All of lore.kernel.org
 help / color / mirror / Atom feed
From: ville.syrjala@linux.intel.com
To: intel-gfx@lists.freedesktop.org
Subject: [PATCH 06/15] drm/i915: Document CxSR
Date: Fri, 21 Apr 2017 21:14:23 +0300	[thread overview]
Message-ID: <20170421181432.15216-7-ville.syrjala@linux.intel.com> (raw)
In-Reply-To: <20170421181432.15216-1-ville.syrjala@linux.intel.com>

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Add some documentation explaining what CxSR actually is.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_pm.c | 37 +++++++++++++++++++++++++++++++++++++
 1 file changed, 37 insertions(+)

diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index da2072728df2..7bd4c8688acb 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -386,6 +386,43 @@ static bool _intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enabl
 	return was_enabled;
 }
 
+/**
+ * intel_set_memory_cxsr - Configure CxSR state
+ * @dev_priv: i915 device
+ * @enable: Allow vs. disallow CxSR
+ *
+ * Allow or disallow the system to enter a special CxSR
+ * (C-state self refresh) state. What typically happens in CxSR mode
+ * is that several display FIFOs may get combined into a single larger
+ * FIFO for a particular plane (so called max FIFO mode) to allow the
+ * system to defer memory fetches longer, and the memory will enter
+ * self refresh.
+ *
+ * Note that enabling CxSR does not guarantee that the system enter
+ * this special mode, nor does it guarantee that the system stays
+ * in that mode once entered. So this just allows/disallows the system
+ * to autonomously utilize the CxSR mode. Other factors such as core
+ * C-states will affect when/if the system actually enters/exits the
+ * CxSR mode.
+ *
+ * Note that on VLV/CHV this actually only controls the max FIFO mode,
+ * and the system is free to enter/exit memory self refresh at any time
+ * even when the use of CxSR has been disallowed.
+ *
+ * While the system is actually in the CxSR/max FIFO mode, some plane
+ * control registers will not get latched on vblank. Thus in order to
+ * guarantee the system will respond to changes in the plane registers
+ * we must always disallow CxSR prior to making changes to those registers.
+ * Unfortunately the system will re-evaluate the CxSR conditions at
+ * frame start which happens after vblank start (which is when the plane
+ * registers would get latched), so we can't proceed with the plane update
+ * during the same frame where we disallowed CxSR.
+ *
+ * Certain platforms also have a deeper HPLL SR mode. Fortunately the
+ * HPLL SR mode depends on CxSR itself, so we don't have to hand hold
+ * the hardware w.r.t. HPLL SR when writing to plane registers.
+ * Disallowing just CxSR is sufficient.
+ */
 bool intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enable)
 {
 	bool ret;
-- 
2.10.2

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  parent reply	other threads:[~2017-04-21 18:14 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-04-21 18:14 [PATCH 00/15] drm/i915: Two stage watermarks for g4x ville.syrjala
2017-04-21 18:14 ` [PATCH 01/15] drm/i915: s/vlv_plane_wm_compute/vlv_raw_plane_wm_compute/ etc ville.syrjala
2017-04-21 18:14 ` [PATCH 02/15] drm/i915: Drop the debug message from vlv_get_fifo_size() ville.syrjala
2017-04-21 18:14 ` [PATCH 03/15] drm/i915: s/vlv_num_wm_levels/intel_wm_num_levels/ ville.syrjala
2017-04-21 18:14 ` [PATCH 04/15] drm/i915: Rename bunch of vlv_ watermark structures to g4x_ ville.syrjala
2017-04-21 18:14 ` [PATCH 05/15] drm/i915: Make vlv/chv watermark debug print less cryptic ville.syrjala
2017-04-21 18:14 ` ville.syrjala [this message]
2017-04-21 18:14 ` [PATCH 07/15] drm/i915: Fix cursor 'cpp' in watermark calculatins for old platforms ville.syrjala
2017-04-21 18:14 ` [PATCH 08/15] drm/i915: Fix the g4x watermark TLB miss workaround ville.syrjala
2017-04-21 18:14 ` [PATCH 09/15] drm/i915: Refactor the g4x TLB miss w/a to a helper ville.syrjala
2017-04-21 18:14 ` [PATCH 10/15] drm/i915: Refactor wm calculations ville.syrjala
2017-04-21 18:14 ` [PATCH 11/15] drm/i915: Apply the g4x TLB miss w/a to SR watermarks as well ville.syrjala
2017-04-21 18:14 ` [PATCH 12/15] drm/i915: Two stage watermarks for g4x ville.syrjala
2017-04-24  7:34   ` Maarten Lankhorst
2017-04-24 13:16     ` Ville Syrjälä
2017-05-10 16:40     ` Ville Syrjälä
2017-04-21 18:14 ` [PATCH 13/15] drm/i915: Enable HPLL watermarks on g4x ville.syrjala
2017-04-21 18:14 ` [PATCH 14/15] drm/i915: Add g4x watermark tracepoint ville.syrjala
2017-04-21 18:14 ` [PATCH 15/15] drm/i915: Add support for sprites on g4x ville.syrjala
2017-04-21 18:36 ` ✓ Fi.CI.BAT: success for drm/i915: Two stage watermarks for g4x Patchwork
2017-04-24 14:20 ` [PATCH 00/15] " Lofstedt, Marta

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=20170421181432.15216-7-ville.syrjala@linux.intel.com \
    --to=ville.syrjala@linux.intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    /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.