[dpdk-stable] [PATCH v10 3/4] pipeline: Fix compilation error with gcc ASan

[Dumitrescu, Cristian]

> -----Original Message-----
> From: Peng, ZhihongX <zhihongx.peng at intel.com>
> Sent: Monday, October 18, 2021 1:55 PM
> To: Dumitrescu, Cristian <cristian.dumitrescu at intel.com>;
> david.marchand at redhat.com; Burakov, Anatoly
> <anatoly.burakov at intel.com>; Ananyev, Konstantin
> <konstantin.ananyev at intel.com>; stephen at networkplumber.org;
> Mcnamara, John <john.mcnamara at intel.com>
> Cc: dev at dpdk.org; Lin, Xueqin <xueqin.lin at intel.com>; stable at dpdk.org
> Subject: RE: [PATCH v10 3/4] pipeline: Fix compilation error with gcc ASan
> 
> > -----Original Message-----
> > From: Dumitrescu, Cristian <cristian.dumitrescu at intel.com>
> > Sent: Monday, October 18, 2021 8:22 PM
> > To: Peng, ZhihongX <zhihongx.peng at intel.com>;
> > david.marchand at redhat.com; Burakov, Anatoly
> > <anatoly.burakov at intel.com>; Ananyev, Konstantin
> > <konstantin.ananyev at intel.com>; stephen at networkplumber.org;
> > Mcnamara, John <john.mcnamara at intel.com>
> > Cc: dev at dpdk.org; Lin, Xueqin <xueqin.lin at intel.com>; stable at dpdk.org
> > Subject: RE: [PATCH v10 3/4] pipeline: Fix compilation error with gcc ASan
> >
> >
> >
> > > -----Original Message-----
> > > From: Peng, ZhihongX <zhihongx.peng at intel.com>
> > > Sent: Friday, October 15, 2021 4:11 PM
> > > To: david.marchand at redhat.com; Burakov, Anatoly
> > > <anatoly.burakov at intel.com>; Ananyev, Konstantin
> > > <konstantin.ananyev at intel.com>; stephen at networkplumber.org;
> > > Dumitrescu, Cristian <cristian.dumitrescu at intel.com>; Mcnamara, John
> > > <john.mcnamara at intel.com>
> > > Cc: dev at dpdk.org; Lin, Xueqin <xueqin.lin at intel.com>; Peng, ZhihongX
> > > <zhihongx.peng at intel.com>; stable at dpdk.org
> > > Subject: [PATCH v10 3/4] pipeline: Fix compilation error with gcc ASan
> > >
> > > From: Zhihong Peng <zhihongx.peng at intel.com>
> > >
> > > The gcc will check code more stricter when ASan enabled.
> > > "Control reaches end of non-void function" error occurs here.
> > >
> > > Fixes: f38913b7fb8e ("pipeline: add meter array to SWX")
> > > Cc: stable at dpdk.org
> > >
> > > Signed-off-by: Xueqin Lin <xueqin.lin at intel.com>
> > > Signed-off-by: Zhihong Peng <zhihongx.peng at intel.com>
> > > ---
> > > v7: no change
> > > v8: no change
> > > v9: Modify the submit log
> > > v10:no change
> > > ---
> > >  lib/pipeline/rte_swx_pipeline.c | 4 ++--
> > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/lib/pipeline/rte_swx_pipeline.c
> > > b/lib/pipeline/rte_swx_pipeline.c index 1cd09a4b44..0acd6c6752 100644
> > > --- a/lib/pipeline/rte_swx_pipeline.c
> > > +++ b/lib/pipeline/rte_swx_pipeline.c
> > > @@ -4642,7 +4642,7 @@ instr_meter_translate(struct rte_swx_pipeline
> *p,
> > >  		return 0;
> > >  	}
> > >
> > > -	CHECK(0, EINVAL);
> > > +	return -EINVAL;
> > >  }
> > >
> > >  static inline void
> > > @@ -5937,7 +5937,7 @@ instr_translate(struct rte_swx_pipeline *p,
> > >  					      instr,
> > >  					      data);
> > >
> > > -	CHECK(0, EINVAL);
> > > +	return -EINVAL;
> > >  }
> > >
> > >  static struct instruction_data *
> > > --
> > > 2.25.1
> >
> > NACK.
> >
> > This is a false issue, no bug is here. CHECK(0, EINVAL) translates to an
> > unconditional return -EINVAL.
> > Does this tool work correctly when macros are present? Maybe the tool
> > should parse the preprocessed C code as opposed to initial C code?
> 
> Yes, this is not a bug, it just solves the problem that cannot be passed after
> adding the asan compiler option.
> Only part of the macro reports errors, which may be caused by the tool itself,
> but this tool is part of gcc and clang, so we still have to make the code not
> report errors.
> > Regards,
> > Cristian

Hi Zhihong,

If this is not a bug in the pipeline library, why then does your patch has fix in the tile, has the Fixes label and CC-es stable at dpdk.org? Please remove and rephrase accordingly.

I agree this is not a bug, and based on your statements I understand this is a sort of issue or limitation with the tool . I would prefer we fix the tool rather than fixing correct code in order to please the tool. This is likely not going to be an isolated case, but a recurring issue.

Hence, I am reluctantly OK to ack this patch in order to allow the tool in (after the "fix" claim is removed), hopefully the tool will prove its benefit to DPDK.

Regards,
Cristian