From: "Michał Grzelak" <michal.grzelak@intel.com>
To: Suraj Kandpal <suraj.kandpal@intel.com>
Cc: intel-xe@lists.freedesktop.org, intel-gfx@lists.freedesktop.org,
ankit.k.nautiyal@intel.com, mika.kahola@intel.com
Subject: Re: [PATCH 1/3] drm/i915/cx0: Split PLL enabling/disabling in two parts
Date: Wed, 7 Jan 2026 16:22:11 +0100 (CET) [thread overview]
Message-ID: <0205146a-4832-f242-ea67-e9b5fc31d940@intel.com> (raw)
In-Reply-To: <20251230083142.70064-2-suraj.kandpal@intel.com>
[-- Attachment #1: Type: text/plain, Size: 9599 bytes --]
Hi Suraj,
On Tue, 30 Dec 2025, Suraj Kandpal wrote:
> From: Mika Kahola <mika.kahola@intel.com>
>
> Split PLL enabling/disabling in two parts - one for pll setting
> pll dividers and second one to enable/disable pll clock. PLL
> clock enabling/disbling happens via encoder->enable_clock/disable_clock
> function hook. The reason for doing this is that we need to make sure
> the clock enablement happens after PPS ON step to be inline with the
> sequences which we end up violating otherwise. As a result of this
> violation we end up in a hanged state if machine stays idle for more
> that 15 mins.
>
> PLL state verification happens now earlier than the clock is enabled
> which causes a drm warn to be thrown. Silence this warning by
> allowing this check for only earlier platforms than MeteorLake.
>
> Bspec: 49190
> Signed-off-by: Mika Kahola <mika.kahola@intel.com>
> Signed-off-by: Suraj Kandpal <suraj.kandpal@intel.com>
> ---
> drivers/gpu/drm/i915/display/intel_cx0_phy.c | 87 ++++++++++++-------
> drivers/gpu/drm/i915/display/intel_dpll_mgr.c | 12 +--
> 2 files changed, 64 insertions(+), 35 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/display/intel_cx0_phy.c b/drivers/gpu/drm/i915/display/intel_cx0_phy.c
> index 7288065d2461..f3baba264e88 100644
> --- a/drivers/gpu/drm/i915/display/intel_cx0_phy.c
> +++ b/drivers/gpu/drm/i915/display/intel_cx0_phy.c
> @@ -3225,11 +3225,8 @@ static void intel_cx0pll_enable(struct intel_encoder *encoder,
> {
> int port_clock = pll_state->use_c10 ? pll_state->c10.clock : pll_state->c20.clock;
> struct intel_display *display = to_intel_display(encoder);
> - enum phy phy = intel_encoder_to_phy(encoder);
> struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
> bool lane_reversal = dig_port->lane_reversal;
> - u8 maxpclk_lane = lane_reversal ? INTEL_CX0_LANE1 :
> - INTEL_CX0_LANE0;
> struct ref_tracker *wakeref = intel_cx0_phy_transaction_begin(encoder);
>
> /*
> @@ -3284,27 +3281,6 @@ static void intel_cx0pll_enable(struct intel_encoder *encoder,
> */
> intel_de_write(display, DDI_CLK_VALFREQ(encoder->port), port_clock);
>
> - /*
> - * 9. Set PORT_CLOCK_CTL register PCLK PLL Request
> - * LN<Lane for maxPCLK> to "1" to enable PLL.
> - */
> - intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port),
> - intel_cx0_get_pclk_pll_request(INTEL_CX0_BOTH_LANES),
> - intel_cx0_get_pclk_pll_request(maxpclk_lane));
> -
> - /* 10. Poll on PORT_CLOCK_CTL PCLK PLL Ack LN<Lane for maxPCLK> == "1". */
> - if (intel_de_wait_us(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port),
> - intel_cx0_get_pclk_pll_ack(INTEL_CX0_BOTH_LANES),
> - intel_cx0_get_pclk_pll_ack(maxpclk_lane),
> - XELPDP_PCLK_PLL_ENABLE_TIMEOUT_US, NULL))
> - drm_warn(display->drm, "Port %c PLL not locked\n",
> - phy_name(phy));
> -
> - /*
> - * 11. Follow the Display Voltage Frequency Switching Sequence After
> - * Frequency Change. We handle this step in bxt_set_cdclk().
> - */
> -
> /*
> * 12. Toggle powerdown if HDMI is enabled on C10 PHY.
> *
> @@ -3403,6 +3379,42 @@ static int intel_mtl_tbt_clock_select(struct intel_display *display,
> }
> }
>
> +static void intel_cx0pll_enable_clock(struct intel_encoder *encoder)
> +{
> + struct intel_display *display = to_intel_display(encoder);
> + enum phy phy = intel_encoder_to_phy(encoder);
> + struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
> + bool lane_reversal = dig_port->lane_reversal;
> + INTEL_CX0_LANE0;
Line above breaks the build. My current understanding is that it causes
compiler to expand INTEL_CX0_LANE0 into BIT(0), which expands into 1UL
<< 0. Finally, compiler parses it more or less as:
1;
Since it is a valid statement but doesn't do anything, it triggers
unused value warning, which is turned into error.
IOW, with the line removed build works.
The explanation seems consistent with the output:
In file included from ./include/linux/bits.h:5,
from ./include/linux/bitops.h:6,
from ./include/linux/log2.h:12,
from drivers/gpu/drm/i915/display/intel_cx0_phy.c:6:
drivers/gpu/drm/i915/display/intel_cx0_phy.c: In function ‘intel_cx0pll_enable_clock’:
./include/vdso/bits.h:7:33: error: statement with no effect [-Werror=unused-value]
7 | #define BIT(nr) (UL(1) << (nr))
| ^
drivers/gpu/drm/i915/display/intel_cx0_phy.c:32:33: note: in expansion of macro ‘BIT’
32 | #define INTEL_CX0_LANE0 BIT(0)
| ^~~
drivers/gpu/drm/i915/display/intel_cx0_phy.c:3378:43: note: in expansion of macro ‘INTEL_CX0_LANE0’
3378 | INTEL_CX0_LANE0;
| ^~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
Besides of this, for the whole series:
Reviewed-by: Michał Grzelak <michal.grzelak@intel.com>
BR,
Michał
> + u8 maxpclk_lane = lane_reversal ? INTEL_CX0_LANE1 :
> + INTEL_CX0_LANE0;
> +
> + struct ref_tracker *wakeref = intel_cx0_phy_transaction_begin(encoder);
> +
> + /*
> + * 9. Set PORT_CLOCK_CTL register PCLK PLL Request
> + * LN<Lane for maxPCLK> to "1" to enable PLL.
> + */
> + intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port),
> + intel_cx0_get_pclk_pll_request(INTEL_CX0_BOTH_LANES),
> + intel_cx0_get_pclk_pll_request(maxpclk_lane));
> +
> + /* 10. Poll on PORT_CLOCK_CTL PCLK PLL Ack LN<Lane for maxPCLK> == "1". */
> + if (intel_de_wait_us(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port),
> + intel_cx0_get_pclk_pll_ack(INTEL_CX0_BOTH_LANES),
> + intel_cx0_get_pclk_pll_ack(maxpclk_lane),
> + XELPDP_PCLK_PLL_ENABLE_TIMEOUT_US, NULL))
> + drm_warn(display->drm, "Port %c PLL not locked\n",
> + phy_name(phy));
> +
> + /*
> + * 11. Follow the Display Voltage Frequency Switching Sequence After
> + * Frequency Change. We handle this step in bxt_set_cdclk().
> + */
> +
> + intel_cx0_phy_transaction_end(encoder, wakeref);
> +}
> +
> void intel_mtl_tbt_pll_enable_clock(struct intel_encoder *encoder, int port_clock)
> {
> struct intel_display *display = to_intel_display(encoder);
> @@ -3472,6 +3484,8 @@ void intel_mtl_pll_enable_clock(struct intel_encoder *encoder,
>
> if (intel_tc_port_in_tbt_alt_mode(dig_port))
> intel_mtl_tbt_pll_enable_clock(encoder, crtc_state->port_clock);
> + else
> + intel_cx0pll_enable_clock(encoder);
> }
>
> /*
> @@ -3567,12 +3581,6 @@ static void intel_cx0pll_disable(struct intel_encoder *encoder)
> * Frequency Change. We handle this step in bxt_set_cdclk().
> */
>
> - /* 7. Program PORT_CLOCK_CTL register to disable and gate clocks. */
> - intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port),
> - XELPDP_DDI_CLOCK_SELECT_MASK(display), 0);
> - intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port),
> - XELPDP_FORWARD_CLOCK_UNGATE, 0);
> -
> intel_cx0_phy_transaction_end(encoder, wakeref);
> }
>
> @@ -3586,6 +3594,20 @@ static bool intel_cx0_pll_is_enabled(struct intel_encoder *encoder)
> intel_cx0_get_pclk_pll_request(lane);
> }
>
> +static void intel_cx0pll_disable_clock(struct intel_encoder *encoder)
> +{
> + struct intel_display *display = to_intel_display(encoder);
> + struct ref_tracker *wakeref = intel_cx0_phy_transaction_begin(encoder);
> +
> + /* 7. Program PORT_CLOCK_CTL register to disable and gate clocks. */
> + intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port),
> + XELPDP_DDI_CLOCK_SELECT_MASK(display), 0);
> + intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port),
> + XELPDP_FORWARD_CLOCK_UNGATE, 0);
> +
> + intel_cx0_phy_transaction_end(encoder, wakeref);
> +}
> +
> void intel_mtl_tbt_pll_disable_clock(struct intel_encoder *encoder)
> {
> struct intel_display *display = to_intel_display(encoder);
> @@ -3635,6 +3657,9 @@ void intel_mtl_pll_disable_clock(struct intel_encoder *encoder)
>
> if (intel_tc_port_in_tbt_alt_mode(dig_port))
> intel_mtl_tbt_pll_disable_clock(encoder);
> + else
> + intel_cx0pll_disable_clock(encoder);
> +
> }
>
> enum icl_port_dpll_id
> @@ -3783,6 +3808,8 @@ void intel_cx0_pll_power_save_wa(struct intel_display *display)
> encoder->base.base.id, encoder->base.name);
>
> intel_cx0pll_enable(encoder, &pll_state);
> + intel_cx0pll_enable_clock(encoder);
> intel_cx0pll_disable(encoder);
> + intel_cx0pll_disable_clock(encoder);
> }
> }
> diff --git a/drivers/gpu/drm/i915/display/intel_dpll_mgr.c b/drivers/gpu/drm/i915/display/intel_dpll_mgr.c
> index 9aa84a430f09..59395076103c 100644
> --- a/drivers/gpu/drm/i915/display/intel_dpll_mgr.c
> +++ b/drivers/gpu/drm/i915/display/intel_dpll_mgr.c
> @@ -186,11 +186,13 @@ void assert_dpll(struct intel_display *display,
> "asserting DPLL %s with no DPLL\n", str_on_off(state)))
> return;
>
> - cur_state = intel_dpll_get_hw_state(display, pll, &hw_state);
> - INTEL_DISPLAY_STATE_WARN(display, cur_state != state,
> - "%s assertion failure (expected %s, current %s)\n",
> - pll->info->name, str_on_off(state),
> - str_on_off(cur_state));
> + if (DISPLAY_VER(display) < 14) {
> + cur_state = intel_dpll_get_hw_state(display, pll, &hw_state);
> + INTEL_DISPLAY_STATE_WARN(display, cur_state != state,
> + "%s assertion failure (expected %s, current %s)\n",
> + pll->info->name, str_on_off(state),
> + str_on_off(cur_state));
> + }
> }
>
> static enum tc_port icl_pll_id_to_tc_port(enum intel_dpll_id id)
> --
> 2.34.1
>
>
next prev parent reply other threads:[~2026-01-07 15:22 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-12-30 8:31 [PATCH 0/3] Fix Cx0 Suspend Resume issue Suraj Kandpal
2025-12-30 8:31 ` [PATCH 1/3] drm/i915/cx0: Split PLL enabling/disabling in two parts Suraj Kandpal
2025-12-30 15:20 ` Gustavo Sousa
2025-12-31 5:07 ` Kandpal, Suraj
2026-01-05 14:13 ` Gustavo Sousa
2026-01-06 6:19 ` Kandpal, Suraj
2026-01-08 8:51 ` Kahola, Mika
2025-12-30 21:23 ` kernel test robot
2026-01-07 15:22 ` Michał Grzelak [this message]
2025-12-30 8:31 ` [PATCH 2/3] drm/i915/cx0: Move step 12 to enable clock hook Suraj Kandpal
2025-12-30 15:22 ` Gustavo Sousa
2025-12-31 5:10 ` Kandpal, Suraj
2026-01-05 14:37 ` Gustavo Sousa
2026-01-06 6:18 ` Kandpal, Suraj
2025-12-30 8:31 ` [PATCH 3/3] drm/i915/cx0: Clear response ready & error bit Suraj Kandpal
2025-12-30 15:29 ` Jani Nikula
2025-12-30 17:36 ` Gustavo Sousa
2025-12-31 4:59 ` Kandpal, Suraj
2026-01-05 15:14 ` Gustavo Sousa
2026-01-06 6:21 ` Kandpal, Suraj
2025-12-30 8:38 ` ✓ CI.KUnit: success for Fix Cx0 Suspend Resume issue Patchwork
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=0205146a-4832-f242-ea67-e9b5fc31d940@intel.com \
--to=michal.grzelak@intel.com \
--cc=ankit.k.nautiyal@intel.com \
--cc=intel-gfx@lists.freedesktop.org \
--cc=intel-xe@lists.freedesktop.org \
--cc=mika.kahola@intel.com \
--cc=suraj.kandpal@intel.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