Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1640520 > unrolled thread

[PATCH] mdio: mux: fix device_node_continue.cocci warnings

Back to article view | Back to linux.kernel

Contents

  [PATCH] mdio: mux: fix device_node_continue.cocci warnings Julia Lawall <julia.lawall@lip6.fr> - 2017-05-12 17:00 +0200
    Re: [PATCH] mdio: mux: fix device_node_continue.cocci warnings David Miller <davem@davemloft.net> - 2017-05-12 18:30 +0200
      Re: [PATCH] mdio: mux: fix device_node_continue.cocci warnings Florian Fainelli <f.fainelli@gmail.com> - 2017-05-13 01:00 +0200
        Re: [PATCH] mdio: mux: fix device_node_continue.cocci warnings Julia Lawall <julia.lawall@lip6.fr> - 2017-05-13 02:50 +0200
        Re: [PATCH] mdio: mux: fix device_node_continue.cocci warnings Jon Mason <jon.mason@broadcom.com> - 2017-05-15 19:40 +0200
          Re: [PATCH] mdio: mux: fix device_node_continue.cocci warnings David Miller <davem@davemloft.net> - 2017-05-15 20:30 +0200

#1640520 — [PATCH] mdio: mux: fix device_node_continue.cocci warnings

Device node iterators put the previous value of the index variable, so an
explicit put causes a double put.

In particular, of_mdiobus_register can fail before doing anything
interesting, so one could view it as a no-op from the reference count
point of view.

Generated by: scripts/coccinelle/iterators/device_node_continue.cocci

CC: Jon Mason <jon.mason@broadcom.com>
Signed-off-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
---

tree:
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git
master
head:   8785ded64cfb68b8d8b2583c7c1fc611f99eabf2
commit: b60161668199ac62011c024adc9e66713b9554e7 [13999/14120] mdio: mux:

 mdio-mux.c |    1 -
 1 file changed, 1 deletion(-)

--- a/drivers/net/phy/mdio-mux.c
+++ b/drivers/net/phy/mdio-mux.c
@@ -169,7 +169,6 @@ int mdio_mux_init(struct device *dev,
 		if (r) {
 			mdiobus_free(cb->mii_bus);
 			devm_kfree(dev, cb);
-			of_node_put(child_bus_node);
 		} else {
 			cb->next = pb->children;
 			pb->children = cb;

[toc] | [next] | [standalone]

#1640570

From: Julia Lawall <julia.lawall@lip6.fr>
Date: Fri, 12 May 2017 22:54:23 +0800 (SGT)

> Device node iterators put the previous value of the index variable, so an
> explicit put causes a double put.
 ...
> @@ -169,7 +169,6 @@ int mdio_mux_init(struct device *dev,
>  		if (r) {
>  			mdiobus_free(cb->mii_bus);
>  			devm_kfree(dev, cb);
> -			of_node_put(child_bus_node);
>  		} else {

I think we're instead simply missing a break; statement here.

[toc] | [prev] | [next] | [standalone]

#1640795

On 05/12/2017 09:22 AM, David Miller wrote:
> From: Julia Lawall <julia.lawall@lip6.fr>
> Date: Fri, 12 May 2017 22:54:23 +0800 (SGT)
> 
>> Device node iterators put the previous value of the index variable, so an
>> explicit put causes a double put.
>  ...
>> @@ -169,7 +169,6 @@ int mdio_mux_init(struct device *dev,
>>  		if (r) {
>>  			mdiobus_free(cb->mii_bus);
>>  			devm_kfree(dev, cb);
>> -			of_node_put(child_bus_node);
>>  		} else {
> 
> I think we're instead simply missing a break; statement here.

It's kind of questionable, if we have an error initializing one of our
child MDIO bus controller (child from the perspective of the MDIO mux,
boy this is getting complicated...), should we keep on going, or should
we abort entirely and rollback what we have successfully registered?

I don't think Julia's patch makes thing worse, in that if we had to
rollback, we would not be doing this correctly now anyway.

Jon, what do you think?
-- 
Florian

[toc] | [prev] | [next] | [standalone]

#1640817

On Fri, 12 May 2017, Florian Fainelli wrote:

> On 05/12/2017 09:22 AM, David Miller wrote:
> > From: Julia Lawall <julia.lawall@lip6.fr>
> > Date: Fri, 12 May 2017 22:54:23 +0800 (SGT)
> >
> >> Device node iterators put the previous value of the index variable, so an
> >> explicit put causes a double put.
> >  ...
> >> @@ -169,7 +169,6 @@ int mdio_mux_init(struct device *dev,
> >>  		if (r) {
> >>  			mdiobus_free(cb->mii_bus);
> >>  			devm_kfree(dev, cb);
> >> -			of_node_put(child_bus_node);
> >>  		} else {
> >
> > I think we're instead simply missing a break; statement here.
>
> It's kind of questionable, if we have an error initializing one of our
> child MDIO bus controller (child from the perspective of the MDIO mux,
> boy this is getting complicated...), should we keep on going, or should
> we abort entirely and rollback what we have successfully registered?
>
> I don't think Julia's patch makes thing worse, in that if we had to
> rollback, we would not be doing this correctly now anyway.

Just to be clear, if you want the break instead, then you need to keep the
put.

julia

>
> Jon, what do you think?
> --
> Florian
>

[toc] | [prev] | [next] | [standalone]

#1641926

On Fri, May 12, 2017 at 6:52 PM, Florian Fainelli <f.fainelli@gmail.com> wrote:
> On 05/12/2017 09:22 AM, David Miller wrote:
>> From: Julia Lawall <julia.lawall@lip6.fr>
>> Date: Fri, 12 May 2017 22:54:23 +0800 (SGT)
>>
>>> Device node iterators put the previous value of the index variable, so an
>>> explicit put causes a double put.
>>  ...
>>> @@ -169,7 +169,6 @@ int mdio_mux_init(struct device *dev,
>>>              if (r) {
>>>                      mdiobus_free(cb->mii_bus);
>>>                      devm_kfree(dev, cb);
>>> -                    of_node_put(child_bus_node);
>>>              } else {
>>
>> I think we're instead simply missing a break; statement here.
>
> It's kind of questionable, if we have an error initializing one of our
> child MDIO bus controller (child from the perspective of the MDIO mux,
> boy this is getting complicated...), should we keep on going, or should
> we abort entirely and rollback what we have successfully registered?
>
> I don't think Julia's patch makes thing worse, in that if we had to
> rollback, we would not be doing this correctly now anyway.
>
> Jon, what do you think?

If every other case is fatal, then it is odd that this one is
permissive.  I think we should go 100% one way or the other.  So, the
options here are to:
1.  Encounter an error, unroll any mallocs, etc created by this entry,
but continue on to the next entry and return success if any are
created
2.  Encounter an error, unroll any mallocs, etc created by this entry
and any others that were created, and return an error
3.  Encounter an error, unroll any mallocs, etc created by this entry,
exit and return success if any are created

#1 would be the most accepting of any errors encountered
#2 would identify any poorly written DTs by breaking their currently
working functionality (though we should add some error messages to let
them know why)
#3 matches the suggestion by David Miller, and would be a hybrid of #1
and #2 in outcome

I would prefer #1, as I would not want to break something that was
currently working.  However, I think we should add much error logging
here to let people know their DT is hosed (instead of silently
working).  So, this would mean applying Julia's patch, and I'll do a
follow-on to change the breaks to continues and add the error logging
(assuming others agree with me).

Thanks,
Jon

[toc] | [prev] | [next] | [standalone]

#1641965

From: Jon Mason <jon.mason@broadcom.com>
Date: Mon, 15 May 2017 13:37:09 -0400

> I would prefer #1, as I would not want to break something that was
> currently working.  However, I think we should add much error logging
> here to let people know their DT is hosed (instead of silently
> working).  So, this would mean applying Julia's patch, and I'll do a
> follow-on to change the breaks to continues and add the error logging
> (assuming others agree with me).

Ok, I've applied Julia's patch.

I agree that we shouldn't fail the whole list just because one does.
And yes, we should emit enough diagnostics so that people can figure
out what the problem is.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web