Re: [PATCH net-next v3] net: mana: Allow variable size indirection table

[Leon Romanovsky <leon@kernel.org>]

On Mon, Jun 03, 2024 at 10:36:48PM -0700, Shradha Gupta wrote:
> On Mon, Jun 03, 2024 at 11:41:22AM +0300, Leon Romanovsky wrote:
> > On Fri, May 31, 2024 at 08:37:41AM -0700, Shradha Gupta wrote:
> > > Allow variable size indirection table allocation in MANA instead
> > > of using a constant value MANA_INDIRECT_TABLE_SIZE.
> > > The size is now derived from the MANA_QUERY_VPORT_CONFIG and the
> > > indirection table is allocated dynamically.
> > > 
> > > Signed-off-by: Shradha Gupta <shradhagupta@linux.microsoft.com>
> > > Reviewed-by: Dexuan Cui <decui@microsoft.com>
> > > Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com>
> > > ---
> > >  Changes in v3:
> > >  * Fixed the memory leak(save_table) in mana_set_rxfh()
> > > 
> > >  Changes in v2:
> > >  * Rebased to latest net-next tree
> > >  * Rearranged cleanup code in mana_probe_port to avoid extra operations
> > > ---
> > >  drivers/infiniband/hw/mana/qp.c               | 10 +--
> > >  drivers/net/ethernet/microsoft/mana/mana_en.c | 68 ++++++++++++++++---
> > >  .../ethernet/microsoft/mana/mana_ethtool.c    | 27 +++++---
> > >  include/net/mana/gdma.h                       |  4 +-
> > >  include/net/mana/mana.h                       |  9 +--
> > >  5 files changed, 89 insertions(+), 29 deletions(-)
> > 
> > <...>
> > 
> > > +free_indir:
> > > +	apc->indir_table_sz = 0;
> > > +	kfree(apc->indir_table);
> > > +	apc->indir_table = NULL;
> > > +	kfree(apc->rxobj_table);
> > > +	apc->rxobj_table = NULL;
> > >  reset_apc:
> > >  	kfree(apc->rxqs);
> > >  	apc->rxqs = NULL;
> > > @@ -2897,6 +2936,7 @@ void mana_remove(struct gdma_dev *gd, bool suspending)
> > >  {
> > 
> > <...>
> > 
> > > @@ -2931,6 +2972,11 @@ void mana_remove(struct gdma_dev *gd, bool suspending)
> > >  		}
> > >  
> > >  		unregister_netdevice(ndev);
> > > +		apc->indir_table_sz = 0;
> > > +		kfree(apc->indir_table);
> > > +		apc->indir_table = NULL;
> > > +		kfree(apc->rxobj_table);
> > > +		apc->rxobj_table = NULL;
> > 
> > Why do you need to NULLify here? Will apc is going to be accessible
> > after call to mana_remove? or port probe failure?
> Right, they won't be accessed. This is just for the sake of completeness
> and to prevent double free in case there are code bug in other place.

This coding patter is called defensive programming, which is discouraged
in the kernel. You are not preventing double free, but hiding bugs which
were possible to be found by various static analysis tools.

Please don't do it.

Thanks

> 
> Regards,
> Shradha.
> > 
> > Thanks