mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: alexander.levin@verizon.com
To: "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"stable@vger.kernel.org" <stable@vger.kernel.org>
Cc: "Ville Syrjälä" <ville.syrjala@linux.intel.com>,
	"Maarten Lankhorst" <maarten.lankhorst@linux.intel.com>,
	alexander.levin@verizon.com
Subject: [PATCH AUTOSEL for 4.9 41/56] drm/i915: Protect dev_priv->atomic_cdclk_freq with all the crtc locks
Date: Wed, 15 Nov 2017 02:45:44 +0000	[thread overview]
Message-ID: <20171115024521.5884-41-alexander.levin@verizon.com> (raw)
In-Reply-To: <20171115024521.5884-1-alexander.levin@verizon.com>

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

[ Upstream commit 8d96561a0a31974fec33bd3f3853d6211b7ab738 ]

A modeset on one pipe can update dev_priv->atomic_cdclk_freq without
actually touching the hardware, in which case we won't force a modeset
on all the pipes, and thus won't lock any of the other pipes either.
That means a parallel plane update on another pipe could be looking at
a stale dev_priv->atomic_cdcdlk_freq and thus fail to notice when the
plane configuration is invalid, or potentially reject a valid update.

To overcome this we must protect writes to atomic_cdclk_freq with
all the crtc locks, and thus for reads any single crtc lock will
be sufficient protection.

Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1479141311-11904-3-git-send-email-ville.syrjala@linux.intel.com
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
---
 drivers/gpu/drm/i915/i915_drv.h      |  9 +++++++-
 drivers/gpu/drm/i915/intel_display.c | 41 +++++++++++++++++++++++++++++++-----
 2 files changed, 44 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index d120797aec17..b9e759ba16c2 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1841,7 +1841,14 @@ struct drm_i915_private {
 
 	unsigned int fsb_freq, mem_freq, is_ddr3;
 	unsigned int skl_preferred_vco_freq;
-	unsigned int cdclk_freq, max_cdclk_freq, atomic_cdclk_freq;
+	unsigned int cdclk_freq, max_cdclk_freq;
+
+	/*
+	 * For reading holding any crtc lock is sufficient,
+	 * for writing must hold all of them.
+	 */
+	unsigned int atomic_cdclk_freq;
+
 	unsigned int max_dotclk_freq;
 	unsigned int rawclk_freq;
 	unsigned int hpll_freq;
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index ce32303b3013..bd560a8792ba 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -13894,13 +13894,32 @@ static int haswell_mode_set_planes_workaround(struct drm_atomic_state *state)
 	return 0;
 }
 
+static int intel_lock_all_pipes(struct drm_atomic_state *state)
+{
+	struct drm_crtc *crtc;
+
+	/* Add all pipes to the state */
+	for_each_crtc(state->dev, crtc) {
+		struct drm_crtc_state *crtc_state;
+
+		crtc_state = drm_atomic_get_crtc_state(state, crtc);
+		if (IS_ERR(crtc_state))
+			return PTR_ERR(crtc_state);
+	}
+
+	return 0;
+}
+
 static int intel_modeset_all_pipes(struct drm_atomic_state *state)
 {
 	struct drm_crtc *crtc;
 	struct drm_crtc_state *crtc_state;
 	int ret = 0;
 
-	/* add all active pipes to the state */
+	/*
+	 * Add all pipes to the state, and force
+	 * a modeset on all the active ones.
+	 */
 	for_each_crtc(state->dev, crtc) {
 		crtc_state = drm_atomic_get_crtc_state(state, crtc);
 		if (IS_ERR(crtc_state))
@@ -13966,12 +13985,24 @@ static int intel_modeset_checks(struct drm_atomic_state *state)
 		if (ret < 0)
 			return ret;
 
+		/*
+		 * Writes to dev_priv->atomic_cdclk_freq must protected by
+		 * holding all the crtc locks, even if we don't end up
+		 * touching the hardware
+		 */
+		if (intel_state->cdclk != dev_priv->atomic_cdclk_freq) {
+			ret = intel_lock_all_pipes(state);
+			if (ret < 0)
+				return ret;
+		}
+
+		/* All pipes must be switched off while we change the cdclk. */
 		if (intel_state->dev_cdclk != dev_priv->cdclk_freq ||
-		    intel_state->cdclk_pll_vco != dev_priv->cdclk_pll.vco)
+		    intel_state->cdclk_pll_vco != dev_priv->cdclk_pll.vco) {
 			ret = intel_modeset_all_pipes(state);
-
-		if (ret < 0)
-			return ret;
+			if (ret < 0)
+				return ret;
+		}
 
 		DRM_DEBUG_KMS("New cdclk calculated to be atomic %u, actual %u\n",
 			      intel_state->cdclk, intel_state->dev_cdclk);
-- 
2.11.0

  parent reply	other threads:[~2017-11-15  3:08 UTC|newest]

Thread overview: 83+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-11-15  2:45 [PATCH AUTOSEL for 4.9 01/56] ACPICA: Resources: Not a valid resource if buffer length too long alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 04/56] RDS: RDMA: fix the ib_map_mr_sg_zbva() argument alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 02/56] RDS: make message size limit compliant with spec alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 05/56] PCI: Apply _HPX settings only to relevant devices alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 03/56] RDS: RDMA: return appropriate error on rdma map failures alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 08/56] clk: sunxi-ng: fix PLL_CPUX adjusting on A33 alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 07/56] clk: sunxi-ng: A31: Fix spdif clock register alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 06/56] drm/sun4i: Fix a return value in case of error alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 12/56] net: Allow IP_MULTICAST_IF to set index to L3 slave alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 11/56] fscrypt: use ENOTDIR when setting encryption policy on nondirectory alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 10/56] fscrypt: use ENOKEY when file cannot be created w/o key alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 09/56] dmaengine: zx: set DMA_CYCLIC cap_mask bit alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 13/56] net: 3com: typhoon: typhoon_init_one: make return values more specific alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 15/56] drm/armada: Fix compile fail alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 17/56] adm80211: return an error if adm8211_alloc_rings() fails alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 14/56] net: 3com: typhoon: typhoon_init_one: fix incorrect return values alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 16/56] rt2800: set minimum MPDU and PSDU lengths to sane values alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 18/56] mwifiex: sdio: fix use after free issue for save_adapter alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 20/56] ath10k: ignore configuring the incorrect board_id alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 19/56] ath10k: fix incorrect txpower set by P2P_DEVICE interface alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 21/56] ath10k: fix potential memory leak in ath10k_wmi_tlv_op_pull_fw_stats() alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 24/56] ath10k: set CTS protection VDEV param only if VDEV is up alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 23/56] bnxt_en: Set default completion ring for async events alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 25/56] ALSA: hda - Apply ALC269_FIXUP_NO_SHUTUP on HDA_FIXUP_ACT_PROBE alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 22/56] pinctrl: sirf: atlas7: Add missing 'of_node_put()' alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 28/56] drm/915: Parsing the missed out DTD fields from the VBT alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 29/56] clk: qcom: ipq4019: Add all the frequencies for apss cpu alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 27/56] drm: Apply range restriction after color adjustment when allocation alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 26/56] gpio: mockup: dynamically allocate memory for chip name alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 33/56] mac80211: Remove invalid flag operations in mesh TSF synchronization alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 32/56] drm/mediatek: don't use drm_put_dev alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 30/56] drm/atomic: Unconditionally call prepare_fb alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 31/56] ASoC: da7218: Set DAI output pin high impedance when not in use alexander.levin
2017-11-15 12:19   ` Mark Brown
2017-11-15 16:43     ` alexander.levin
2017-11-15 16:55       ` Adam Thomson
2017-11-16 10:46         ` Mark Brown
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 37/56] adm80211: add checks for dma mapping errors alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 35/56] drm/i915: Protect DSPARB registers with a spinlock alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 34/56] mac80211: Suppress NEW_PEER_CANDIDATE event if no room alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 36/56] drm/i915: Fix the level 0 max_wm hack on VLV/CHV alexander.levin
2017-11-15 11:08   ` Ville Syrjälä
2017-11-15 16:44     ` alexander.levin
2017-11-15 17:03       ` Ville Syrjälä
2017-11-17 11:28         ` Jani Nikula
2017-11-17 12:41           ` Greg KH
2017-11-17 12:53             ` Ville Syrjälä
2017-11-17 12:59               ` Greg KH
2017-11-17 13:13                 ` [Intel-gfx] " Emil Velikov
2017-11-17 13:58                   ` Greg KH
2017-11-17 13:01             ` Jani Nikula
2017-11-17 13:57               ` Greg KH
2017-11-18  3:15               ` alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 38/56] iio: light: fix improper return value alexander.levin
2017-11-15  2:45 ` alexander.levin [this message]
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 40/56] drm/i915: Assert no external observers when unwind a failed request alloc alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 39/56] staging: iio: cdc: fix improper return value alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 42/56] drm/i915/bxt: Correct dual-link MIPI port control alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 44/56] ASoC: dwc: Fix PIO mode initialization alexander.levin
2017-11-15 10:54   ` Mark Brown
2017-11-15 12:46     ` Jose Abreu
2017-11-15 13:20       ` Mark Brown
2017-11-15 14:06         ` Jose Abreu
2017-11-15 16:41           ` alexander.levin
2017-11-16 10:51             ` Mark Brown
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 45/56] netfilter: nft_queue: use raw_smp_processor_id() alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 43/56] spi: SPI_FSL_DSPI should depend on HAS_DMA alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 49/56] btrfs: return the actual error value from from btrfs_uuid_tree_iterate alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 47/56] ASoC: rsnd: don't double free kctrl alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 46/56] netfilter: nf_tables: fix oob access alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 48/56] crypto: marvell - Copy IVDIG before launching partial DMA ahash requests alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 50/56] ASoC: wm_adsp: Don't overrun firmware file buffer when reading region data alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 51/56] s390/kbuild: enable modversions for symbols exported from asm alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 53/56] [media] cec: CEC_MSG_GIVE_FEATURES should abort for CEC version < 2 alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 52/56] [media] cec: when canceling a message, don't overwrite old status info alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 56/56] xen: xenbus driver must not accept invalid transaction ids alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 54/56] [media] cec: update log_addr[] before finishing configuration alexander.levin
2017-11-15  2:45 ` [PATCH AUTOSEL for 4.9 55/56] nvmet: fix KATO offset in Set Features alexander.levin
     [not found] ` <94F2FBAB4432B54E8AACC7DFDE6C92E37E619757@ORSMSX110.amr.corp.intel.com>
2017-11-15 16:39   ` [PATCH AUTOSEL for 4.9 01/56] ACPICA: Resources: Not a valid resource if buffer length too long alexander.levin
2017-11-15 17:05     ` Schmauss, Erik
2017-11-17  0:56       ` alexander.levin
2017-11-20 17:22         ` Schmauss, Erik
2017-11-22 23:24           ` alexander.levin

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=20171115024521.5884-41-alexander.levin@verizon.com \
    --to=alexander.levin@verizon.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=maarten.lankhorst@linux.intel.com \
    --cc=stable@vger.kernel.org \
    --cc=ville.syrjala@linux.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

all inboxes | Powered by JetHome®