Re: [PATCH 04/20] [SCSI] mpt3sas: Remove redundancy code while freeing the controller resources.

[Johannes Thumshirn <jthumshirn@suse.de>]

On Mon, Jun 15, 2015 at 03:56:56PM +0530, Sreekanth Reddy wrote:
> On Fri, Jun 12, 2015 at 6:10 PM, Johannes Thumshirn <jthumshirn@suse.de> wrote:
> > On Fri, Jun 12, 2015 at 05:48:56PM +0530, Sreekanth Reddy wrote:
> >> On Fri, Jun 12, 2015 at 4:58 PM, Johannes Thumshirn <jthumshirn@suse.de> wrote:
> >> > On Fri, Jun 12, 2015 at 03:12:16PM +0530, Sreekanth Reddy wrote:
> >> >> Removed the redundancy code while freeing the controller resources.
> >> >>
> >> >> Signed-off-by: Sreekanth Reddy <Sreekanth.Reddy@avagotech.com>
> >> >> ---
> >> >>  drivers/scsi/mpt3sas/mpt3sas_base.c | 57 +++++++++++++++++++++----------------
> > [...]
> >> >> +     _base_free_irq(ioc);
> >> >> +     _base_disable_msix(ioc);
> >> >> +
> >> >> +     if (ioc->msix96_vector)
> >> >> +             kfree(ioc->replyPostRegisterIndex);
> >> >
> >> > kfree() already checks for zero or a NULL pointer.
> >>
> >> Sorry Johannes, I didn't get you. If I understand this correctly, you
> >> are suggesting to check for NULL pointer before calling kree() API as
> >> shown below,
> >>
> >> if (ioc->msix96_vector && (ioc->replyPostRegisterIndex != NULL))
> >>         kfree(ioc->replyPostRegisterIndex);
> >
> > Correct me if I'm wrong, but I thought you don't need the if
> > (ioc->msix96_vector) before the kfree(). ioc->replyPostRegisterIndex should be
> > NULL if ioc->msix96_vector is 0, as far as I can see.
> >
> > In _scsih_probe() you have:
> > shost = scsi_host_alloc(&scsih_driver_template,
> >             sizeof(struct MPT3SAS_ADAPTER));
> > if (!shost)
> >            return -ENODEV;
> >
> > /* init local params */
> > ioc = shost_priv(shost);
> >
> > and scsi_host_alloc() does a kzalloc() for shost.
> >
> > so ioc->replyPortRegisterIndex is NULL.
> >
> > Or am I thinking wrong here?
> 
> Yes, ioc->replyPostRegisterIndex will be NULL if ioc->msix96_vector is 0,
> We have added this checks as a precautionary. since calling this
> function (i.e kfree()) on memory not previously allocated with
> kmalloc(), or on memory which has already been freed, may results in
> very bad things, such as freeing memory belonging to another part of
> the kernel.
>

OK, then please leave it in.

> >
> >>
> >> Regards,
> >> Sreekanth
> >
> > --
> > Johannes Thumshirn                                       Storage
> > jthumshirn@suse.de                             +49 911 74053 689
> > SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
> > GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
> > HRB 21284 (AG Nürnberg)
> 
> 
> 
> -- 
> 
> Regards,
> Sreekanth

-- 
Johannes Thumshirn                                       Storage
jthumshirn@suse.de                             +49 911 74053 689
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)