Re: [PATCH v4 2nd 1/4] ramoops: Add __ramoops_init_prz() as generic function

[Mark Salyzyn]

On 02/07/2017 12:51 AM, 岩松信洋 / IWAMATSU，NOBUHIRO wrote:
> Hi,
>
> Thanks for your review.
>
> > -----Original Message-----
> > From: Mark Salyzyn [mailto:salyzyn@xxxxxxxxxxx]
> > Sent: Friday, February 03, 2017 7:13 AM
> > To: 岩松信洋 / IWAMATSU，NOBUHIRO; Anton Vorontsov; Colin Cross; Kees Cook;
> > Tony Luck
> > Cc: linux-kernel@xxxxxxxxxxxxxxx; Hiraku Toyooka; 阿口誠司 / AGUCHI，
> > SEIJI; Shuah Khan
> > Subject: Re: [PATCH v4 2nd 1/4] ramoops: Add __ramoops_init_prz() as generic
> > function
> >
> > On 01/30/2017 05:58 PM, Nobuhiro Iwamatsu wrote:
> > > +
> > > +	if (zap && *paddr + sz - cxt->phys_addr > cxt->size) {
> > > +		dev_err(dev, "no room for %s mem region (0x%zx@0x%llx) in
> > (0x%lx@0x%llx)\n",
> > > +			name, sz, (unsigned long long)*paddr,
> > > +			cxt->size, (unsigned long long)cxt->phys_addr);
> > > +		return -ENOMEM;
> > > +	}
> > > +
> > Why not allow this limit check for ramoops_init_prsz call?
> This code is controlled by flag of zap.
> If zap is false, __ramoops_init_prz() is worked same as original ramoops_init_przs().
>
> Best regards,
>    Nobuhiro

I agree, preserve functionality as is. Please consider removing the need  
for zap && in a future patch, as I _believe_ (untested, and no deep  
dive) that the limit check is relevant for the zap is false path as well  
and may have been an oversight.

-- Mark