linux-kernel - Re: [PATCH v2] media: hantro: HEVC: unconditionnaly set pps_{cb/cr}_qp

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] [day] [month] [year] [list]

Date:   Tue, 3 May 2022 15:32:11 +0200
From:   Benjamin Gaignard <benjamin.gaignard@...labora.com>
To:     Ezequiel Garcia <ezequiel@...guardiasur.com.ar>
Cc:     Philipp Zabel <p.zabel@...gutronix.de>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Greg KH <gregkh@...uxfoundation.org>,
        linux-media <linux-media@...r.kernel.org>,
        "open list:ARM/Rockchip SoC..." <linux-rockchip@...ts.infradead.org>,
        "open list:STAGING SUBSYSTEM" <linux-staging@...ts.linux.dev>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        jon@...ocrew.net, Adam Ford <aford173@...il.com>,
        Collabora Kernel ML <kernel@...labora.com>
Subject: Re: [PATCH v2] media: hantro: HEVC: unconditionnaly set
 pps_{cb/cr}_qp_offset values


Le 03/05/2022 à 15:12, Ezequiel Garcia a écrit :
> Hi Benjamin,
>
> On Tue, Apr 26, 2022 at 10:50 AM Benjamin Gaignard
> <benjamin.gaignard@...labora.com> wrote:
>> Always set pps_cb_qp_offset and pps_cr_qp_offset values in Hantro/G2
>> register whatever is V4L2_HEVC_PPS_FLAG_PPS_SLICE_CHROMA_QP_OFFSETS_PRESENT
>> flag value.
> I would say we need more justification why this is correct, or at least
> checking what the reference vendor implementation is doing (and mentioning
> in the commit description so we can track it in the future).

Yes that is what the vendor implementation is doing.

>
>> This fix CAINIT_G_SHARP_3 test in fluster.
>>
> This could sound like a tad a pedantic detail, but I'd say it's
> important we stop refering to tests
> as "fluster tests", and instead say something more correct as "HEVC
> conformance test CAINIT_G_SHARP_3".

As you want.

>
> Also, when we are fixing conformance tests, let's please add the
> Fluster score (in this case, I think it's
> OK to refer to Fluster).

We are fixing bugs in parallel in the driver, the uAPI and GStreamer
so fluster score evolution reflect that progression and maybe not only
what this patch is fixing.
The best I could says here is that patch fix HEVC conformance test
CAINIT_G_SHARP_3 so fluster score increase by one.

Regards,
Benjamin

>
> PS: Same comments apply to patch "media: hantro: HEVC: Fix reference
> frames management".
>
> Thanks,
> Ezequiel
>
>> Signed-off-by: Benjamin Gaignard <benjamin.gaignard@...labora.com>
>> ---
>>   drivers/staging/media/hantro/hantro_g2_hevc_dec.c | 9 ++-------
>>   1 file changed, 2 insertions(+), 7 deletions(-)
>>
>> diff --git a/drivers/staging/media/hantro/hantro_g2_hevc_dec.c b/drivers/staging/media/hantro/hantro_g2_hevc_dec.c
>> index 6deb31b7b993..503f4b028bc5 100644
>> --- a/drivers/staging/media/hantro/hantro_g2_hevc_dec.c
>> +++ b/drivers/staging/media/hantro/hantro_g2_hevc_dec.c
>> @@ -194,13 +194,8 @@ static void set_params(struct hantro_ctx *ctx)
>>                  hantro_reg_write(vpu, &g2_max_cu_qpd_depth, 0);
>>          }
>>
>> -       if (pps->flags & V4L2_HEVC_PPS_FLAG_PPS_SLICE_CHROMA_QP_OFFSETS_PRESENT) {
>> -               hantro_reg_write(vpu, &g2_cb_qp_offset, pps->pps_cb_qp_offset);
>> -               hantro_reg_write(vpu, &g2_cr_qp_offset, pps->pps_cr_qp_offset);
>> -       } else {
>> -               hantro_reg_write(vpu, &g2_cb_qp_offset, 0);
>> -               hantro_reg_write(vpu, &g2_cr_qp_offset, 0);
>> -       }
>> +       hantro_reg_write(vpu, &g2_cb_qp_offset, pps->pps_cb_qp_offset);
>> +       hantro_reg_write(vpu, &g2_cr_qp_offset, pps->pps_cr_qp_offset);
>>
>>          hantro_reg_write(vpu, &g2_filt_offset_beta, pps->pps_beta_offset_div2);
>>          hantro_reg_write(vpu, &g2_filt_offset_tc, pps->pps_tc_offset_div2);
>> --
>> 2.32.0
>>