From: Marijn Suijten <marijn.suijten@somainline.org>
To: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Cc: freedreno@lists.freedesktop.org, Sean Paul <sean@poorly.run>,
Bjorn Andersson <andersson@kernel.org>,
Abhinav Kumar <quic_abhinavk@quicinc.com>,
dri-devel@lists.freedesktop.org,
Stephen Boyd <swboyd@chromium.org>,
linux-arm-msm@vger.kernel.org
Subject: Re: [PATCH v2 06/22] drm/msm/dpu: simplify peer LM handling
Date: Fri, 16 Jun 2023 01:04:29 +0200 [thread overview]
Message-ID: <cjspdreamae7wvvxo45yiru2b6f57faxguf3er5vdmj62zesej@hwwik7mps7ru> (raw)
In-Reply-To: <20230613001004.3426676-7-dmitry.baryshkov@linaro.org>
On 2023-06-13 03:09:45, Dmitry Baryshkov wrote:
> For each LM there is at max 1 peer LM which can be driven by the same
> CTL, so there no need to have a mask instead of just an ID of the peer
> LM.
>
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Nit: I think you can describe the the patch contents in the title:
Replace LM peer mask with index
Instead of the vague (IMHO) "simplify handling".
> ---
> .../gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c | 2 +-
> .../gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h | 4 +--
> drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c | 34 +++++++------------
> 3 files changed, 15 insertions(+), 25 deletions(-)
>
> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c
> index 0de507d4d7b7..30fb5b1f3966 100644
> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c
> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c
> @@ -394,7 +394,7 @@ static const struct dpu_sspp_sub_blks qcm2290_dma_sblk_0 = _DMA_SBLK("8", 1);
> .features = _fmask, \
> .sblk = _sblk, \
> .pingpong = _pp, \
> - .lm_pair_mask = (1 << _lmpair), \
> + .lm_pair = _lmpair, \
> .dspp = _dspp \
> }
>
> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h
> index b860784ade72..b07caa4b867e 100644
> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h
> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.h
> @@ -554,14 +554,14 @@ struct dpu_sspp_cfg {
> * @features bit mask identifying sub-blocks/features
> * @sblk: LM Sub-blocks information
> * @pingpong: ID of connected PingPong, PINGPONG_NONE if unsupported
> - * @lm_pair_mask: Bitmask of LMs that can be controlled by same CTL
> + * @lm_pair: ID of LM that can be controlled by same CTL
Of *the* LM
By *the* same CTL
But then the rest of these comments have this borked hard-to-read style
as well.
> */
> struct dpu_lm_cfg {
> DPU_HW_BLK_INFO;
> const struct dpu_lm_sub_blks *sblk;
> u32 pingpong;
> u32 dspp;
> - unsigned long lm_pair_mask;
> + unsigned long lm_pair;
> };
>
> /**
> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c
> index 471842bbb950..e333f4eeafc1 100644
> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c
> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c
> @@ -253,28 +253,19 @@ static bool _dpu_rm_needs_split_display(const struct msm_display_topology *top)
> }
>
> /**
> - * _dpu_rm_check_lm_peer - check if a mixer is a peer of the primary
> + * _dpu_rm_get_lm_peer - get the id of a mixer which is a peer of the primary
... mixer?
> * @rm: dpu resource manager handle
> * @primary_idx: index of primary mixer in rm->mixer_blks[]
> - * @peer_idx: index of other mixer in rm->mixer_blks[]
> - * Return: true if rm->mixer_blks[peer_idx] is a peer of
> - * rm->mixer_blks[primary_idx]
> */
> -static bool _dpu_rm_check_lm_peer(struct dpu_rm *rm, int primary_idx,
> - int peer_idx)
> +static int _dpu_rm_get_lm_peer(struct dpu_rm *rm, int primary_idx)
> {
> const struct dpu_lm_cfg *prim_lm_cfg;
> - const struct dpu_lm_cfg *peer_cfg;
>
> prim_lm_cfg = to_dpu_hw_mixer(rm->mixer_blks[primary_idx])->cap;
> - peer_cfg = to_dpu_hw_mixer(rm->mixer_blks[peer_idx])->cap;
>
> - if (!test_bit(peer_cfg->id, &prim_lm_cfg->lm_pair_mask)) {
> - DPU_DEBUG("lm %d not peer of lm %d\n", peer_cfg->id,
> - peer_cfg->id);
> - return false;
> - }
> - return true;
> + if (prim_lm_cfg->lm_pair >= LM_0 && prim_lm_cfg->lm_pair < LM_MAX)
> + return prim_lm_cfg->lm_pair - LM_0;
> + return -EINVAL;
> }
>
> /**
> @@ -351,7 +342,7 @@ static int _dpu_rm_reserve_lms(struct dpu_rm *rm,
> int lm_idx[MAX_BLOCKS];
> int pp_idx[MAX_BLOCKS];
> int dspp_idx[MAX_BLOCKS] = {0};
> - int i, j, lm_count = 0;
> + int i, lm_count = 0;
>
> if (!reqs->topology.num_lm) {
> DPU_ERROR("invalid number of lm: %d\n", reqs->topology.num_lm);
> @@ -376,16 +367,15 @@ static int _dpu_rm_reserve_lms(struct dpu_rm *rm,
> ++lm_count;
>
> /* Valid primary mixer found, find matching peers */
> - for (j = i + 1; j < ARRAY_SIZE(rm->mixer_blks) &&
> - lm_count < reqs->topology.num_lm; j++) {
> - if (!rm->mixer_blks[j])
> + if (lm_count < reqs->topology.num_lm) {
> + int j = _dpu_rm_get_lm_peer(rm, i);
> +
> + /* ignore the peer if there is an error or if the peer was already processed */
I would not call this an "error" (though it is -EINVAL): 0 (out of range
of LM_0 <= x M LM_MAX) is a valid value meaning "LM has no peer" and
maybe another error code is more fitting?
> + if (j < 0 || j < i)
> continue;
>
> - if (!_dpu_rm_check_lm_peer(rm, i, j)) {
> - DPU_DEBUG("lm %d not peer of lm %d\n", LM_0 + j,
> - LM_0 + i);
> + if (!rm->mixer_blks[j])
> continue;
This seems silly now, why would an existing LM have a pair to an LM that
might not be in the catalog? Return -EINVAL?
Nits aside, this is some tight cleanup:
Reviewed-by: Marijn Suijten <marijn.suijten@somainline.org>
- Marijn
> - }
>
> if (!_dpu_rm_check_lm_and_get_connected_blks(rm,
> global_state, enc_id, j,
> --
> 2.39.2
>
next prev parent reply other threads:[~2023-06-15 23:04 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-06-13 0:09 [PATCH v2 00/22]drm/msm/dpu: another catalog rework Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 01/22] drm/msm/dpu: fix sc7280 and sc7180 PINGPONG done interrupts Dmitry Baryshkov
2023-06-13 21:41 ` Abhinav Kumar
2023-06-15 22:19 ` Marijn Suijten
2023-06-13 0:09 ` [PATCH v2 02/22] drm/msm/dpu: correct MERGE_3D length Dmitry Baryshkov
2023-06-13 21:54 ` Abhinav Kumar
2023-06-15 22:22 ` Marijn Suijten
2023-06-16 9:45 ` Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 03/22] drm/msm/dpu: remove unused INTF_NONE interfaces Dmitry Baryshkov
2023-06-13 22:06 ` Abhinav Kumar
2023-06-15 22:26 ` Marijn Suijten
2023-06-16 9:49 ` Dmitry Baryshkov
2023-06-16 12:28 ` Marijn Suijten
2023-06-13 0:09 ` [PATCH v2 04/22] drm/msm: enumerate DSI interfaces Dmitry Baryshkov
2023-06-13 22:07 ` Abhinav Kumar
2023-06-15 22:26 ` Marijn Suijten
2023-06-13 0:09 ` [PATCH v2 05/22] drm/msm/dpu: always use MSM_DP/DSI_CONTROLLER_n Dmitry Baryshkov
2023-06-13 23:47 ` Abhinav Kumar
2023-06-15 22:31 ` Marijn Suijten
2023-06-13 0:09 ` [PATCH v2 06/22] drm/msm/dpu: simplify peer LM handling Dmitry Baryshkov
2023-06-15 23:04 ` Marijn Suijten [this message]
2023-06-13 0:09 ` [PATCH v2 07/22] drm/msm/dpu: drop dpu_mdss_cfg::mdp_count field Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 08/22] drm/msm/dpu: drop enum dpu_mdp and MDP_TOP value Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 09/22] drm/msm/dpu: expand .clk_ctrls definitions Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 10/22] drm/msm/dpu: drop zero features from dpu_mdp_cfg data Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 11/22] drm/msm/dpu: drop zero features from dpu_ctl_cfg data Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 12/22] drm/msm/dpu: correct indentation for CTL definitions Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 13/22] drm/msm/dpu: inline SSPP_BLK macros Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 14/22] drm/msm/dpu: inline DSPP_BLK macros Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 15/22] drm/msm/dpu: inline LM_BLK macros Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 16/22] drm/msm/dpu: inline DSC_BLK and DSC_BLK_1_2 macros Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 16/21] drm/msm/dpu: inline DSC_BLK macros Dmitry Baryshkov
2023-06-15 22:05 ` Marijn Suijten
2023-06-15 22:06 ` Dmitry Baryshkov
2023-06-13 0:09 ` [PATCH v2 17/22] drm/msm/dpu: inline MERGE_3D_BLK macros Dmitry Baryshkov
2023-06-13 0:12 ` [PATCH v2 20/22] drm/msm/dpu: inline INTF_BLK and INTF_BLK_DSI_TE macros Dmitry Baryshkov
2023-06-13 0:12 ` [PATCH v2 21/22] drm/msm/dpu: drop empty features mask MERGE_3D_SM8150_MASK Dmitry Baryshkov
2023-06-13 0:12 ` [PATCH v2 22/22] drm/msm/dpu: drop empty features mask INTF_SDM845_MASK Dmitry Baryshkov
2023-06-13 0:14 ` [PATCH v2 18/22] drm/msm/dpu: inline various PP_BLK_* macros Dmitry Baryshkov
2023-06-13 0:14 ` [PATCH v2 19/22] drm/msm/dpu: inline WB_BLK macros Dmitry Baryshkov
2023-06-13 0:14 ` [PATCH v2 20/22] drm/msm/dpu: inline INTF_BLK and INTF_BLK_DSI_TE macros Dmitry Baryshkov
2023-06-13 0:14 ` [PATCH v2 21/22] drm/msm/dpu: drop empty features mask MERGE_3D_SM8150_MASK Dmitry Baryshkov
2023-06-13 0:14 ` [PATCH v2 22/22] drm/msm/dpu: drop empty features mask INTF_SDM845_MASK Dmitry Baryshkov
2023-06-15 11:31 ` [PATCH v2 00/22]drm/msm/dpu: another catalog rework Dmitry Baryshkov
2023-06-15 22:32 ` Marijn Suijten
2023-06-15 22:09 ` Marijn Suijten
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=cjspdreamae7wvvxo45yiru2b6f57faxguf3er5vdmj62zesej@hwwik7mps7ru \
--to=marijn.suijten@somainline.org \
--cc=andersson@kernel.org \
--cc=dmitry.baryshkov@linaro.org \
--cc=dri-devel@lists.freedesktop.org \
--cc=freedreno@lists.freedesktop.org \
--cc=linux-arm-msm@vger.kernel.org \
--cc=quic_abhinavk@quicinc.com \
--cc=sean@poorly.run \
--cc=swboyd@chromium.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).