All of lore.kernel.org
 help / color / mirror / Atom feed
From: Hans de Goede <hdegoede@redhat.com>
To: "Ville Syrjälä" <ville.syrjala@linux.intel.com>
Cc: Wolfram Sang <wsa@the-dreams.de>,
	"russianneuromancer @ ya . ru" <russianneuromancer@ya.ru>,
	intel-gfx <intel-gfx@lists.freedesktop.org>,
	linux-i2c@vger.kernel.org,
	Jarkko Nikula <jarkko.nikula@linux.intel.com>,
	dri-devel@lists.freedesktop.org,
	"H . Peter Anvin" <hpa@zytor.com>,
	Daniel Vetter <daniel.vetter@intel.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	Mika Westerberg <mika.westerberg@linux.intel.com>,
	Len Brown <lenb@kernel.org>
Subject: Re: [PATCH v2 13/13] drm/i915: Acquire P-Unit access when modifying P-Unit settings
Date: Mon, 30 Jan 2017 16:27:58 +0100	[thread overview]
Message-ID: <dd41cb48-dfa1-e3d9-7850-7886f029944e@redhat.com> (raw)
In-Reply-To: <20170130151144.GP31595@intel.com>

Hi,

On 30-01-17 16:11, Ville Syrjälä wrote:
> On Mon, Jan 30, 2017 at 04:02:19PM +0100, Hans de Goede wrote:
>> Hi,
>>
>> On 30-01-17 14:10, Ville Syrjälä wrote:
>>> On Sat, Jan 28, 2017 at 06:18:45PM +0100, Hans de Goede wrote:
>>>> Hi,
>>>>
>>>> On 01/28/2017 05:25 PM, Hans de Goede wrote:
>>>>> Hi,
>>>>>
>>>>> On 01/27/2017 02:51 PM, Ville Syrjälä wrote:
>>>>>> On Mon, Jan 23, 2017 at 10:09:58PM +0100, Hans de Goede wrote:
>>>>>>> Make sure the P-Unit or the PMIC i2c bus is not in use when we send a
>>>>>>> request to the P-Unit by calling iosf_mbi_punit_acquire() / _release()
>>>>>>> around P-Unit write accesses.
>>>>>>
>>>>>> Can't we just stuff the calls into the actual punit write function
>>>>>> rather than sprinkling them all over the place?
>>>>>
>>>>> punit access is acquired across sections like this:
>>>>>
>>>>>         iosf_mbi_punit_acquire();
>>>>>
>>>>>         val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
>>>>>         val &= ~DSPFREQGUAR_MASK;
>>>>>         val |= (cmd << DSPFREQGUAR_SHIFT);
>>>>>         vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, val);
>>>>>         if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) &
>>>>>                       DSPFREQSTAT_MASK) == (cmd << DSPFREQSTAT_SHIFT),
>>>>>                      50)) {
>>>>>                 DRM_ERROR("timed out waiting for CDclk change\n");
>>>>>         }
>>>>>         iosf_mbi_punit_release();
>>>>>
>>>>> Where we want to wait for the requested change to have taken
>>>>> effect before releasing the punit.
>>>
>>> Hmm. That's somewhat unfortunate. It also highlights a problem with the
>>> patch wrt. RPS. We don't wait for the GPU to actually change frequencies
>>> in set_rps() because that would slow things down too much. So I have to
>>> wonder how much luck is needed to make this workaround really effective.
>>
>> So the history of this patch-set is that I wrote this patch before
>> writing the patch to get FORCEWAKE_ALL before the pmic bus becomes
>> active (patch 12/13). Since a lot of testing was done with this
>> patch included in the patch-set and since it seemed a good idea
>> regardless (given my experience with accessing the punit vs
>> pmic bus accesses) I decided to leave it in.
>>
>> Possibly just the patch to get FORCEWAKE_ALL is enough, that one
>> actually fixed things for me. That is also why I made this the
>> last patch in the set. I asked tagorereddy to test his system
>> without this patch, but he did not get around to that.
>>
>> After all we do tell the punit to not touch the bus by acquiring
>> the pmic bus semaphore from i2c-desigware-baytrail.c, so maybe
>> for RPS freq changes it honors that and properly waits. Maybe it
>> honors that for all punit requests i915 does and the only real
>> problem is the forcewake stuff ?
>>
>> I can try to drop this patch from my queue and run without it
>> for a while and see if things don't regress. And also ask
>> tagorereddy again to test his system that way.
>>
>> Does that (dropping this patch for now) sound like a good idea?
>
> More test results couldn't hurt at least. It also makes me wonder if
> just bumping the timeouts to some ridiculously high value would fix
> the problem as well.

I've already tried bumping the forcewake timeout from 50 to 250ms,
before writing the patch to just get forcewake_all before the pmic
bus access begins, that does not fix things, and since we busy wait
for this timeout from non-sleeping context 250ms already is way too
high.

>>>>>> + a comment would be nice why it's there.
>>>>>
>>>>> I will add comments to the acquire calls.
>>>>>
>>>>>> Do we need a kconfig select/depends on the iosf_mbi thing? Or some
>>>>>> ifdefs?
>>>>>
>>>>> No, the iosf_mbi header defines empty inline versions of
>>>>> iosf_mbi_punit_acquire / _release if IOSF_MBI is disabled,
>>>>> this does mean that iosf_mbi must be builtin if the i915
>>>>> driver is. I'll add:
>>>>>
>>>>>     depends on DRM_I915=IOSF_MBI || IOSF_MBI=y
>>>>>
>>>>> To the i915 Kconfig to enforce this.
>>>>
>>>> Hmm, ok so that does not work (long cyclic dependency through the
>>>> selection of ACPI_VIDEO).
>>>>
>>>> So I've now added this instead:
>>>>
>>>> 	# iosf_mbi needs to be builtin if we are builtin
>>>> 	select IOSF_MBI if DRM_I915=y
>>>
>>> That's probably not going to help anyone since i915 is usually a module.
>>
>> Right, that is fine, then either the IOSF_MBI symbols are available,
>> or IOSF_MBI is disabled and we get the inline nops from the header.
>>
>> The problem scenario is DRM_I915=y and IOSF_MBI=m, which is not very
>> realistic IMHO, but will get triggered by the random-config testing
>> several contributors do and lead to an unresolved symbol error there.
>
> Well, from the user POV anything with IOSF_MBI==n can be a problem.
> So I'm not sure if we should allow that.

So you're suggesting we just add an unconditional "select IOSF_MBI"
to the i915 Kconfig entry?

Regards,

Hans



>>>>>>> BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=155241
>>>>>>> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
>>>>>>> Tested-by: tagorereddy <tagore.chandan@gmail.com>
>>>>>>> ---
>>>>>>> Changes in v2:
>>>>>>> -Spelling: P-Unit, PMIC
>>>>>>> -Adjust for iosf_mbi_punit_lock/_unlock to _acquire/_release rename
>>>>>>> ---
>>>>>>>  drivers/gpu/drm/i915/intel_display.c    | 6 ++++++
>>>>>>>  drivers/gpu/drm/i915/intel_pm.c         | 9 +++++++++
>>>>>>>  drivers/gpu/drm/i915/intel_runtime_pm.c | 9 +++++++++
>>>>>>>  3 files changed, 24 insertions(+)
>>>>>>>
>>>>>>> diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
>>>>>>> index 5604701..13e5152 100644
>>>>>>> --- a/drivers/gpu/drm/i915/intel_display.c
>>>>>>> +++ b/drivers/gpu/drm/i915/intel_display.c
>>>>>>> @@ -47,6 +47,7 @@
>>>>>>>  #include <drm/drm_rect.h>
>>>>>>>  #include <linux/dma_remapping.h>
>>>>>>>  #include <linux/reservation.h>
>>>>>>> +#include <asm/iosf_mbi.h>
>>>>>>>
>>>>>>>  static bool is_mmio_work(struct intel_flip_work *work)
>>>>>>>  {
>>>>>>> @@ -6421,6 +6422,8 @@ static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
>>>>>>>          cmd = 0;
>>>>>>>
>>>>>>>      mutex_lock(&dev_priv->rps.hw_lock);
>>>>>>> +    iosf_mbi_punit_acquire();
>>>>>>> +
>>>>>>>      val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
>>>>>>>      val &= ~DSPFREQGUAR_MASK;
>>>>>>>      val |= (cmd << DSPFREQGUAR_SHIFT);
>>>>>>> @@ -6430,6 +6433,7 @@ static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
>>>>>>>               50)) {
>>>>>>>          DRM_ERROR("timed out waiting for CDclk change\n");
>>>>>>>      }
>>>>>>> +    iosf_mbi_punit_release();
>>>>>>>      mutex_unlock(&dev_priv->rps.hw_lock);
>>>>>>>
>>>>>>>      mutex_lock(&dev_priv->sb_lock);
>>>>>>> @@ -6497,6 +6501,7 @@ static void cherryview_set_cdclk(struct drm_device *dev, int cdclk)
>>>>>>>      cmd = DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1, cdclk) - 1;
>>>>>>>
>>>>>>>      mutex_lock(&dev_priv->rps.hw_lock);
>>>>>>> +    iosf_mbi_punit_acquire();
>>>>>>>      val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
>>>>>>>      val &= ~DSPFREQGUAR_MASK_CHV;
>>>>>>>      val |= (cmd << DSPFREQGUAR_SHIFT_CHV);
>>>>>>> @@ -6506,6 +6511,7 @@ static void cherryview_set_cdclk(struct drm_device *dev, int cdclk)
>>>>>>>               50)) {
>>>>>>>          DRM_ERROR("timed out waiting for CDclk change\n");
>>>>>>>      }
>>>>>>> +    iosf_mbi_punit_release();
>>>>>>>      mutex_unlock(&dev_priv->rps.hw_lock);
>>>>>>>
>>>>>>>      intel_update_cdclk(dev_priv);
>>>>>>> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
>>>>>>> index 249623d..adff84a 100644
>>>>>>> --- a/drivers/gpu/drm/i915/intel_pm.c
>>>>>>> +++ b/drivers/gpu/drm/i915/intel_pm.c
>>>>>>> @@ -32,6 +32,7 @@
>>>>>>>  #include "../../../platform/x86/intel_ips.h"
>>>>>>>  #include <linux/module.h>
>>>>>>>  #include <drm/drm_atomic_helper.h>
>>>>>>> +#include <asm/iosf_mbi.h>
>>>>>>>
>>>>>>>  /**
>>>>>>>   * DOC: RC6
>>>>>>> @@ -276,6 +277,7 @@ static void chv_set_memory_dvfs(struct drm_i915_private *dev_priv, bool enable)
>>>>>>>      u32 val;
>>>>>>>
>>>>>>>      mutex_lock(&dev_priv->rps.hw_lock);
>>>>>>> +    iosf_mbi_punit_acquire();
>>>>>>>
>>>>>>>      val = vlv_punit_read(dev_priv, PUNIT_REG_DDR_SETUP2);
>>>>>>>      if (enable)
>>>>>>> @@ -290,6 +292,7 @@ static void chv_set_memory_dvfs(struct drm_i915_private *dev_priv, bool enable)
>>>>>>>                FORCE_DDR_FREQ_REQ_ACK) == 0, 3))
>>>>>>>          DRM_ERROR("timed out waiting for Punit DDR DVFS request\n");
>>>>>>>
>>>>>>> +    iosf_mbi_punit_release();
>>>>>>>      mutex_unlock(&dev_priv->rps.hw_lock);
>>>>>>>  }
>>>>>>>
>>>>>>> @@ -298,6 +301,7 @@ static void chv_set_memory_pm5(struct drm_i915_private *dev_priv, bool enable)
>>>>>>>      u32 val;
>>>>>>>
>>>>>>>      mutex_lock(&dev_priv->rps.hw_lock);
>>>>>>> +    iosf_mbi_punit_acquire();
>>>>>>>
>>>>>>>      val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
>>>>>>>      if (enable)
>>>>>>> @@ -306,6 +310,7 @@ static void chv_set_memory_pm5(struct drm_i915_private *dev_priv, bool enable)
>>>>>>>          val &= ~DSP_MAXFIFO_PM5_ENABLE;
>>>>>>>      vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, val);
>>>>>>>
>>>>>>> +    iosf_mbi_punit_release();
>>>>>>>      mutex_unlock(&dev_priv->rps.hw_lock);
>>>>>>>  }
>>>>>>>
>>>>>>> @@ -4553,6 +4558,7 @@ void vlv_wm_get_hw_state(struct drm_device *dev)
>>>>>>>
>>>>>>>      if (IS_CHERRYVIEW(dev_priv)) {
>>>>>>>          mutex_lock(&dev_priv->rps.hw_lock);
>>>>>>> +        iosf_mbi_punit_acquire();
>>>>>>>
>>>>>>>          val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
>>>>>>>          if (val & DSP_MAXFIFO_PM5_ENABLE)
>>>>>>> @@ -4582,6 +4588,7 @@ void vlv_wm_get_hw_state(struct drm_device *dev)
>>>>>>>                  wm->level = VLV_WM_LEVEL_DDR_DVFS;
>>>>>>>          }
>>>>>>>
>>>>>>> +        iosf_mbi_punit_release();
>>>>>>>          mutex_unlock(&dev_priv->rps.hw_lock);
>>>>>>>      }
>>>>>>>
>>>>>>> @@ -4988,7 +4995,9 @@ static void valleyview_set_rps(struct drm_i915_private *dev_priv, u8 val)
>>>>>>>      I915_WRITE(GEN6_PMINTRMSK, gen6_rps_pm_mask(dev_priv, val));
>>>>>>>
>>>>>>>      if (val != dev_priv->rps.cur_freq) {
>>>>>>> +        iosf_mbi_punit_acquire();
>>>>>>>          vlv_punit_write(dev_priv, PUNIT_REG_GPU_FREQ_REQ, val);
>>>>>>> +        iosf_mbi_punit_release();
>>>>>>>          if (!IS_CHERRYVIEW(dev_priv))
>>>>>>>              gen6_set_rps_thresholds(dev_priv, val);
>>>>>>>      }
>>>>>>> diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
>>>>>>> index c0b7e95..e66bcc8 100644
>>>>>>> --- a/drivers/gpu/drm/i915/intel_runtime_pm.c
>>>>>>> +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
>>>>>>> @@ -28,6 +28,7 @@
>>>>>>>
>>>>>>>  #include <linux/pm_runtime.h>
>>>>>>>  #include <linux/vgaarb.h>
>>>>>>> +#include <asm/iosf_mbi.h>
>>>>>>>
>>>>>>>  #include "i915_drv.h"
>>>>>>>  #include "intel_drv.h"
>>>>>>> @@ -1027,6 +1028,8 @@ static void vlv_set_power_well(struct drm_i915_private *dev_priv,
>>>>>>>      if (COND)
>>>>>>>          goto out;
>>>>>>>
>>>>>>> +    iosf_mbi_punit_acquire();
>>>>>>> +
>>>>>>>      ctrl = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL);
>>>>>>>      ctrl &= ~mask;
>>>>>>>      ctrl |= state;
>>>>>>> @@ -1037,6 +1040,8 @@ static void vlv_set_power_well(struct drm_i915_private *dev_priv,
>>>>>>>                state,
>>>>>>>                vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL));
>>>>>>>
>>>>>>> +    iosf_mbi_punit_release();
>>>>>>> +
>>>>>>>  #undef COND
>>>>>>>
>>>>>>>  out:
>>>>>>> @@ -1643,6 +1648,8 @@ static void chv_set_pipe_power_well(struct drm_i915_private *dev_priv,
>>>>>>>      if (COND)
>>>>>>>          goto out;
>>>>>>>
>>>>>>> +    iosf_mbi_punit_acquire();
>>>>>>> +
>>>>>>>      ctrl = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
>>>>>>>      ctrl &= ~DP_SSC_MASK(pipe);
>>>>>>>      ctrl |= enable ? DP_SSC_PWR_ON(pipe) : DP_SSC_PWR_GATE(pipe);
>>>>>>> @@ -1653,6 +1660,8 @@ static void chv_set_pipe_power_well(struct drm_i915_private *dev_priv,
>>>>>>>                state,
>>>>>>>                vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ));
>>>>>>>
>>>>>>> +    iosf_mbi_punit_release();
>>>>>>> +
>>>>>>>  #undef COND
>>>>>>>
>>>>>>>  out:
>>>>>>> --
>>>>>>> 2.9.3
>>>>>>
>>>
>
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

  reply	other threads:[~2017-01-30 15:27 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-01-23 21:09 [PATCH v2 00/13] coordinate cht i2c-pmic and i915-punit accesses Hans de Goede
2017-01-23 21:09 ` [PATCH v2 01/13] x86/platform/intel/iosf_mbi: Add a mutex for P-Unit access Hans de Goede
2017-01-23 21:09 ` [PATCH v2 02/13] x86/platform/intel/iosf_mbi: Add a PMIC bus access notifier Hans de Goede
2017-01-23 21:09 ` [PATCH v2 03/13] i2c: designware: Rename accessor_flags to flags Hans de Goede
2017-01-23 21:09 ` [PATCH v2 04/13] i2c: designware-baytrail: Pass dw_i2c_dev into helper functions Hans de Goede
2017-01-23 21:09 ` [PATCH v2 05/13] i2c: designware-baytrail: Only check iosf_mbi_available() for shared hosts Hans de Goede
2017-01-23 21:09 ` [PATCH v2 06/13] i2c: designware-baytrail: Disallow the CPU to enter C6 or C7 while holding the punit semaphore Hans de Goede
2017-01-24  9:51   ` Andy Shevchenko
2017-01-24 16:48     ` Hans de Goede
2017-01-23 21:09 ` [PATCH v2 07/13] i2c: designware-baytrail: Fix race when resetting the semaphore Hans de Goede
2017-01-23 21:09 ` [PATCH v2 08/13] i2c: designware-baytrail: Add support for cherrytrail Hans de Goede
2017-01-23 21:09 ` [PATCH v2 09/13] i2c: designware-baytrail: Acquire P-Unit access on bus acquire Hans de Goede
2017-01-27 11:29   ` Jarkko Nikula
2017-01-23 21:09 ` [PATCH v2 10/13] i2c: designware-baytrail: Call pmic_bus_access_notifier_chain Hans de Goede
2017-01-27 11:35   ` Jarkko Nikula
2017-01-23 21:09 ` [PATCH v2 11/13] drm/i915: Add intel_uncore_suspend / resume functions Hans de Goede
2017-01-27 13:45   ` Ville Syrjälä
2017-01-28 16:05     ` Hans de Goede
2017-01-23 21:09 ` [PATCH v2 12/13] drm/i915: Listen for PMIC bus access notifications Hans de Goede
2017-01-27 13:52   ` Ville Syrjälä
2017-01-28 17:39     ` Hans de Goede
2017-01-23 21:09 ` [PATCH v2 13/13] drm/i915: Acquire P-Unit access when modifying P-Unit settings Hans de Goede
2017-01-27 13:51   ` Ville Syrjälä
2017-01-28 16:25     ` Hans de Goede
2017-01-28 17:18       ` Hans de Goede
2017-01-30 13:10         ` Ville Syrjälä
2017-01-30 15:02           ` Hans de Goede
2017-01-30 15:11             ` Ville Syrjälä
2017-01-30 15:27               ` Hans de Goede [this message]
2017-01-30 15:38                 ` Ville Syrjälä
2017-01-30 16:33                   ` Hans de Goede
2017-02-10 10:19               ` Hans de Goede
2017-01-25 20:18 ` [PATCH v2 00/13] coordinate cht i2c-pmic and i915-punit accesses Wolfram Sang

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=dd41cb48-dfa1-e3d9-7850-7886f029944e@redhat.com \
    --to=hdegoede@redhat.com \
    --cc=andriy.shevchenko@linux.intel.com \
    --cc=daniel.vetter@intel.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=hpa@zytor.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=jarkko.nikula@linux.intel.com \
    --cc=lenb@kernel.org \
    --cc=linux-i2c@vger.kernel.org \
    --cc=mika.westerberg@linux.intel.com \
    --cc=russianneuromancer@ya.ru \
    --cc=tglx@linutronix.de \
    --cc=ville.syrjala@linux.intel.com \
    --cc=wsa@the-dreams.de \
    /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.