LinuxLists.cc - [PATCH] IB/qib: remove superfluous fallthrough statements

2020-08-25 15:55:40

Subject: [PATCH] IB/qib: remove superfluous fallthrough statements

Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
were later converted to fallthrough statements by commit df561f6688fe
("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
warning about unreachable code.

Remove the fallthrough statements and replace the mass of gotos with
simple return statements to make the code terser and less bug-prone.

Addresses-Coverity: ("Unreachable code")
Fixes: 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
Signed-off-by: Alex Dewar <[email protected]>
---
drivers/infiniband/hw/qib/qib_mad.c | 52 ++++++++---------------------
1 file changed, 13 insertions(+), 39 deletions(-)

diff --git a/drivers/infiniband/hw/qib/qib_mad.c b/drivers/infiniband/hw/qib/qib_mad.c
index e7789e724f56..f83e331977f8 100644
--- a/drivers/infiniband/hw/qib/qib_mad.c
+++ b/drivers/infiniband/hw/qib/qib_mad.c
@@ -2293,76 +2293,50 @@ static int process_cc(struct ib_device *ibdev, int mad_flags,
struct ib_mad *out_mad)
{
struct ib_cc_mad *ccp = (struct ib_cc_mad *)out_mad;
- int ret;
-
*out_mad = *in_mad;

if (ccp->class_version != 2) {
ccp->status |= IB_SMP_UNSUP_VERSION;
- ret = reply((struct ib_smp *)ccp);
- goto bail;
+ return reply((struct ib_smp *)ccp);
}

switch (ccp->method) {
case IB_MGMT_METHOD_GET:
switch (ccp->attr_id) {
case IB_CC_ATTR_CLASSPORTINFO:
- ret = cc_get_classportinfo(ccp, ibdev);
- goto bail;
-
+ return cc_get_classportinfo(ccp, ibdev);
case IB_CC_ATTR_CONGESTION_INFO:
- ret = cc_get_congestion_info(ccp, ibdev, port);
- goto bail;
-
+ return cc_get_congestion_info(ccp, ibdev, port);
case IB_CC_ATTR_CA_CONGESTION_SETTING:
- ret = cc_get_congestion_setting(ccp, ibdev, port);
- goto bail;
-
+ return cc_get_congestion_setting(ccp, ibdev, port);
case IB_CC_ATTR_CONGESTION_CONTROL_TABLE:
- ret = cc_get_congestion_control_table(ccp, ibdev, port);
- goto bail;
-
- fallthrough;
+ return cc_get_congestion_control_table(ccp, ibdev, port);
default:
ccp->status |= IB_SMP_UNSUP_METH_ATTR;
- ret = reply((struct ib_smp *) ccp);
- goto bail;
+ return reply((struct ib_smp *) ccp);
}
-
case IB_MGMT_METHOD_SET:
switch (ccp->attr_id) {
case IB_CC_ATTR_CA_CONGESTION_SETTING:
- ret = cc_set_congestion_setting(ccp, ibdev, port);
- goto bail;
-
+ return cc_set_congestion_setting(ccp, ibdev, port);
case IB_CC_ATTR_CONGESTION_CONTROL_TABLE:
- ret = cc_set_congestion_control_table(ccp, ibdev, port);
- goto bail;
-
- fallthrough;
+ return cc_set_congestion_control_table(ccp, ibdev, port);
default:
ccp->status |= IB_SMP_UNSUP_METH_ATTR;
- ret = reply((struct ib_smp *) ccp);
- goto bail;
+ return reply((struct ib_smp *) ccp);
}
-
case IB_MGMT_METHOD_GET_RESP:
/*
* The ib_mad module will call us to process responses
* before checking for other consumers.
* Just tell the caller to process it normally.
*/
- ret = IB_MAD_RESULT_SUCCESS;
- goto bail;
-
- case IB_MGMT_METHOD_TRAP:
- default:
- ccp->status |= IB_SMP_UNSUP_METHOD;
- ret = reply((struct ib_smp *) ccp);
+ return IB_MAD_RESULT_SUCCESS;
}

-bail:
- return ret;
+ /* method is unsupported */
+ ccp->status |= IB_SMP_UNSUP_METHOD;
+ return reply((struct ib_smp *) ccp);
}

/**
--
2.28.0

2020-08-25 16:15:37

by Gustavo A. R. Silva

[permalink] [raw]

Subject: Re: [PATCH] IB/qib: remove superfluous fallthrough statements

On 8/25/20 10:51, Alex Dewar wrote:
> Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
> were later converted to fallthrough statements by commit df561f6688fe
> ("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
> warning about unreachable code.
>
> Remove the fallthrough statements and replace the mass of gotos with
> simple return statements to make the code terser and less bug-prone.
>

This should be split up into two separate patches: one to address the
fallthrough markings, and another one for the gotos.

Thanks
--
Gustavo

2020-08-25 16:30:15

by Joe Perches

[permalink] [raw]

Subject: Re: [PATCH] IB/qib: remove superfluous fallthrough statements

On Tue, 2020-08-25 at 11:19 -0500, Gustavo A. R. Silva wrote:
>
> On 8/25/20 10:51, Alex Dewar wrote:
> > Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> > erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
> > were later converted to fallthrough statements by commit df561f6688fe
> > ("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
> > warning about unreachable code.
> >
> > Remove the fallthrough statements and replace the mass of gotos with
> > simple return statements to make the code terser and less bug-prone.
> >
>
> This should be split up into two separate patches: one to address the
> fallthrough markings, and another one for the gotos.

I don't think it's necessary to break this into multiple patches.
Logical changes in a single patch are just fine, micro patches
aren't that useful.

2020-08-25 16:44:32

by Gustavo A. R. Silva

[permalink] [raw]

Subject: Re: [PATCH] IB/qib: remove superfluous fallthrough statements

On 8/25/20 11:26, Joe Perches wrote:
> On Tue, 2020-08-25 at 11:19 -0500, Gustavo A. R. Silva wrote:
>>
>> On 8/25/20 10:51, Alex Dewar wrote:
>>> Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
>>> erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
>>> were later converted to fallthrough statements by commit df561f6688fe
>>> ("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
>>> warning about unreachable code.
>>>
>>> Remove the fallthrough statements and replace the mass of gotos with
>>> simple return statements to make the code terser and less bug-prone.
>>>
>>
>> This should be split up into two separate patches: one to address the
>> fallthrough markings, and another one for the gotos.
>
> I don't think it's necessary to break this into multiple patches.
> Logical changes in a single patch are just fine, micro patches
> aren't that useful.
>

There is a reason for this. Read the changelog text and review the patch.

Thanks
--
Gustavo

2020-08-25 16:51:07

by Joe Perches

[permalink] [raw]

Subject: Re: [PATCH] IB/qib: remove superfluous fallthrough statements

On Tue, 2020-08-25 at 11:49 -0500, Gustavo A. R. Silva wrote:
>
> On 8/25/20 11:26, Joe Perches wrote:
> > On Tue, 2020-08-25 at 11:19 -0500, Gustavo A. R. Silva wrote:
> > > On 8/25/20 10:51, Alex Dewar wrote:
> > > > Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> > > > erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
> > > > were later converted to fallthrough statements by commit df561f6688fe
> > > > ("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
> > > > warning about unreachable code.
> > > >
> > > > Remove the fallthrough statements and replace the mass of gotos with
> > > > simple return statements to make the code terser and less bug-prone.
> > > >
> > >
> > > This should be split up into two separate patches: one to address the
> > > fallthrough markings, and another one for the gotos.
> >
> > I don't think it's necessary to break this into multiple patches.
> > Logical changes in a single patch are just fine, micro patches
> > aren't that useful.
> >
>
> There is a reason for this. Read the changelog text and review the patch.

What makes you think I didn't already do that?

I think your desire for micropatches is unnecessary.

2020-08-25 16:57:06

by Gustavo A. R. Silva

[permalink] [raw]

Subject: Re: [PATCH] IB/qib: remove superfluous fallthrough statements

On 8/25/20 11:47, Joe Perches wrote:
> On Tue, 2020-08-25 at 11:49 -0500, Gustavo A. R. Silva wrote:
>>
>> On 8/25/20 11:26, Joe Perches wrote:
>>> On Tue, 2020-08-25 at 11:19 -0500, Gustavo A. R. Silva wrote:
>>>> On 8/25/20 10:51, Alex Dewar wrote:
>>>>> Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
>>>>> erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
>>>>> were later converted to fallthrough statements by commit df561f6688fe
>>>>> ("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
>>>>> warning about unreachable code.
>>>>>
>>>>> Remove the fallthrough statements and replace the mass of gotos with
>>>>> simple return statements to make the code terser and less bug-prone.
>>>>>
>>>>
>>>> This should be split up into two separate patches: one to address the
>>>> fallthrough markings, and another one for the gotos.
>>>
>>> I don't think it's necessary to break this into multiple patches.
>>> Logical changes in a single patch are just fine, micro patches
>>> aren't that useful.
>>>
>>
>> There is a reason for this. Read the changelog text and review the patch.
>
> What makes you think I didn't already do that?
>

You would have noticed this should be two patches.

> I think your desire for micropatches is unnecessary.
>

You might be generalizing. My 'desire' here is justified and specific.

Thanks
--
Gustavo

2020-08-25 17:01:30

by Alex Dewar

[permalink] [raw]

Subject: Re: [PATCH] IB/qib: remove superfluous fallthrough statements

On 25/08/2020 17:49, Gustavo A. R. Silva wrote:
>
> On 8/25/20 11:26, Joe Perches wrote:
>> On Tue, 2020-08-25 at 11:19 -0500, Gustavo A. R. Silva wrote:
>>> On 8/25/20 10:51, Alex Dewar wrote:
>>>> Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
>>>> erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
>>>> were later converted to fallthrough statements by commit df561f6688fe
>>>> ("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
>>>> warning about unreachable code.
>>>>
>>>> Remove the fallthrough statements and replace the mass of gotos with
>>>> simple return statements to make the code terser and less bug-prone.
>>>>
>>> This should be split up into two separate patches: one to address the
>>> fallthrough markings, and another one for the gotos.
>> I don't think it's necessary to break this into multiple patches.
>> Logical changes in a single patch are just fine, micro patches
>> aren't that useful.
>>
> There is a reason for this. Read the changelog text and review the patch.
>
> Thanks
> --
> Gustavo
I think it probably does make sense as two patches. I'll do a resend.

2020-08-25 17:15:19

by Alex Dewar

[permalink] [raw]

Subject: [PATCH v2 1/2] IB/qib: remove superfluous fallthrough statements

2020-08-25 17:16:09

by Alex Dewar

[permalink] [raw]

Subject: [PATCH v2 2/2] IB/qib: tidy up process_cc()

This function has a lot of gotos which could be replaced by simple
returns, making the function tidier and less bug prone.

Signed-off-by: Alex Dewar <[email protected]>
---
drivers/infiniband/hw/qib/qib_mad.c | 50 ++++++++---------------------
1 file changed, 13 insertions(+), 37 deletions(-)

diff --git a/drivers/infiniband/hw/qib/qib_mad.c b/drivers/infiniband/hw/qib/qib_mad.c
index f972e559a8a7..f83e331977f8 100644
--- a/drivers/infiniband/hw/qib/qib_mad.c
+++ b/drivers/infiniband/hw/qib/qib_mad.c
@@ -2293,74 +2293,50 @@ static int process_cc(struct ib_device *ibdev, int mad_flags,
struct ib_mad *out_mad)
{
struct ib_cc_mad *ccp = (struct ib_cc_mad *)out_mad;
- int ret;
-
*out_mad = *in_mad;

if (ccp->class_version != 2) {
ccp->status |= IB_SMP_UNSUP_VERSION;
- ret = reply((struct ib_smp *)ccp);
- goto bail;
+ return reply((struct ib_smp *)ccp);
}

switch (ccp->method) {
case IB_MGMT_METHOD_GET:
switch (ccp->attr_id) {
case IB_CC_ATTR_CLASSPORTINFO:
- ret = cc_get_classportinfo(ccp, ibdev);
- goto bail;
-
+ return cc_get_classportinfo(ccp, ibdev);
case IB_CC_ATTR_CONGESTION_INFO:
- ret = cc_get_congestion_info(ccp, ibdev, port);
- goto bail;
-
+ return cc_get_congestion_info(ccp, ibdev, port);
case IB_CC_ATTR_CA_CONGESTION_SETTING:
- ret = cc_get_congestion_setting(ccp, ibdev, port);
- goto bail;
-
+ return cc_get_congestion_setting(ccp, ibdev, port);
case IB_CC_ATTR_CONGESTION_CONTROL_TABLE:
- ret = cc_get_congestion_control_table(ccp, ibdev, port);
- goto bail;
-
+ return cc_get_congestion_control_table(ccp, ibdev, port);
default:
ccp->status |= IB_SMP_UNSUP_METH_ATTR;
- ret = reply((struct ib_smp *) ccp);
- goto bail;
+ return reply((struct ib_smp *) ccp);
}
-
case IB_MGMT_METHOD_SET:
switch (ccp->attr_id) {
case IB_CC_ATTR_CA_CONGESTION_SETTING:
- ret = cc_set_congestion_setting(ccp, ibdev, port);
- goto bail;
-
+ return cc_set_congestion_setting(ccp, ibdev, port);
case IB_CC_ATTR_CONGESTION_CONTROL_TABLE:
- ret = cc_set_congestion_control_table(ccp, ibdev, port);
- goto bail;
-
+ return cc_set_congestion_control_table(ccp, ibdev, port);
default:
ccp->status |= IB_SMP_UNSUP_METH_ATTR;
- ret = reply((struct ib_smp *) ccp);
- goto bail;
+ return reply((struct ib_smp *) ccp);
}
-
case IB_MGMT_METHOD_GET_RESP:
/*
* The ib_mad module will call us to process responses
* before checking for other consumers.
* Just tell the caller to process it normally.
*/
- ret = IB_MAD_RESULT_SUCCESS;
- goto bail;
-
- case IB_MGMT_METHOD_TRAP:
- default:
- ccp->status |= IB_SMP_UNSUP_METHOD;
- ret = reply((struct ib_smp *) ccp);
+ return IB_MAD_RESULT_SUCCESS;
}

-bail:
- return ret;
+ /* method is unsupported */
+ ccp->status |= IB_SMP_UNSUP_METHOD;
+ return reply((struct ib_smp *) ccp);
}

/**
--
2.28.0

2020-08-25 19:30:59

by Gustavo A. R. Silva

[permalink] [raw]

Subject: Re: [PATCH v2 1/2] IB/qib: remove superfluous fallthrough statements

On Tue, Aug 25, 2020 at 06:12:42PM +0100, Alex Dewar wrote:
> Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
> were later converted to fallthrough statements by commit df561f6688fe
> ("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
> warning about unreachable code.
>

It's worth mentioning that this warning is triggered only by compilers
that don't support __attribute__((__fallthrough__)), which has been
supported since GCC 7.1.

> Remove the fallthrough statements.
>
> Addresses-Coverity: ("Unreachable code")
> Fixes: 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> Signed-off-by: Alex Dewar <[email protected]>

Reviewed-by: Gustavo A. R. Silva <[email protected]>

> ---
> v2: Do refactoring in a separate patch (Gustavo)
> ---

^^^
These dashes are not needed.

Thanks
--
Gustavo

> drivers/infiniband/hw/qib/qib_mad.c | 2 --
> 1 file changed, 2 deletions(-)
>
> diff --git a/drivers/infiniband/hw/qib/qib_mad.c b/drivers/infiniband/hw/qib/qib_mad.c
> index e7789e724f56..f972e559a8a7 100644
> --- a/drivers/infiniband/hw/qib/qib_mad.c
> +++ b/drivers/infiniband/hw/qib/qib_mad.c
> @@ -2322,7 +2322,6 @@ static int process_cc(struct ib_device *ibdev, int mad_flags,
> ret = cc_get_congestion_control_table(ccp, ibdev, port);
> goto bail;
>
> - fallthrough;
> default:
> ccp->status |= IB_SMP_UNSUP_METH_ATTR;
> ret = reply((struct ib_smp *) ccp);
> @@ -2339,7 +2338,6 @@ static int process_cc(struct ib_device *ibdev, int mad_flags,
> ret = cc_set_congestion_control_table(ccp, ibdev, port);
> goto bail;
>
> - fallthrough;
> default:
> ccp->status |= IB_SMP_UNSUP_METH_ATTR;
> ret = reply((struct ib_smp *) ccp);
> --
> 2.28.0
>

2020-08-25 20:58:36

by Joe Perches

[permalink] [raw]

Subject: Re: [PATCH] IB/qib: remove superfluous fallthrough statements

On Tue, 2020-08-25 at 12:01 -0500, Gustavo A. R. Silva wrote:
> On 8/25/20 11:47, Joe Perches wrote
[]
> You would have noticed this should be two patches.

That's interpretational.

> > I think your desire for micropatches is unnecessary.
> >
> You might be generalizing. My 'desire' here is justified and specific.

And to date undescribed.

2020-08-26 19:14:19

by Gustavo A. R. Silva

[permalink] [raw]

Subject: Re: [PATCH v2 1/2] IB/qib: remove superfluous fallthrough statements

Hi,

On Tue, Aug 25, 2020 at 02:33:27PM -0500, Gustavo A. R. Silva wrote:
> On Tue, Aug 25, 2020 at 06:12:42PM +0100, Alex Dewar wrote:
> > Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> > erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
> > were later converted to fallthrough statements by commit df561f6688fe
> > ("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
> > warning about unreachable code.
> >
>
> It's worth mentioning that this warning is triggered only by compilers
> that don't support __attribute__((__fallthrough__)), which has been
> supported since GCC 7.1.
>
> > Remove the fallthrough statements.
> >
> > Addresses-Coverity: ("Unreachable code")
> > Fixes: 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> > Signed-off-by: Alex Dewar <[email protected]>
>
> Reviewed-by: Gustavo A. R. Silva <[email protected]>
>

I can take this in my tree for 5.9-rc3.

Thanks
--
Gustavo

2020-08-27 00:13:28

by Jason Gunthorpe

[permalink] [raw]

Subject: Re: [PATCH v2 1/2] IB/qib: remove superfluous fallthrough statements

On Wed, Aug 26, 2020 at 02:18:59PM -0500, Gustavo A. R. Silva wrote:
> Hi,
>
> On Tue, Aug 25, 2020 at 02:33:27PM -0500, Gustavo A. R. Silva wrote:
> > On Tue, Aug 25, 2020 at 06:12:42PM +0100, Alex Dewar wrote:
> > > Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> > > erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
> > > were later converted to fallthrough statements by commit df561f6688fe
> > > ("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
> > > warning about unreachable code.
> > >
> >
> > It's worth mentioning that this warning is triggered only by compilers
> > that don't support __attribute__((__fallthrough__)), which has been
> > supported since GCC 7.1.
> >
> > > Remove the fallthrough statements.
> > >
> > > Addresses-Coverity: ("Unreachable code")
> > > Fixes: 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> > > Signed-off-by: Alex Dewar <[email protected]>
> >
> > Reviewed-by: Gustavo A. R. Silva <[email protected]>
> >
>
> I can take this in my tree for 5.9-rc3.

That would make conflicts for the 2nd patch, lets just send them all
through the rdma tree please. Is there a reason this is -rc material?

Jason

2020-08-27 01:38:51

by Gustavo A. R. Silva

[permalink] [raw]

Subject: Re: [PATCH v2 1/2] IB/qib: remove superfluous fallthrough statements

On Wed, Aug 26, 2020 at 09:11:49PM -0300, Jason Gunthorpe wrote:
> On Wed, Aug 26, 2020 at 02:18:59PM -0500, Gustavo A. R. Silva wrote:
> > Hi,
> >
> > On Tue, Aug 25, 2020 at 02:33:27PM -0500, Gustavo A. R. Silva wrote:
> > > On Tue, Aug 25, 2020 at 06:12:42PM +0100, Alex Dewar wrote:
> > > > Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> > > > erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
> > > > were later converted to fallthrough statements by commit df561f6688fe
> > > > ("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
> > > > warning about unreachable code.
> > > >
> > >
> > > It's worth mentioning that this warning is triggered only by compilers
> > > that don't support __attribute__((__fallthrough__)), which has been
> > > supported since GCC 7.1.
> > >
> > > > Remove the fallthrough statements.
> > > >
> > > > Addresses-Coverity: ("Unreachable code")
> > > > Fixes: 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> > > > Signed-off-by: Alex Dewar <[email protected]>
> > >
> > > Reviewed-by: Gustavo A. R. Silva <[email protected]>
> > >
> >
> > I can take this in my tree for 5.9-rc3.
>
> That would make conflicts for the 2nd patch, lets just send them all
> through the rdma tree please.

OK.

> Is there a reason this is -rc material?

It's just that this warning is currently in mainline.

Thanks
--
Gustavo

2020-08-27 10:41:45

by Alex Dewar

[permalink] [raw]

Subject: Re: [PATCH v2 1/2] IB/qib: remove superfluous fallthrough statements

On Wed, Aug 26, 2020 at 08:41:20PM -0500, Gustavo A. R. Silva wrote:
> On Wed, Aug 26, 2020 at 09:11:49PM -0300, Jason Gunthorpe wrote:
> > On Wed, Aug 26, 2020 at 02:18:59PM -0500, Gustavo A. R. Silva wrote:
> > > Hi,
> > >
> > > On Tue, Aug 25, 2020 at 02:33:27PM -0500, Gustavo A. R. Silva wrote:
> > > > On Tue, Aug 25, 2020 at 06:12:42PM +0100, Alex Dewar wrote:
> > > > > Commit 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> > > > > erroneously marked a couple of switch cases as /* FALLTHROUGH */, which
> > > > > were later converted to fallthrough statements by commit df561f6688fe
> > > > > ("treewide: Use fallthrough pseudo-keyword"). This triggered a Coverity
> > > > > warning about unreachable code.
> > > > >
> > > >
> > > > It's worth mentioning that this warning is triggered only by compilers
> > > > that don't support __attribute__((__fallthrough__)), which has been
> > > > supported since GCC 7.1.
> > > >
> > > > > Remove the fallthrough statements.
> > > > >
> > > > > Addresses-Coverity: ("Unreachable code")
> > > > > Fixes: 36a8f01cd24b ("IB/qib: Add congestion control agent implementation")
> > > > > Signed-off-by: Alex Dewar <[email protected]>
> > > >
> > > > Reviewed-by: Gustavo A. R. Silva <[email protected]>
> > > >
> > >
> > > I can take this in my tree for 5.9-rc3.
> >
> > That would make conflicts for the 2nd patch, lets just send them all
> > through the rdma tree please.
>
> OK.
>
> > Is there a reason this is -rc material?
>
> It's just that this warning is currently in mainline.

FYI this issue was found with Coverity, not a compiler. I just built the
unfixed version from mainline (with gcc 10.2.0) and didn't get any
warnings.

>
> Thanks
> --
> Gustavo

2020-08-31 20:25:41

by Jason Gunthorpe

[permalink] [raw]