linux-kernel - Re: [PATCH v3 2/3] media: hantro: Support color conversion via post-processing

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <272dedf2cc25053475e9536b446b434f4754f62c.camel@pengutronix.de>
Date:   Fri, 15 Nov 2019 17:10:59 +0100
From:   Philipp Zabel <p.zabel@...gutronix.de>
To:     Ezequiel Garcia <ezequiel@...labora.com>,
        linux-media@...r.kernel.org
Cc:     kernel@...labora.com, Tomasz Figa <tfiga@...omium.org>,
        linux-rockchip@...ts.infradead.org,
        Heiko Stuebner <heiko@...ech.de>,
        Jonas Karlman <jonas@...boo.se>,
        Boris Brezillon <boris.brezillon@...labora.com>,
        Chris Healy <cphealy@...il.com>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 2/3] media: hantro: Support color conversion via
 post-processing

Hi Ezequiel,

On Fri, 2019-11-15 at 12:44 -0300, Ezequiel Garcia wrote:
> Hello Philipp,
> 
> Thanks for reviewing.
> 
> On Thu, 2019-11-14 at 10:48 +0100, Philipp Zabel wrote:
[...]
> > Why isn't PP enabled in prepare_run? Does this mean the first frame is
> > not post-processed?
> > 
> 
> No, because hantro_finish_run is called (despite its name)
> before the decoding operation is actually triggered.
> 
> I guess this hantro_finish_run name adds some confusion:
> prepare_run and finish_run should be something along
> start_prepare_run, end_prepare_run. 

Ah, ok then. I was confused because I just came from looking at coda-vpu 
code, where finish_run is a callback called after the device has
finished processing. Maybe I should rename that as well.

> And also, perhaps disabling the post-processor in prepare_run
> works just fine. I need to check that.

Ok.

[...]
> > > +#define HANTRO_PP_REG_WRITE_S(vpu, reg_name, val) \
> > > +	do { \
> > > +		if ((vpu)->variant->postproc_regs->(reg_name).base)	\
> > > +			hantro_reg_write((vpu), \
> > > +					 &(vpu)->variant->postproc_regs->(reg_name), \
> > > +					 (val)); \
> > > +	} while (0)
> > 
> > Why all these checks, are any of the register fields optional?
> > 
> 
> That was the plan. Perhaps now it makes less sense,
> but maybe it's safer this way, if it's extended?
> 
> OTOH, we might want to make sure the driver fails (or warns).

I think that would be better than silently ignoring them.

Although I don't quite see the point in repeatedly checking the presence
of mandatory register fields at runtime.

regards
Philipp