All of lore.kernel.org
 help / color / mirror / Atom feed
From: Anthoine Bourgeois <anthoine.bourgeois@gmail.com>
To: "Deucher, Alexander" <Alexander.Deucher@amd.com>,
	"Koenig, Christian" <Christian.Koenig@amd.com>,
	Jerome Glisse <jglisse@redhat.com>
Cc: dri-devel@lists.freedesktop.org
Subject: Re: [PATCH] drm/radeon/dpm: add debugfs support for RS780/RS880 (v2)
Date: Wed, 17 Jul 2013 23:23:13 +0200	[thread overview]
Message-ID: <20130717212235.GA2456@gmail.com> (raw)
In-Reply-To: <A3397C8B8B789E45844E7EC5DEAD89D02E510DA1@sausexdag04.amd.com>

On Wed, Jul 17, 2013 at 08:54:33PM +0000, Deucher, Alexander wrote:
>Good catch, I screwed up the ref divider calculation.  How about the attached patch?

It works too. But here was my logic, tell me if I'm wrong:

The output frequency of the PLL is equal to the VCO frequency (FVCO)
divided by the post-scale counter (C).
Fout = FVCO / C
And VCO frequency is equal to feedback (M) times the reference frequency
(FREF).
FVCO = FREF x M
Where reference clock (FREF) is equal to the input clock (Fin) divided
by the pre-scale counter (N)
FREF = Fin / N
then
FVCO = Fin x M / N
and
Fout = (Fin x M) / (N x C) = FREF x M / C

where:
FVCO = VCO frequency
Fin = input frequency
Fout = output frequency
FREF = reference frequency (rdev->clock.spll.reference_freq in the code)
M = counter (multiplier), part of the clock feedback path (current_fb_div)
N = counter (divider), part of the clock reference path (ref_div)
C = post-scale counter (divider) (post_div)

No need to ref_div. Or reference_freq == Fin and the varname confuses me ?

In my test:
# cat /sys/kernel/debug/dri/0/radeon_pm_info
uvd    vclk: 0 dclk: 0
power level 0    sclk: 20000 vddc_index: 2

when my system is idle and sclk == 19690 and:
# cat /sys/kernel/debug/dri/0/radeon_pm_info
uvd    vclk: 0 dclk: 0
power level 1    sclk: 50000 vddc_index: 2

when I play supertuxkart with sclk == 49762.

Thanks for yesturday, I was in a wrong way,
Anthoine

>
>Alex
>
>> -----Original Message-----
>> From: Anthoine Bourgeois [mailto:anthoine.bourgeois@gmail.com]
>> Sent: Wednesday, July 17, 2013 4:43 PM
>> To: Deucher, Alexander; Koenig, Christian; Jerome Glisse; Anthoine
>> Bourgeois
>> Cc: dri-devel@lists.freedesktop.org
>> Subject: [PATCH] drm/radeon/dpm: add debugfs support for RS780/RS880
>> (v2)
>>
>> This allows you to look at the current DPM state via
>> debugfs.
>>
>> Due to the way the hardware works on these asics, there's
>> no way to look up exactly what power state we are in, so
>> we make the best guess we can based on the current sclk.
>>
>> v2: fix sclk equation
>>
>> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
>> Signed-off-by: Anthoine Bourgeois <anthoine.bourgeois@gmail.com>
>> ---
>>  drivers/gpu/drm/radeon/radeon_asic.c |  1 +
>>  drivers/gpu/drm/radeon/radeon_asic.h |  2 ++
>>  drivers/gpu/drm/radeon/rs780_dpm.c   | 23 +++++++++++++++++++++++
>>  drivers/gpu/drm/radeon/rs780d.h      |  3 +++
>>  4 files changed, 29 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/radeon/radeon_asic.c
>> b/drivers/gpu/drm/radeon/radeon_asic.c
>> index fea997e..78bec1a 100644
>> --- a/drivers/gpu/drm/radeon/radeon_asic.c
>> +++ b/drivers/gpu/drm/radeon/radeon_asic.c
>> @@ -1270,6 +1270,7 @@ static struct radeon_asic rs780_asic = {
>>  		.get_sclk = &rs780_dpm_get_sclk,
>>  		.get_mclk = &rs780_dpm_get_mclk,
>>  		.print_power_state = &rs780_dpm_print_power_state,
>> +		.debugfs_print_current_performance_level =
>> &rs780_dpm_debugfs_print_current_performance_level,
>>  	},
>>  	.pflip = {
>>  		.pre_page_flip = &rs600_pre_page_flip,
>> diff --git a/drivers/gpu/drm/radeon/radeon_asic.h
>> b/drivers/gpu/drm/radeon/radeon_asic.h
>> index b04b578..ca18957 100644
>> --- a/drivers/gpu/drm/radeon/radeon_asic.h
>> +++ b/drivers/gpu/drm/radeon/radeon_asic.h
>> @@ -433,6 +433,8 @@ u32 rs780_dpm_get_sclk(struct radeon_device *rdev,
>> bool low);
>>  u32 rs780_dpm_get_mclk(struct radeon_device *rdev, bool low);
>>  void rs780_dpm_print_power_state(struct radeon_device *rdev,
>>  				 struct radeon_ps *ps);
>> +void rs780_dpm_debugfs_print_current_performance_level(struct
>> radeon_device *rdev,
>> +						       struct seq_file *m);
>>
>>  /* uvd */
>>  int r600_uvd_init(struct radeon_device *rdev);
>> diff --git a/drivers/gpu/drm/radeon/rs780_dpm.c
>> b/drivers/gpu/drm/radeon/rs780_dpm.c
>> index bef832a..7f128c2 100644
>> --- a/drivers/gpu/drm/radeon/rs780_dpm.c
>> +++ b/drivers/gpu/drm/radeon/rs780_dpm.c
>> @@ -28,6 +28,7 @@
>>  #include "r600_dpm.h"
>>  #include "rs780_dpm.h"
>>  #include "atom.h"
>> +#include <linux/seq_file.h>
>>
>>  static struct igp_ps *rs780_get_ps(struct radeon_ps *rps)
>>  {
>> @@ -961,3 +962,25 @@ u32 rs780_dpm_get_mclk(struct radeon_device
>> *rdev, bool low)
>>
>>  	return pi->bootup_uma_clk;
>>  }
>> +
>> +void rs780_dpm_debugfs_print_current_performance_level(struct
>> radeon_device *rdev,
>> +						       struct seq_file *m)
>> +{
>> +	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
>> +	struct igp_ps *ps = rs780_get_ps(rps);
>> +	u32 current_fb_div = RREG32(FVTHROT_STATUS_REG0) &
>> CURRENT_FEEDBACK_DIV_MASK;
>> +	u32 func_cntl = RREG32(CG_SPLL_FUNC_CNTL);
>> +	u32 post_div = ((func_cntl & SPLL_SW_HILEN_MASK) >>
>> SPLL_SW_HILEN_SHIFT) + 1 +
>> +		((func_cntl & SPLL_SW_LOLEN_MASK) >>
>> SPLL_SW_LOLEN_SHIFT) + 1;
>> +	u32 sclk = (rdev->clock.spll.reference_freq * current_fb_div) /
>> post_div;
>> +
>> +	seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
>> +
>> +	/* guess based on the current sclk */
>> +	if (sclk < (ps->sclk_low + 500))
>> +		seq_printf(m, "power level 0    sclk: %u vddc_index: %d\n",
>> +			   ps->sclk_low, ps->min_voltage);
>> +	else
>> +		seq_printf(m, "power level 1    sclk: %u vddc_index: %d\n",
>> +			   ps->sclk_high, ps->max_voltage);
>> +}
>> diff --git a/drivers/gpu/drm/radeon/rs780d.h
>> b/drivers/gpu/drm/radeon/rs780d.h
>> index b1142ed..cfbe9a4 100644
>> --- a/drivers/gpu/drm/radeon/rs780d.h
>> +++ b/drivers/gpu/drm/radeon/rs780d.h
>> @@ -28,6 +28,7 @@
>>  #       define SPLL_SLEEP                                (1 << 1)
>>  #       define SPLL_REF_DIV(x)                           ((x) << 2)
>>  #       define SPLL_REF_DIV_MASK                         (7 << 2)
>> +#       define SPLL_REF_DIV_SHIFT                        2
>>  #       define SPLL_FB_DIV(x)                            ((x) << 5)
>>  #       define SPLL_FB_DIV_MASK                          (0xff << 2)
>>  #       define SPLL_FB_DIV_SHIFT                         2
>> @@ -36,8 +37,10 @@
>>  #       define SPLL_PULSENUM_MASK                        (3 << 14)
>>  #       define SPLL_SW_HILEN(x)                          ((x) << 16)
>>  #       define SPLL_SW_HILEN_MASK                        (0xf << 16)
>> +#       define SPLL_SW_HILEN_SHIFT                       16
>>  #       define SPLL_SW_LOLEN(x)                          ((x) << 20)
>>  #       define SPLL_SW_LOLEN_MASK                        (0xf << 20)
>> +#       define SPLL_SW_LOLEN_SHIFT                       20
>>  #       define SPLL_DIVEN                                (1 << 24)
>>  #       define SPLL_BYPASS_EN                            (1 << 25)
>>  #       define SPLL_CHG_STATUS                           (1 << 29)
>> --
>> 1.8.1.5
>>
>

  reply	other threads:[~2013-07-17 21:20 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-07-17 20:42 [PATCH] drm/radeon/dpm: add debugfs support for RS780/RS880 (v2) Anthoine Bourgeois
2013-07-17 20:54 ` Deucher, Alexander
2013-07-17 21:23   ` Anthoine Bourgeois [this message]
2013-07-17 21:44     ` Deucher, Alexander
2013-07-17 21:55       ` Anthoine Bourgeois

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=20130717212235.GA2456@gmail.com \
    --to=anthoine.bourgeois@gmail.com \
    --cc=Alexander.Deucher@amd.com \
    --cc=Christian.Koenig@amd.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=jglisse@redhat.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.