From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stanimir Varbanov Subject: Re: [PATCH 08/10] venus: vdec_ctrls: get real minimum buffers for capture Date: Thu, 24 Jan 2019 11:59:01 +0200 Message-ID: References: <20190117162008.25217-1-stanimir.varbanov@linaro.org> <20190117162008.25217-9-stanimir.varbanov@linaro.org> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Content-Language: en-US Sender: linux-kernel-owner@vger.kernel.org To: Alexandre Courbot Cc: Linux Media Mailing List , Mauro Carvalho Chehab , Hans Verkuil , LKML , linux-arm-msm@vger.kernel.org, Vikash Garodia , Tomasz Figa , Malathi Gottam List-Id: linux-arm-msm@vger.kernel.org Hi Alex, Thanks for the comments! On 1/24/19 10:43 AM, Alexandre Courbot wrote: > On Fri, Jan 18, 2019 at 1:21 AM Stanimir Varbanov > wrote: >> >> Until now we returned num_output_bufs set during reqbuf but >> that could be wrong when we implement stateful Codec API. So >> get the minimum buffers for capture from HFI. This is supposed >> to be called after stream header parsing, i.e. after dequeue >> v4l2 event for change resolution. >> >> Signed-off-by: Stanimir Varbanov >> --- >> drivers/media/platform/qcom/venus/vdec_ctrls.c | 7 ++++++- >> 1 file changed, 6 insertions(+), 1 deletion(-) >> >> diff --git a/drivers/media/platform/qcom/venus/vdec_ctrls.c b/drivers/media/platform/qcom/venus/vdec_ctrls.c >> index f4604b0cd57e..e1da87bf52bc 100644 >> --- a/drivers/media/platform/qcom/venus/vdec_ctrls.c >> +++ b/drivers/media/platform/qcom/venus/vdec_ctrls.c >> @@ -16,6 +16,7 @@ >> #include >> >> #include "core.h" >> +#include "helpers.h" >> #include "vdec.h" >> >> static int vdec_op_s_ctrl(struct v4l2_ctrl *ctrl) >> @@ -47,7 +48,9 @@ static int vdec_op_g_volatile_ctrl(struct v4l2_ctrl *ctrl) >> { >> struct venus_inst *inst = ctrl_to_inst(ctrl); >> struct vdec_controls *ctr = &inst->controls.dec; >> + struct hfi_buffer_requirements bufreq; >> union hfi_get_property hprop; >> + enum hfi_version ver = inst->core->res->hfi_version; >> u32 ptype = HFI_PROPERTY_PARAM_PROFILE_LEVEL_CURRENT; >> int ret; >> >> @@ -71,7 +74,9 @@ static int vdec_op_g_volatile_ctrl(struct v4l2_ctrl *ctrl) >> ctrl->val = ctr->post_loop_deb_mode; >> break; >> case V4L2_CID_MIN_BUFFERS_FOR_CAPTURE: >> - ctrl->val = inst->num_output_bufs; >> + ret = venus_helper_get_bufreq(inst, HFI_BUFFER_OUTPUT, &bufreq); >> + if (!ret) >> + ctrl->val = HFI_BUFREQ_COUNT_MIN(&bufreq, ver); > > What happens if venus_helper_get_bufreq() returns an error? It seems > that we just happily continue with whatever the previous value of > ctrl->val was. It seems like we do the same with other controls as > well. I agree that this is wrong, I think no-one userspace client used that g_ctrls controls :) > > I think you can fix this globally by initializing ret to 0 at the > beginning of the function, and then returning ret instead of 0 at the > end. That way all errors would be propagated. Of course please check > that this is relevant for other controls following this scheme before > doing so. :) > yes, I will do! -- regards, Stan