From: "Niklas Söderlund" <niklas.soderlund+renesas@ragnatech.se>
To: Mauro Carvalho Chehab <mchehab@kernel.org>,
Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>,
Jacopo Mondi <jacopo.mondi@ideasonboard.com>,
Laurent Pinchart <laurent.pinchart@ideasonboard.com>,
linux-media@vger.kernel.org, linux-renesas-soc@vger.kernel.org,
linux-kernel@vger.kernel.org
Cc: "Niklas Söderlund" <niklas.soderlund+renesas@ragnatech.se>,
"Marek Vasut" <marek.vasut+renesas@mailbox.org>
Subject: [PATCH v6 09/12] media: rppx1: Add support for Lens Shade Correction
Date: Sat, 14 Mar 2026 22:59:41 +0100 [thread overview]
Message-ID: <20260314215944.3674865-10-niklas.soderlund+renesas@ragnatech.se> (raw)
In-Reply-To: <20260314215944.3674865-1-niklas.soderlund+renesas@ragnatech.se>
Extend the RPPX1 driver to allow setting the Lens Shade Correction (LSC)
configuration using the RkISP1 parameter buffer format. It uses the
RPPX1 framework for parameters and its writer abstraction to allow the
user to control how (and when) configuration is applied to the RPPX1.
As the RkISP1 parameters buffer only describes one quadrant of the lens,
this is due to the RkISP1 hardware only allowing one quadrant to be
programmed to hardware. The hardware then extrapolates this to the other
three quadrants of the lens. For RPP all four quadrants are individually
programmable.
To compensate for the driver need to extrapolate all four quadrants from
the RkISP1 parameters buffer information.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Tested-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
---
.../platform/dreamchip/rppx1/rpp_params.c | 4 +
.../platform/dreamchip/rppx1/rppx1_lsc.c | 126 ++++++++++++++++++
2 files changed, 130 insertions(+)
diff --git a/drivers/media/platform/dreamchip/rppx1/rpp_params.c b/drivers/media/platform/dreamchip/rppx1/rpp_params.c
index c807ccfef628..0314cb732844 100644
--- a/drivers/media/platform/dreamchip/rppx1/rpp_params.c
+++ b/drivers/media/platform/dreamchip/rppx1/rpp_params.c
@@ -19,6 +19,7 @@ rkisp1_ext_params_blocks_info[] = {
RKISP1_PARAMS_BLOCK_INFO(BLS, bls),
RKISP1_PARAMS_BLOCK_INFO(AWB_GAIN, awb_gain),
RKISP1_PARAMS_BLOCK_INFO(CTK, ctk),
+ RKISP1_PARAMS_BLOCK_INFO(LSC, lsc),
RKISP1_PARAMS_BLOCK_INFO(AWB_MEAS, awb_meas),
RKISP1_PARAMS_BLOCK_INFO(HST_MEAS, hst),
RKISP1_PARAMS_BLOCK_INFO(AEC_MEAS, aec),
@@ -65,6 +66,9 @@ int rppx1_params(struct rppx1 *rpp, struct vb2_buffer *vb, size_t max_size,
case RKISP1_EXT_PARAMS_BLOCK_TYPE_CTK:
module = &rpp->post.ccor;
break;
+ case RKISP1_EXT_PARAMS_BLOCK_TYPE_LSC:
+ module = &rpp->pre1.lsc;
+ break;
case RKISP1_EXT_PARAMS_BLOCK_TYPE_AWB_MEAS:
module = &rpp->post.wbmeas;
break;
diff --git a/drivers/media/platform/dreamchip/rppx1/rppx1_lsc.c b/drivers/media/platform/dreamchip/rppx1/rppx1_lsc.c
index e8acdf744956..00bdcc4aedb3 100644
--- a/drivers/media/platform/dreamchip/rppx1/rppx1_lsc.c
+++ b/drivers/media/platform/dreamchip/rppx1/rppx1_lsc.c
@@ -54,6 +54,10 @@
#define LSC_TABLE_SEL_REG 0x00a8
#define LSC_STATUS_REG 0x00ac
+#define LSC_R_TABLE_DATA_VALUE(v1, v2) (((v1) & 0xfff) | (((v2) & 0xfff) << 12))
+#define LSC_GRAD_VALUE(v1, v2) (((v1) & 0xfff) | (((v2) & 0xfff) << 16))
+#define LSC_SIZE_VALUE(v1, v2) (((v1) & 0x1ff) | (((v2) & 0x1ff) << 16))
+
static int rppx1_lsc_probe(struct rpp_module *mod)
{
/* Version check. */
@@ -63,6 +67,128 @@ static int rppx1_lsc_probe(struct rpp_module *mod)
return 0;
}
+static int
+rppx1_lsc_param_rkisp1(struct rpp_module *mod,
+ const union rppx1_params_rkisp1_config *block,
+ rppx1_reg_write write, void *priv)
+{
+ const struct rkisp1_ext_params_lsc_config *cfg = &block->lsc;
+ const __u16 *v;
+
+ /* Always disable module as it needs be disabled before configuring. */
+ write(priv, mod->base + LSC_CTRL_REG, 0);
+ if (cfg->header.flags & RKISP1_EXT_PARAMS_FL_BLOCK_DISABLE)
+ return 0;
+
+ /*
+ * Program the color correction sectors.
+ *
+ * There are two tables to one can program and switch between. As the
+ * RPPX1 supports preparing a buffer of commands to be applied later
+ * only use table 0. This works as long as the ISP is not used in
+ * inline-mode.
+ *
+ * For inline-mode support using DMA for configuration is not possible
+ * so this is not an issue, but needs to be address if inline-mode
+ * support is added to the driver.
+ */
+
+ /* Start writing at beginning of table 0. */
+ write(priv, mod->base + LSC_R_TABLE_ADDR_REG, 0);
+ write(priv, mod->base + LSC_GR_TABLE_ADDR_REG, 0);
+ write(priv, mod->base + LSC_B_TABLE_ADDR_REG, 0);
+ write(priv, mod->base + LSC_GB_TABLE_ADDR_REG, 0);
+
+ /* Program data tables. */
+ for (unsigned int i = 0; i < RKISP1_CIF_ISP_LSC_SAMPLES_MAX; i++) {
+ const __u16 *r = cfg->config.r_data_tbl[i];
+ const __u16 *gr = cfg->config.gr_data_tbl[i];
+ const __u16 *b = cfg->config.b_data_tbl[i];
+ const __u16 *gb = cfg->config.gb_data_tbl[i];
+ unsigned int j;
+
+ for (j = 0; j < RKISP1_CIF_ISP_LSC_SAMPLES_MAX - 1; j += 2) {
+ write(priv, mod->base + LSC_R_TABLE_DATA_REG,
+ LSC_R_TABLE_DATA_VALUE(r[j], r[j + 1]));
+ write(priv, mod->base + LSC_GR_TABLE_DATA_REG,
+ LSC_R_TABLE_DATA_VALUE(gr[j], gr[j + 1]));
+ write(priv, mod->base + LSC_B_TABLE_DATA_REG,
+ LSC_R_TABLE_DATA_VALUE(b[j], b[j + 1]));
+ write(priv, mod->base + LSC_GB_TABLE_DATA_REG,
+ LSC_R_TABLE_DATA_VALUE(gb[j], gb[j + 1]));
+ }
+
+ write(priv, mod->base + LSC_R_TABLE_DATA_REG,
+ LSC_R_TABLE_DATA_VALUE(r[j], 0));
+ write(priv, mod->base + LSC_GR_TABLE_DATA_REG,
+ LSC_R_TABLE_DATA_VALUE(gr[j], 0));
+ write(priv, mod->base + LSC_B_TABLE_DATA_REG,
+ LSC_R_TABLE_DATA_VALUE(b[j], 0));
+ write(priv, mod->base + LSC_GB_TABLE_DATA_REG,
+ LSC_R_TABLE_DATA_VALUE(gb[j], 0));
+ }
+
+ /* Activate table 0. */
+ write(priv, mod->base + LSC_TABLE_SEL_REG, 0);
+
+ /*
+ * Program X- and Y- sizes, and gradients.
+ *
+ * The RPP ISP can describe each quarter of the lens individually, this
+ * differs from the Rk1ISP which can only describe one quarter of lens
+ * with software and then extrapolates the other three.
+ *
+ * To adjust for this extrapolate the three missing quadrants using
+ * software for the RPP ISP.
+ */
+
+ v = cfg->config.x_grad_tbl;
+ write(priv, mod->base + LSC_XGRAD_01_REG, LSC_GRAD_VALUE(v[0], v[1]));
+ write(priv, mod->base + LSC_XGRAD_23_REG, LSC_GRAD_VALUE(v[2], v[3]));
+ write(priv, mod->base + LSC_XGRAD_45_REG, LSC_GRAD_VALUE(v[4], v[5]));
+ write(priv, mod->base + LSC_XGRAD_67_REG, LSC_GRAD_VALUE(v[6], v[7]));
+ write(priv, mod->base + LSC_XGRAD_89_REG, LSC_GRAD_VALUE(v[7], v[6]));
+ write(priv, mod->base + LSC_XGRAD_1011_REG, LSC_GRAD_VALUE(v[5], v[4]));
+ write(priv, mod->base + LSC_XGRAD_1213_REG, LSC_GRAD_VALUE(v[3], v[2]));
+ write(priv, mod->base + LSC_XGRAD_1415_REG, LSC_GRAD_VALUE(v[1], v[0]));
+
+ v = cfg->config.y_grad_tbl;
+ write(priv, mod->base + LSC_YGRAD_01_REG, LSC_GRAD_VALUE(v[0], v[1]));
+ write(priv, mod->base + LSC_YGRAD_23_REG, LSC_GRAD_VALUE(v[2], v[3]));
+ write(priv, mod->base + LSC_YGRAD_45_REG, LSC_GRAD_VALUE(v[4], v[5]));
+ write(priv, mod->base + LSC_YGRAD_67_REG, LSC_GRAD_VALUE(v[6], v[7]));
+ write(priv, mod->base + LSC_YGRAD_89_REG, LSC_GRAD_VALUE(v[7], v[6]));
+ write(priv, mod->base + LSC_YGRAD_1011_REG, LSC_GRAD_VALUE(v[5], v[4]));
+ write(priv, mod->base + LSC_YGRAD_1213_REG, LSC_GRAD_VALUE(v[3], v[2]));
+ write(priv, mod->base + LSC_YGRAD_1415_REG, LSC_GRAD_VALUE(v[1], v[0]));
+
+ v = cfg->config.x_size_tbl;
+ write(priv, mod->base + LSC_XSIZE_01_REG, LSC_GRAD_VALUE(v[0], v[1]));
+ write(priv, mod->base + LSC_XSIZE_23_REG, LSC_GRAD_VALUE(v[2], v[3]));
+ write(priv, mod->base + LSC_XSIZE_45_REG, LSC_GRAD_VALUE(v[4], v[5]));
+ write(priv, mod->base + LSC_XSIZE_67_REG, LSC_GRAD_VALUE(v[6], v[7]));
+ write(priv, mod->base + LSC_XSIZE_89_REG, LSC_GRAD_VALUE(v[7], v[6]));
+ write(priv, mod->base + LSC_XSIZE_1011_REG, LSC_GRAD_VALUE(v[5], v[4]));
+ write(priv, mod->base + LSC_XSIZE_1213_REG, LSC_GRAD_VALUE(v[3], v[2]));
+ write(priv, mod->base + LSC_XSIZE_1415_REG, LSC_GRAD_VALUE(v[1], v[0]));
+
+ v = cfg->config.y_size_tbl;
+ write(priv, mod->base + LSC_YSIZE_01_REG, LSC_GRAD_VALUE(v[0], v[1]));
+ write(priv, mod->base + LSC_YSIZE_23_REG, LSC_GRAD_VALUE(v[2], v[3]));
+ write(priv, mod->base + LSC_YSIZE_45_REG, LSC_GRAD_VALUE(v[4], v[5]));
+ write(priv, mod->base + LSC_YSIZE_67_REG, LSC_GRAD_VALUE(v[6], v[7]));
+ write(priv, mod->base + LSC_YSIZE_89_REG, LSC_GRAD_VALUE(v[7], v[6]));
+ write(priv, mod->base + LSC_YSIZE_1011_REG, LSC_GRAD_VALUE(v[5], v[4]));
+ write(priv, mod->base + LSC_YSIZE_1213_REG, LSC_GRAD_VALUE(v[3], v[2]));
+ write(priv, mod->base + LSC_YSIZE_1415_REG, LSC_GRAD_VALUE(v[1], v[0]));
+
+ /* Enable module. */
+ write(priv, mod->base + LSC_CTRL_REG, LSC_CTRL_LSC_EN);
+
+ return 0;
+}
+
const struct rpp_module_ops rppx1_lsc_ops = {
.probe = rppx1_lsc_probe,
+ .param_rkisp1 = rppx1_lsc_param_rkisp1,
};
--
2.53.0
next prev parent reply other threads:[~2026-03-14 22:11 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-14 21:59 [PATCH v6 00/12] media: Add support for R-Car ISP using Dreamchip RPPX1 ISP Niklas Söderlund
2026-03-14 21:59 ` [PATCH v6 01/12] media: rppx1: Add framework to support " Niklas Söderlund
2026-03-14 21:59 ` [PATCH v6 02/12] media: rcar-isp: Add support for ISPCORE Niklas Söderlund
2026-03-14 21:59 ` [PATCH v6 03/12] media: rppx1: Add support for AWB measurement parameters and statistics Niklas Söderlund
2026-03-14 21:59 ` [PATCH v6 04/12] media: rppx1: Add support for AWB gain settings Niklas Söderlund
2026-03-14 21:59 ` [PATCH v6 05/12] media: rppx1: Add support for Auto Exposure Measurement Niklas Söderlund
2026-03-14 21:59 ` [PATCH v6 06/12] media: rppx1: Add support for Histogram Measurement Niklas Söderlund
2026-03-14 21:59 ` [PATCH v6 07/12] media: rppx1: Add support for Black Level Subtraction Niklas Söderlund
2026-03-14 21:59 ` [PATCH v6 08/12] media: rppx1: Add support for Color Correction Matrix Niklas Söderlund
2026-03-14 21:59 ` Niklas Söderlund [this message]
2026-03-14 21:59 ` [PATCH v6 10/12] media: rppx1: Add support for Gamma Correction Niklas Söderlund
2026-03-14 21:59 ` [PATCH v6 11/12] media: rppx1: Add support for Bayer Demosaicing Niklas Söderlund
2026-03-14 21:59 ` [PATCH v6 12/12] media: rppx1: Add support for Bilateral Denoising Niklas Söderlund
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=20260314215944.3674865-10-niklas.soderlund+renesas@ragnatech.se \
--to=niklas.soderlund+renesas@ragnatech.se \
--cc=jacopo.mondi@ideasonboard.com \
--cc=kuninori.morimoto.gx@renesas.com \
--cc=laurent.pinchart@ideasonboard.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-media@vger.kernel.org \
--cc=linux-renesas-soc@vger.kernel.org \
--cc=marek.vasut+renesas@mailbox.org \
--cc=mchehab@kernel.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
all inboxes | Powered by JetHome®