From: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
To: Mark Tseng <chun-jen.tseng@mediatek.com>,
Rob Herring <robh+dt@kernel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
Conor Dooley <conor+dt@kernel.org>,
Matthias Brugger <matthias.bgg@gmail.com>,
Roger Lu <roger.lu@mediatek.com>,
Kevin Hilman <khilman@kernel.org>
Cc: devicetree@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-arm-kernel@lists.infradead.org,
linux-mediatek@lists.infradead.org,
Project_Global_Chrome_Upstream_Group@mediatek.com
Subject: Re: [PATCH v1 2/2] soc: mediatek: svs: add support for mt8188
Date: Fri, 15 Sep 2023 11:20:49 +0200 [thread overview]
Message-ID: <a6ae9a08-d295-8c1a-840a-84a1b447421b@collabora.com> (raw)
In-Reply-To: <20230915075003.1552-3-chun-jen.tseng@mediatek.com>
Il 15/09/23 09:50, Mark Tseng ha scritto:
> MT8188 svs gpu uses 2-line high bank and low bank to optimize the
> voltage of opp table for higher and lower frequency respectively.
>
> Signed-off-by: Mark Tseng <chun-jen.tseng@mediatek.com>
> ---
> drivers/soc/mediatek/mtk-svs.c | 177 ++++++++++++++++++++++++++++++++-
> 1 file changed, 172 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/soc/mediatek/mtk-svs.c b/drivers/soc/mediatek/mtk-svs.c
> index 3a2f97cd5272..fafc72df99ee 100644
> --- a/drivers/soc/mediatek/mtk-svs.c
> +++ b/drivers/soc/mediatek/mtk-svs.c
> @@ -407,6 +407,7 @@ struct svs_platform_data {
> * @dcbdet: svs efuse data
> * @dcmdet: svs efuse data
> * @turn_pt: 2-line turn point tells which opp_volt calculated by high/low bank
> + * @vbin_turn_pt: voltage bin turn point helps know which svsb_volt should be overridden
> * @type: bank type to represent it is 2-line (high/low) bank or 1-line bank
> *
> * Svs bank will generate suitalbe voltages by below general math equation
> @@ -469,6 +470,7 @@ struct svs_bank {
> u32 dcbdet;
> u32 dcmdet;
> u32 turn_pt;
> + u32 vbin_turn_pt;
> u32 type;
> };
>
> @@ -751,11 +753,12 @@ static int svs_status_debug_show(struct seq_file *m, void *v)
>
> ret = thermal_zone_get_temp(svsb->tzd, &tzone_temp);
> if (ret)
> - seq_printf(m, "%s: temperature ignore, turn_pt = %u\n",
> - svsb->name, svsb->turn_pt);
> + seq_printf(m, "%s: temperature ignore, vbin_turn_pt = %u, turn_pt = %u\n",
> + svsb->name, svsb->vbin_turn_pt, svsb->turn_pt);
> else
> - seq_printf(m, "%s: temperature = %d, turn_pt = %u\n",
> - svsb->name, tzone_temp, svsb->turn_pt);
> + seq_printf(m, "%s: temperature = %d, vbin_turn_pt = %u, turn_pt = %u\n",
> + svsb->name, tzone_temp, svsb->vbin_turn_pt,
> + svsb->turn_pt);
>
> for (i = 0; i < svsb->opp_count; i++) {
> opp = dev_pm_opp_find_freq_exact(svsb->opp_dev,
> @@ -952,6 +955,29 @@ static void svs_get_bank_volts_v3(struct svs_platform *svsp)
> for (i = opp_start; i < opp_stop; i++)
> if (svsb->volt_flags & SVSB_REMOVE_DVTFIXED_VOLT)
> svsb->volt[i] -= svsb->dvt_fixed;
> +
> + /* For voltage bin support */
> + if (svsb->opp_dfreq[0] > svsb->freq_base) {
> + svsb->volt[0] = svs_opp_volt_to_bank_volt(svsb->opp_dvolt[0],
> + svsb->volt_step,
> + svsb->volt_base);
> +
> + /* Find voltage bin turn point */
> + for (i = 0; i < svsb->opp_count; i++) {
> + if (svsb->opp_dfreq[i] <= svsb->freq_base) {
> + svsb->vbin_turn_pt = i;
> + break;
> + }
> + }
> +
> + /* Override svs bank voltages */
> + for (i = 1; i < svsb->vbin_turn_pt; i++)
> + svsb->volt[i] = interpolate(svsb->freq_pct[0],
> + svsb->freq_pct[svsb->vbin_turn_pt],
> + svsb->volt[0],
> + svsb->volt[svsb->vbin_turn_pt],
> + svsb->freq_pct[i]);
> + }
This looks like being a new feature that applies to more than just MT8188?
Goes into a different commit.
> }
>
> static void svs_set_bank_freq_pct_v3(struct svs_platform *svsp)
> @@ -1808,6 +1834,80 @@ static bool svs_mt8192_efuse_parsing(struct svs_platform *svsp)
> return true;
> }
>
> +static bool svs_mt8188_efuse_parsing(struct svs_platform *svsp)
> +{
> + struct svs_bank *svsb;
> + struct nvmem_cell *cell;
> + u32 idx, i, golden_temp;
> +
> + for (i = 0; i < svsp->efuse_max; i++)
> + if (svsp->efuse[i])
> + dev_info(svsp->dev, "M_HW_RES%d: 0x%08x\n",
> + i, svsp->efuse[i]);
> +
> + if (!svsp->efuse[5]) {
> + dev_notice(svsp->dev, "svs_efuse[5] = 0x0?\n");
> + return false;
> + }
> +
> + /* Svs efuse parsing */
> + for (idx = 0; idx < svsp->bank_max; idx++) {
> + svsb = &svsp->banks[idx];
> +
> + if (svsb->type == SVSB_LOW) {
> + svsb->mtdes = svsp->efuse[5] & GENMASK(7, 0);
> + svsb->bdes = (svsp->efuse[5] >> 16) & GENMASK(7, 0);
> + svsb->mdes = (svsp->efuse[5] >> 24) & GENMASK(7, 0);
> + svsb->dcbdet = (svsp->efuse[15] >> 16) & GENMASK(7, 0);
> + svsb->dcmdet = (svsp->efuse[15] >> 24) & GENMASK(7, 0);
> + } else if (svsb->type == SVSB_HIGH) {
> + svsb->mtdes = svsp->efuse[4] & GENMASK(7, 0);
> + svsb->bdes = (svsp->efuse[4] >> 16) & GENMASK(7, 0);
> + svsb->mdes = (svsp->efuse[4] >> 24) & GENMASK(7, 0);
> + svsb->dcbdet = svsp->efuse[14] & GENMASK(7, 0);
> + svsb->dcmdet = (svsp->efuse[14] >> 8) & GENMASK(7, 0);
> + }
> +
> + svsb->vmax += svsb->dvt_fixed;
> + }
The only change between mt8192, 8195 and 8188 is this for loop; can we please
commonize the rest of the function?
> +
> + /* Thermal efuse parsing */
> + cell = nvmem_cell_get(svsp->dev, "t-calibration-data");
> + if (IS_ERR_OR_NULL(cell)) {
> + dev_err(svsp->dev, "no \"t-calibration-data\"? %ld\n",
> + PTR_ERR(cell));
> + return false;
> + }
> +
> + svsp->tefuse = nvmem_cell_read(cell, &svsp->tefuse_max);
> + if (IS_ERR(svsp->tefuse)) {
> + dev_err(svsp->dev, "cannot read thermal efuse: %ld\n",
> + PTR_ERR(svsp->tefuse));
> + nvmem_cell_put(cell);
> + return false;
> + }
> +
> + svsp->tefuse_max /= sizeof(u32);
> + nvmem_cell_put(cell);
> +
> + for (i = 0; i < svsp->tefuse_max; i++)
> + if (svsp->tefuse[i] != 0)
> + break;
> +
> + if (i == svsp->tefuse_max)
> + golden_temp = 50; /* All thermal efuse data are 0 */
> + else
> + golden_temp = (svsp->tefuse[0] & GENMASK(31, 24)) >> 24;
> +
> + for (idx = 0; idx < svsp->bank_max; idx++) {
> + svsb = &svsp->banks[idx];
> + svsb->mts = 500;
> + svsb->bts = (((500 * golden_temp + 250460) / 1000) - 25) * 4;
> + }
> +
> + return true;
> +}
> +
> static bool svs_mt8183_efuse_parsing(struct svs_platform *svsp)
> {
> struct svs_bank *svsb;
> @@ -2052,7 +2152,7 @@ static int svs_mt8192_platform_probe(struct svs_platform *svsp)
> return dev_err_probe(svsp->dev, PTR_ERR(svsp->rst),
> "cannot get svs reset control\n");
>
> - dev = svs_add_device_link(svsp, "lvts");
> + dev = svs_add_device_link(svsp, "thermal-sensor");
This is a fix that must go in a different commit with a Fixes tag.
> if (IS_ERR(dev))
> return dev_err_probe(svsp->dev, PTR_ERR(dev),
> "failed to get lvts device\n");
> @@ -2173,6 +2273,61 @@ static struct svs_bank svs_mt8192_banks[] = {
> },
> };
>
Regards,
Angelo
next prev parent reply other threads:[~2023-09-15 9:20 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-09-15 7:50 [PATCH v1 0/2] soc: mediatek: svs: add support for mt8188 Mark Tseng
2023-09-15 7:50 ` [PATCH v1 1/2] dt-bindings: soc: mediatek: add mt8188 svs dt-bindings Mark Tseng
2023-09-15 7:52 ` Krzysztof Kozlowski
2023-09-15 9:22 ` AngeloGioacchino Del Regno
2023-09-15 9:24 ` AngeloGioacchino Del Regno
2023-09-15 7:50 ` [PATCH v1 2/2] soc: mediatek: svs: add support for mt8188 Mark Tseng
2023-09-15 9:20 ` AngeloGioacchino Del Regno [this message]
2023-09-18 3:00 ` Chun-Jen Tseng (曾俊仁)
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=a6ae9a08-d295-8c1a-840a-84a1b447421b@collabora.com \
--to=angelogioacchino.delregno@collabora.com \
--cc=Project_Global_Chrome_Upstream_Group@mediatek.com \
--cc=chun-jen.tseng@mediatek.com \
--cc=conor+dt@kernel.org \
--cc=devicetree@vger.kernel.org \
--cc=khilman@kernel.org \
--cc=krzysztof.kozlowski+dt@linaro.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mediatek@lists.infradead.org \
--cc=matthias.bgg@gmail.com \
--cc=robh+dt@kernel.org \
--cc=roger.lu@mediatek.com \
/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