[Intel-gfx] [PATCH v3 3/9] drm/i915/dp: Move vswing/pre-emphasis adjustment calculation
Manna, Animesh
animesh.manna at intel.com
Mon Jan 6 10:02:09 UTC 2020
On 04-01-2020 05:18, Manasi Navare wrote:
> On Thu, Jan 02, 2020 at 03:56:09PM +0530, Manna, Animesh wrote:
>> On 02-01-2020 14:48, Jani Nikula wrote:
>>> On Mon, 30 Dec 2019, Animesh Manna <animesh.manna at intel.com> wrote:
>>>> vswing/pre-emphasis adjustment calculation is needed in processing
>>>> of auto phy compliance request other than link training, so moved
>>>> the same function in intel_dp.c.
>>> I guess I'm still asking why you think this is better located in
>>> intel_dp.c than intel_dp_link_training.c, as the function has been moved
>>> once in the other direction already to split out stuff from intel_dp.c
>>> and to make the file smaller. Even the file name suggests it should
>>> really be in intel_dp_link_training.c, right?
>> Just a thought, can we change the name to "intel_dp_link_config.c" from "intel_dp_link_training.c" which will provide little wider scope
>> and all the function playing with link configuration can be under it and also exposed through header file.
>>
>> AFAIK, processing phy compliance request always do not need link training. I understood link training is very specific process consisting of clock recovery + channel eq.
>> So I am afraid of exposing intel_get_adjust_train() from intel_dp_link_training.c which is not only specific to link-training. Need your suggestion.
>>
>> Regards,
>> Animesh
>>
> I agree with Jani here and I think I had even suggested this earlier that instead of moving this function to intel_dp.c
> we should make it non static so it can be used even for PHY compliance but since this function still deals
> with adjusting training patterns IMHO it should still stay in intel_dp_link_training.c
>
> Manasi
Sure Manasi, I will make intel_get_adjust_train() non-static and keep in intel_dp_link_training.c.
Now as suggested by Jani before (https://patchwork.freedesktop.org/patch/345823/?series=71121&rev=1#comment_640087) other non static functions are not following the rule,
should I change the function name or keep it as it is?
Regards,
Animesh
>
>>> BR,
>>> Jani.
>>>
>>>
>>>> No functional change.
>>>>
>>>> v1: initial patch.
>>>> v2:
>>>> - used "intel_dp" prefix in function name. (Jani)
>>>> - used array notation instead pointer for link_status. (Ville)
>>>>
>>>> Signed-off-by: Animesh Manna <animesh.manna at intel.com>
>>>> ---
>>>> drivers/gpu/drm/i915/display/intel_dp.c | 34 ++++++++++++++++++
>>>> drivers/gpu/drm/i915/display/intel_dp.h | 4 +++
>>>> .../drm/i915/display/intel_dp_link_training.c | 36 ++-----------------
>>>> 3 files changed, 40 insertions(+), 34 deletions(-)
>>>>
>>>> diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
>>>> index 991f343579ef..2a27ee106089 100644
>>>> --- a/drivers/gpu/drm/i915/display/intel_dp.c
>>>> +++ b/drivers/gpu/drm/i915/display/intel_dp.c
>>>> @@ -4110,6 +4110,40 @@ ivb_cpu_edp_signal_levels(u8 train_set)
>>>> }
>>>> }
>>>> +void
>>>> +intel_dp_get_adjust_train(struct intel_dp *intel_dp,
>>>> + const u8 link_status[DP_LINK_STATUS_SIZE])
>>>> +{
>>>> + u8 v = 0;
>>>> + u8 p = 0;
>>>> + int lane;
>>>> + u8 voltage_max;
>>>> + u8 preemph_max;
>>>> +
>>>> + for (lane = 0; lane < intel_dp->lane_count; lane++) {
>>>> + u8 this_v = drm_dp_get_adjust_request_voltage(link_status,
>>>> + lane);
>>>> + u8 this_p = drm_dp_get_adjust_request_pre_emphasis(link_status,
>>>> + lane);
>>>> +
>>>> + if (this_v > v)
>>>> + v = this_v;
>>>> + if (this_p > p)
>>>> + p = this_p;
>>>> + }
>>>> +
>>>> + voltage_max = intel_dp_voltage_max(intel_dp);
>>>> + if (v >= voltage_max)
>>>> + v = voltage_max | DP_TRAIN_MAX_SWING_REACHED;
>>>> +
>>>> + preemph_max = intel_dp_pre_emphasis_max(intel_dp, v);
>>>> + if (p >= preemph_max)
>>>> + p = preemph_max | DP_TRAIN_MAX_PRE_EMPHASIS_REACHED;
>>>> +
>>>> + for (lane = 0; lane < 4; lane++)
>>>> + intel_dp->train_set[lane] = v | p;
>>>> +}
>>>> +
>>>> void
>>>> intel_dp_set_signal_levels(struct intel_dp *intel_dp)
>>>> {
>>>> diff --git a/drivers/gpu/drm/i915/display/intel_dp.h b/drivers/gpu/drm/i915/display/intel_dp.h
>>>> index 3da166054788..83eadc87af26 100644
>>>> --- a/drivers/gpu/drm/i915/display/intel_dp.h
>>>> +++ b/drivers/gpu/drm/i915/display/intel_dp.h
>>>> @@ -9,6 +9,7 @@
>>>> #include <linux/types.h>
>>>> #include <drm/i915_drm.h>
>>>> +#include <drm/drm_dp_helper.h>
>>>> #include "i915_reg.h"
>>>> @@ -91,6 +92,9 @@ void
>>>> intel_dp_program_link_training_pattern(struct intel_dp *intel_dp,
>>>> u8 dp_train_pat);
>>>> void
>>>> +intel_dp_get_adjust_train(struct intel_dp *intel_dp,
>>>> + const u8 link_status[DP_LINK_STATUS_SIZE]);
>>>> +void
>>>> intel_dp_set_signal_levels(struct intel_dp *intel_dp);
>>>> void intel_dp_set_idle_link_train(struct intel_dp *intel_dp);
>>>> u8
>>>> diff --git a/drivers/gpu/drm/i915/display/intel_dp_link_training.c b/drivers/gpu/drm/i915/display/intel_dp_link_training.c
>>>> index 2a1130dd1ad0..e8ff9e279800 100644
>>>> --- a/drivers/gpu/drm/i915/display/intel_dp_link_training.c
>>>> +++ b/drivers/gpu/drm/i915/display/intel_dp_link_training.c
>>>> @@ -34,38 +34,6 @@ intel_dp_dump_link_status(const u8 link_status[DP_LINK_STATUS_SIZE])
>>>> link_status[3], link_status[4], link_status[5]);
>>>> }
>>>> -static void
>>>> -intel_get_adjust_train(struct intel_dp *intel_dp,
>>>> - const u8 link_status[DP_LINK_STATUS_SIZE])
>>>> -{
>>>> - u8 v = 0;
>>>> - u8 p = 0;
>>>> - int lane;
>>>> - u8 voltage_max;
>>>> - u8 preemph_max;
>>>> -
>>>> - for (lane = 0; lane < intel_dp->lane_count; lane++) {
>>>> - u8 this_v = drm_dp_get_adjust_request_voltage(link_status, lane);
>>>> - u8 this_p = drm_dp_get_adjust_request_pre_emphasis(link_status, lane);
>>>> -
>>>> - if (this_v > v)
>>>> - v = this_v;
>>>> - if (this_p > p)
>>>> - p = this_p;
>>>> - }
>>>> -
>>>> - voltage_max = intel_dp_voltage_max(intel_dp);
>>>> - if (v >= voltage_max)
>>>> - v = voltage_max | DP_TRAIN_MAX_SWING_REACHED;
>>>> -
>>>> - preemph_max = intel_dp_pre_emphasis_max(intel_dp, v);
>>>> - if (p >= preemph_max)
>>>> - p = preemph_max | DP_TRAIN_MAX_PRE_EMPHASIS_REACHED;
>>>> -
>>>> - for (lane = 0; lane < 4; lane++)
>>>> - intel_dp->train_set[lane] = v | p;
>>>> -}
>>>> -
>>>> static bool
>>>> intel_dp_set_link_train(struct intel_dp *intel_dp,
>>>> u8 dp_train_pat)
>>>> @@ -215,7 +183,7 @@ intel_dp_link_training_clock_recovery(struct intel_dp *intel_dp)
>>>> voltage = intel_dp->train_set[0] & DP_TRAIN_VOLTAGE_SWING_MASK;
>>>> /* Update training set as requested by target */
>>>> - intel_get_adjust_train(intel_dp, link_status);
>>>> + intel_dp_get_adjust_train(intel_dp, link_status);
>>>> if (!intel_dp_update_link_train(intel_dp)) {
>>>> DRM_ERROR("failed to update link training\n");
>>>> return false;
>>>> @@ -325,7 +293,7 @@ intel_dp_link_training_channel_equalization(struct intel_dp *intel_dp)
>>>> }
>>>> /* Update training set as requested by target */
>>>> - intel_get_adjust_train(intel_dp, link_status);
>>>> + intel_dp_get_adjust_train(intel_dp, link_status);
>>>> if (!intel_dp_update_link_train(intel_dp)) {
>>>> DRM_ERROR("failed to update link training\n");
>>>> break;
More information about the Intel-gfx
mailing list