From: Markus Elfring <[email protected]>
Date: Sat, 2 Mar 2024 11:44:17 +0100
Add a jump target so that a bit of exception handling can be better reused
at the end of this function implementation.
This issue was transformed by using the Coccinelle software.
Signed-off-by: Markus Elfring <[email protected]>
---
drivers/input/misc/iqs626a.c | 13 +++++++------
1 file changed, 7 insertions(+), 6 deletions(-)
diff --git a/drivers/input/misc/iqs626a.c b/drivers/input/misc/iqs626a.c
index 0dab54d3a060..fa9570755f7b 100644
--- a/drivers/input/misc/iqs626a.c
+++ b/drivers/input/misc/iqs626a.c
@@ -530,8 +530,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
dev_err(&client->dev,
"Invalid input type: %u\n",
val);
- fwnode_handle_put(ev_node);
- return -EINVAL;
+ goto put_fwnode;
}
iqs626->kp_type[ch_id][i] = val;
@@ -545,8 +544,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
dev_err(&client->dev,
"Invalid %s channel hysteresis: %u\n",
fwnode_get_name(ch_node), val);
- fwnode_handle_put(ev_node);
- return -EINVAL;
+ goto put_fwnode;
}
if (i == IQS626_EVENT_DEEP_DN ||
@@ -566,8 +564,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
dev_err(&client->dev,
"Invalid %s channel threshold: %u\n",
fwnode_get_name(ch_node), val);
- fwnode_handle_put(ev_node);
- return -EINVAL;
+ goto put_fwnode;
}
if (ch_id == IQS626_CH_HALL)
@@ -580,6 +577,10 @@ iqs626_parse_events(struct iqs626_private *iqs626,
}
return 0;
+
+put_fwnode:
+ fwnode_handle_put(ev_node);
+ return -EINVAL;
}
static noinline_for_stack int
--
2.44.0
Hi Markus,
On Sat, Mar 02, 2024 at 12:42:08PM +0100, Markus Elfring wrote:
> From: Markus Elfring <[email protected]>
> Date: Sat, 2 Mar 2024 11:44:17 +0100
>
> Add a jump target so that a bit of exception handling can be better reused
> at the end of this function implementation.
>
> This issue was transformed by using the Coccinelle software.
>
> Signed-off-by: Markus Elfring <[email protected]>
> ---
> drivers/input/misc/iqs626a.c | 13 +++++++------
> 1 file changed, 7 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/input/misc/iqs626a.c b/drivers/input/misc/iqs626a.c
> index 0dab54d3a060..fa9570755f7b 100644
> --- a/drivers/input/misc/iqs626a.c
> +++ b/drivers/input/misc/iqs626a.c
> @@ -530,8 +530,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> dev_err(&client->dev,
> "Invalid input type: %u\n",
> val);
> - fwnode_handle_put(ev_node);
> - return -EINVAL;
> + goto put_fwnode;
> }
>
> iqs626->kp_type[ch_id][i] = val;
> @@ -545,8 +544,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> dev_err(&client->dev,
> "Invalid %s channel hysteresis: %u\n",
> fwnode_get_name(ch_node), val);
> - fwnode_handle_put(ev_node);
> - return -EINVAL;
> + goto put_fwnode;
> }
>
> if (i == IQS626_EVENT_DEEP_DN ||
> @@ -566,8 +564,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> dev_err(&client->dev,
> "Invalid %s channel threshold: %u\n",
> fwnode_get_name(ch_node), val);
> - fwnode_handle_put(ev_node);
> - return -EINVAL;
> + goto put_fwnode;
> }
>
> if (ch_id == IQS626_CH_HALL)
> @@ -580,6 +577,10 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> }
>
> return 0;
> +
> +put_fwnode:
> + fwnode_handle_put(ev_node);
> + return -EINVAL;
> }
>
> static noinline_for_stack int
> --
> 2.44.0
>
Thank you for this patch, but it seems like a NAK to me. I think this is
a matter of personal preference, and according to mine, it is much more
confusing to insert a goto label after a function's primary return path
than to have 2-3 calls to fwnode_handle_put().
If you feel strongly otherwise, then I would suggest a helper function as
recommended by Dmitry in another thread. However, maybe that helper should
live in the driver core, as I suspect this driver is not the only place we
can avoid calling fwnode_handle_put() in an error path that returns an int.
Kind regards,
Jeff LaBundy
On Sun, Mar 03, 2024 at 09:12:16PM -0600, Jeff LaBundy wrote:
> Hi Markus,
>
> On Sat, Mar 02, 2024 at 12:42:08PM +0100, Markus Elfring wrote:
> > From: Markus Elfring <[email protected]>
> > Date: Sat, 2 Mar 2024 11:44:17 +0100
> >
> > Add a jump target so that a bit of exception handling can be better reused
> > at the end of this function implementation.
> >
> > This issue was transformed by using the Coccinelle software.
> >
> > Signed-off-by: Markus Elfring <[email protected]>
> > ---
> > drivers/input/misc/iqs626a.c | 13 +++++++------
> > 1 file changed, 7 insertions(+), 6 deletions(-)
> >
> > diff --git a/drivers/input/misc/iqs626a.c b/drivers/input/misc/iqs626a.c
> > index 0dab54d3a060..fa9570755f7b 100644
> > --- a/drivers/input/misc/iqs626a.c
> > +++ b/drivers/input/misc/iqs626a.c
> > @@ -530,8 +530,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> > dev_err(&client->dev,
> > "Invalid input type: %u\n",
> > val);
> > - fwnode_handle_put(ev_node);
> > - return -EINVAL;
> > + goto put_fwnode;
> > }
> >
> > iqs626->kp_type[ch_id][i] = val;
> > @@ -545,8 +544,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> > dev_err(&client->dev,
> > "Invalid %s channel hysteresis: %u\n",
> > fwnode_get_name(ch_node), val);
> > - fwnode_handle_put(ev_node);
> > - return -EINVAL;
> > + goto put_fwnode;
> > }
> >
> > if (i == IQS626_EVENT_DEEP_DN ||
> > @@ -566,8 +564,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> > dev_err(&client->dev,
> > "Invalid %s channel threshold: %u\n",
> > fwnode_get_name(ch_node), val);
> > - fwnode_handle_put(ev_node);
> > - return -EINVAL;
> > + goto put_fwnode;
> > }
> >
> > if (ch_id == IQS626_CH_HALL)
> > @@ -580,6 +577,10 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> > }
> >
> > return 0;
> > +
> > +put_fwnode:
> > + fwnode_handle_put(ev_node);
> > + return -EINVAL;
> > }
> >
> > static noinline_for_stack int
> > --
> > 2.44.0
> >
>
> Thank you for this patch, but it seems like a NAK to me. I think this is
> a matter of personal preference, and according to mine, it is much more
> confusing to insert a goto label after a function's primary return path
> than to have 2-3 calls to fwnode_handle_put().
>
> If you feel strongly otherwise, then I would suggest a helper function as
> recommended by Dmitry in another thread. However, maybe that helper should
> live in the driver core, as I suspect this driver is not the only place we
> can avoid calling fwnode_handle_put() in an error path that returns an int.
Yes, it should go into include/linux/fwnode.h, something like
DEFINE_FREE(fwnode, struct fwnode_handle *, if (_T) fwnode_hanlde_put(_T));
Then drivers can do:
struct fwnode_handle *ev_node __free(fwnode) =
fwnode_get_named_child_node(ch_node, ev_name);
and have it automatically be "put" once execution leaves the variable
scope.
Ah, we actually already have it defined in include/linux/property.h, all
the better.
Thanks.
--
Dmitry
On Sun, Mar 03, 2024 at 08:29:49PM -0800, Dmitry Torokhov wrote:
> On Sun, Mar 03, 2024 at 09:12:16PM -0600, Jeff LaBundy wrote:
> > Hi Markus,
> >
> > On Sat, Mar 02, 2024 at 12:42:08PM +0100, Markus Elfring wrote:
> > > From: Markus Elfring <[email protected]>
> > > Date: Sat, 2 Mar 2024 11:44:17 +0100
> > >
> > > Add a jump target so that a bit of exception handling can be better reused
> > > at the end of this function implementation.
> > >
> > > This issue was transformed by using the Coccinelle software.
> > >
> > > Signed-off-by: Markus Elfring <[email protected]>
> > > ---
> > > drivers/input/misc/iqs626a.c | 13 +++++++------
> > > 1 file changed, 7 insertions(+), 6 deletions(-)
> > >
> > > diff --git a/drivers/input/misc/iqs626a.c b/drivers/input/misc/iqs626a.c
> > > index 0dab54d3a060..fa9570755f7b 100644
> > > --- a/drivers/input/misc/iqs626a.c
> > > +++ b/drivers/input/misc/iqs626a.c
> > > @@ -530,8 +530,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> > > dev_err(&client->dev,
> > > "Invalid input type: %u\n",
> > > val);
> > > - fwnode_handle_put(ev_node);
> > > - return -EINVAL;
> > > + goto put_fwnode;
> > > }
> > >
> > > iqs626->kp_type[ch_id][i] = val;
> > > @@ -545,8 +544,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> > > dev_err(&client->dev,
> > > "Invalid %s channel hysteresis: %u\n",
> > > fwnode_get_name(ch_node), val);
> > > - fwnode_handle_put(ev_node);
> > > - return -EINVAL;
> > > + goto put_fwnode;
> > > }
> > >
> > > if (i == IQS626_EVENT_DEEP_DN ||
> > > @@ -566,8 +564,7 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> > > dev_err(&client->dev,
> > > "Invalid %s channel threshold: %u\n",
> > > fwnode_get_name(ch_node), val);
> > > - fwnode_handle_put(ev_node);
> > > - return -EINVAL;
> > > + goto put_fwnode;
> > > }
> > >
> > > if (ch_id == IQS626_CH_HALL)
> > > @@ -580,6 +577,10 @@ iqs626_parse_events(struct iqs626_private *iqs626,
> > > }
> > >
> > > return 0;
> > > +
> > > +put_fwnode:
> > > + fwnode_handle_put(ev_node);
> > > + return -EINVAL;
> > > }
> > >
> > > static noinline_for_stack int
> > > --
> > > 2.44.0
> > >
> >
> > Thank you for this patch, but it seems like a NAK to me. I think this is
> > a matter of personal preference, and according to mine, it is much more
> > confusing to insert a goto label after a function's primary return path
> > than to have 2-3 calls to fwnode_handle_put().
> >
> > If you feel strongly otherwise, then I would suggest a helper function as
> > recommended by Dmitry in another thread. However, maybe that helper should
> > live in the driver core, as I suspect this driver is not the only place we
> > can avoid calling fwnode_handle_put() in an error path that returns an int.
>
> Yes, it should go into include/linux/fwnode.h, something like
>
> DEFINE_FREE(fwnode, struct fwnode_handle *, if (_T) fwnode_hanlde_put(_T));
>
> Then drivers can do:
>
> struct fwnode_handle *ev_node __free(fwnode) =
> fwnode_get_named_child_node(ch_node, ev_name);
>
> and have it automatically be "put" once execution leaves the variable
> scope.
>
> Ah, we actually already have it defined in include/linux/property.h, all
> the better.
It's already there.
DEFINE_FREE(fwnode_handle, struct fwnode_handle *, fwnode_handle_put(_T))
I can send a patch for this. You need to be a bit carefull to move
the declaration into the correct scope for this to work. I should write
some Smatch rules for this...
regards,
dan carpenter
From: Markus Elfring <[email protected]>
Date: Mon, 4 Mar 2024 11:40:04 +0100
Scope-based resource management became supported also for this software
area by contributions of Jonathan Cameron on 2024-02-17.
device property: Add cleanup.h based fwnode_handle_put() scope based cleanup.
https://lore.kernel.org/r/[email protected]
* Thus use the attribute “__free(fwnode_handle)”.
* Reduce the scope for the local variable “ev_node” into a for loop.
Signed-off-by: Markus Elfring <[email protected]>
---
v2:
An other cleanup technique was applied as requested by Dmitry Torokhov
and Jeff LaBundy.
drivers/input/misc/iqs626a.c | 8 ++------
1 file changed, 2 insertions(+), 6 deletions(-)
diff --git a/drivers/input/misc/iqs626a.c b/drivers/input/misc/iqs626a.c
index 0dab54d3a060..86fcb5134f45 100644
--- a/drivers/input/misc/iqs626a.c
+++ b/drivers/input/misc/iqs626a.c
@@ -462,7 +462,6 @@ iqs626_parse_events(struct iqs626_private *iqs626,
{
struct iqs626_sys_reg *sys_reg = &iqs626->sys_reg;
struct i2c_client *client = iqs626->client;
- struct fwnode_handle *ev_node;
const char *ev_name;
u8 *thresh, *hyst;
unsigned int val;
@@ -501,6 +500,8 @@ iqs626_parse_events(struct iqs626_private *iqs626,
if (!iqs626_channels[ch_id].events[i])
continue;
+ struct fwnode_handle *ev_node __free(fwnode_handle);
+
if (ch_id == IQS626_CH_TP_2 || ch_id == IQS626_CH_TP_3) {
/*
* Trackpad touch events are simply described under the
@@ -530,7 +531,6 @@ iqs626_parse_events(struct iqs626_private *iqs626,
dev_err(&client->dev,
"Invalid input type: %u\n",
val);
- fwnode_handle_put(ev_node);
return -EINVAL;
}
@@ -545,7 +545,6 @@ iqs626_parse_events(struct iqs626_private *iqs626,
dev_err(&client->dev,
"Invalid %s channel hysteresis: %u\n",
fwnode_get_name(ch_node), val);
- fwnode_handle_put(ev_node);
return -EINVAL;
}
@@ -566,7 +565,6 @@ iqs626_parse_events(struct iqs626_private *iqs626,
dev_err(&client->dev,
"Invalid %s channel threshold: %u\n",
fwnode_get_name(ch_node), val);
- fwnode_handle_put(ev_node);
return -EINVAL;
}
@@ -575,8 +573,6 @@ iqs626_parse_events(struct iqs626_private *iqs626,
else
*(thresh + iqs626_events[i].th_offs) = val;
}
-
- fwnode_handle_put(ev_node);
}
return 0;
--
2.44.0
> DEFINE_FREE(fwnode_handle, struct fwnode_handle *, fwnode_handle_put(_T))
>
> I can send a patch for this. You need to be a bit carefull to move
> the declaration into the correct scope for this to work. I should write
> some Smatch rules for this...
I became also curious how available development tools will evolve further
for improved handling of scope-based resource management.
Regards,
Markus