-day build testing or individual usage of make coccicheck? Maybe it would make sense to remove the semantic patch from 0-day build testing but leave it in the kernel, perhaps removing the < 0 case because that one in practice doesn't seem to turn up much that is useful?

Individuals running coccicheck on in-kernel code and posting patches to "fix warnings", where the end result is not necessarily an improvement.

But I don't think these warnings should be enabled for 0-day build testing either as it is should be up to the author to decide what style to prefer in each case.

Quoted 3 lines

Perhaps it could also be improved to detect a previous != 0 case and then not return a warning. On some functions, this change can make some nice simplifications.

Yes, that would at least improve things.

I don't think warnings should be generated at all for the following code:

{
	int ret;

	ret = init_a(...);
	if (ret)
		return ret;

	ret = init_b(...);
	if (ret)
		return ret;

	return 0;
}

as it is (at least to me) preferred over:

{
	int ret;

	ret = init_a(...);
	if (ret)
		return ret;

	return init_b(...);
}

for symmetry and readability reasons (e.g. I don't have to look at init_b to figure out what the functions returns). And with a long parameter list to init_b with line breaks, this would look even worse.

But either way, it should be up to the author of the code to decide what style to use.

Thanks,
Johan

Acked-by on v1 1/1
#7 lore raw
In reply to Johan Hovold

Should have acked this message...

Quoted 207 lines · Johan Hovold

This effectively reverts 932058a5d5f9 ("coccinelle: misc: semantic patch to delete overly complex return code processing").

There can be both symmetry and readability reasons for not wanting to do the final function call as part of the return statement and to maintain a clear separation of success and error paths.

Since this is in no way mandated by the coding standard, let's just remove this semantic patch to avoid having "clean up" patches being posted over and over in response to these Coccinelle warnings.

---
 scripts/coccinelle/misc/simple_return.cocci | 180 ----------------------------
 1 file changed, 180 deletions(-)
 delete mode 100644 scripts/coccinelle/misc/simple_return.cocci
scripts/coccinelle/misc/simple_return.cocci deleted +0 -181
deleted file mode 100644
index e8b6313b116f..000000000000
@@ -1,180 +0,0 @@
-/// Simplify a trivial if-return sequence.  Possibly combine with a
-/// preceding function call.
-///
-// Confidence: High
-// Copyright: (C) 2014 Julia Lawall, INRIA/LIP6.  GPLv2.
-// Copyright: (C) 2014 Gilles Muller, INRIA/LiP6.  GPLv2.
-// URL: http://coccinelle.lip6.fr/
-// Comments:
-// Options: --no-includes --include-headers
-
-virtual patch
-virtual context
-virtual org
-virtual report
-
- at r depends on patch@
-local idexpression e;
-identifier i,f,fn;
-@@
-
-fn(...) { <...
-- e at i =
-+ return
-    f(...);
--if (i != 0) return i;
--return 0;
-...> }
-
- at depends on patch@
-identifier r.i;
-type t;
-@@
-
--t i;
- ... when != i
-
- at depends on patch@
-expression e;
-@@
-
--if (e != 0)
-   return e;
--return 0;
-
-// -----------------------------------------------------------------------
-
- at s1 depends on context || org || report@
-local idexpression e;
-identifier i,f,fn;
-position p,p1,p2;
-@@
-
-fn(...) { <...
-* e at i@p = f(...);
-  if (\(i at p1 != 0\|i at p2 < 0\))
-     return i;
-  return 0;
-...> }
-
- at s2 depends on context || org || report forall@
-identifier s1.i;
-type t;
-position q,s1.p;
-expression e,f;
-@@
-
-* t i at q;
-  ... when != i
-  e at p = f(...);
-
- at s3 depends on context || org || report@
-expression e;
-position p1!=s1.p1;
-position p2!=s1.p2;
-@@
-
-*if (\(e at p1 != 0\|e at p2 < 0\))
-   return e;
- return 0;
-
-// -----------------------------------------------------------------------
-
- at script:python depends on org@
-p << s1.p;
-p1 << s1.p1;
-q << s2.q;
-@@
-
-cocci.print_main("decl",q)
-cocci.print_secs("use",p)
-cocci.include_match(False)
-
- at script:python depends on org@
-p << s1.p;
-p2 << s1.p2;
-q << s2.q;
-@@
-
-cocci.print_main("decl",q)
-cocci.print_secs("use with questionable test",p)
-cocci.include_match(False)
-
- at script:python depends on org@
-p << s1.p;
-p1 << s1.p1;
-@@
-
-cocci.print_main("use",p)
-
- at script:python depends on org@
-p << s1.p;
-p2 << s1.p2;
-@@
-
-cocci.print_main("use with questionable test",p)
-
- at script:python depends on org@
-p << s3.p1;
-@@
-
-cocci.print_main("test",p)
-
- at script:python depends on org@
-p << s3.p2;
-@@
-
-cocci.print_main("questionable test",p)
-
-// -----------------------------------------------------------------------
-
- at script:python depends on report@
-p << s1.p;
-p1 << s1.p1;
-q << s2.q;
-@@
-
-msg = "WARNING: end returns can be simpified and declaration on line %s can be dropped" % (q[0].line)
-coccilib.report.print_report(p[0],msg)
-cocci.include_match(False)
-
- at script:python depends on report@
-p << s1.p;
-p1 << s1.p1;
-q << s2.q
-;
-@@
-
-msg = "WARNING: end returns may be simpified if negative or 0 value and declaration on line %s can be dropped" % (q[0].line)
-coccilib.report.print_report(p[0],msg)
-cocci.include_match(False)
-
- at script:python depends on report@
-p << s1.p;
-p1 << s1.p1;
-@@
-
-msg = "WARNING: end returns can be simpified"
-coccilib.report.print_report(p[0],msg)
-
- at script:python depends on report@
-p << s1.p;
-p2 << s1.p2;
-@@
-
-msg = "WARNING: end returns can be simpified if negative or 0 value"
-coccilib.report.print_report(p[0],msg)
-
- at script:python depends on report@
-p << s3.p1;
-@@
-
-msg = "WARNING: end returns can be simpified"
-coccilib.report.print_report(p[0],msg)
-
- at script:python depends on report@
-p << s3.p2;
-@@
-
-msg = "WARNING: end returns can be simpified if tested value is negative or 0"
-coccilib.report.print_report(p[0],msg)
--

2.4.9

#4 lore raw
In reply to Julia Lawall
Quoted 14 lines · Julia Lawall

Do you consider that this function would be better off in two lines?

static int mxt_acquire_irq(struct mxt_data *data)

{
        int error;

        enable_irq(data->irq);

        error = mxt_process_messages_until_invalid(data);
        if (error)
                return error;

        return 0;
}

Actually no, but again I'd say it's up to the author to decide.

Quoted 24 lines

Would simplifying the code at the end of the following function be helpful or not?

static int adnp_gpio_setup(struct adnp *adnp, unsigned int num_gpios)

{
        struct gpio_chip *chip = &adnp->gpio;
        int err;

        adnp->reg_shift = get_count_order(num_gpios) - 3;

        chip->direction_input = adnp_gpio_direction_input;
        chip->direction_output = adnp_gpio_direction_output;
        chip->get = adnp_gpio_get;
        chip->set = adnp_gpio_set;
        chip->can_sleep = true;

        if (IS_ENABLED(CONFIG_DEBUG_FS))
                chip->dbg_show = adnp_gpio_dbg_show;

        chip->base = -1;
        chip->ngpio = num_gpios;
        chip->label = adnp->client->name;
        chip->dev = &adnp->client->dev;
        chip->of_node = chip->dev->of_node;
Quoted 8 lines
        chip->owner = THIS_MODULE;

        err = gpiochip_add(chip);
        if (err)
                return err;

        return 0;
}

I think this is just fine as is as well.

Thanks,
Johan

#8 lore raw
In reply to Julia Lawall
Quoted 18 lines · Julia Lawall

Should have acked this message...

Quoted 12 lines · Johan Hovold

This effectively reverts 932058a5d5f9 ("coccinelle: misc: semantic patch to delete overly complex return code processing").

There can be both symmetry and readability reasons for not wanting to do the final function call as part of the return statement and to maintain a clear separation of success and error paths.

Since this is in no way mandated by the coding standard, let's just remove this semantic patch to avoid having "clean up" patches being posted over and over in response to these Coccinelle warnings.

Is anyone planning on picking this patch up?

Thanks,
Johan

#9 lore raw

Dne 28.10.2015 v 10:54 Johan Hovold napsal(a):

Quoted 14 lines · Julia Lawall
Quoted 13 lines

Should have acked this message...

Quoted 7 lines · Johan Hovold

This effectively reverts 932058a5d5f9 ("coccinelle: misc: semantic patch to delete overly complex return code processing").

There can be both symmetry and readability reasons for not wanting to do the final function call as part of the return statement and to maintain a clear separation of success and error paths.

Quoted 7 lines
Quoted 5 lines
Quoted 5 lines

Since this is in no way mandated by the coding standard, let's just remove this semantic patch to avoid having "clean up" patches being posted over and over in response to these Coccinelle warnings.

Is anyone planning on picking this patch up?

It is already in kbuild.git as commit 1a617a8475e8, I apparenly forgot to reply to this thread.

Michal

#10 lore raw
Quoted 19 lines · Michal Marek

Dne 28.10.2015 v 10:54 Johan Hovold napsal(a):

Quoted 18 lines · Julia Lawall
Quoted 17 lines

Should have acked this message...

Quoted 11 lines · Johan Hovold

This effectively reverts 932058a5d5f9 ("coccinelle: misc: semantic patch to delete overly complex return code processing").

There can be both symmetry and readability reasons for not wanting to do the final function call as part of the return statement and to maintain a clear separation of success and error paths.

Since this is in no way mandated by the coding standard, let's just remove this semantic patch to avoid having "clean up" patches being posted over and over in response to these Coccinelle warnings.

Quoted 6 lines
Quoted 3 lines
Quoted 1 line
Quoted 1 line

Is anyone planning on picking this patch up?

It is already in kbuild.git as commit 1a617a8475e8, I apparenly forgot to reply to this thread.

Ok, great.

Thanks,
Johan