Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Imre Deak <imre.deak@intel.com>
To: intel-gfx@lists.freedesktop.org
Cc: "Jouni Högander" <jouni.hogander@intel.com>
Subject: [CI 2/2] drm/i915: Prevent HW access during init from connector get_modes hooks
Date: Mon, 12 Feb 2024 19:52:37 +0200	[thread overview]
Message-ID: <20240212175237.2625812-2-imre.deak@intel.com> (raw)
In-Reply-To: <20240212175237.2625812-1-imre.deak@intel.com>

Prevent accessing the HW from the get_modes hooks of connectors deriving
the mode list from the display's EDID. drm_edid_connector_add_modes()
will return the mode list based on the EDID which was cached during a
previous detection/get_modes call.

This also fixes the NULL deref problem (10085) which was
introduced/revealed by

commit bab87ef4db9a ("drm/i915: Disable hotplug detection handlers during driver init/shutdown")

After the above change MST connectors will not change state during
driver init/shutdown; thus some of these connectors with no I2C/DDC
adapter registered for them (since the given MST port has no sink
connected) may stay then in the 'unknown' connector status. The
get_modes() hook should not try to use the I2C/DDC adapter in this state
(which would lead to the above NULL deref) which this patch ensures.

v2:
- Remove the redundant check from intel_crt_ddc_get_modes().
- Rebase on latest drm-tip.
- Add Fixes: line / related commit notes.

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/10085
Fixes: bab87ef4db9a ("drm/i915: Disable hotplug detection handlers during driver init/shutdown")
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
---
 drivers/gpu/drm/i915/display/intel_crt.c    | 3 +++
 drivers/gpu/drm/i915/display/intel_dp_mst.c | 4 ++++
 drivers/gpu/drm/i915/display/intel_dvo.c    | 5 +++++
 drivers/gpu/drm/i915/display/intel_sdvo.c   | 4 ++++
 4 files changed, 16 insertions(+)

diff --git a/drivers/gpu/drm/i915/display/intel_crt.c b/drivers/gpu/drm/i915/display/intel_crt.c
index b9733a73e21d4..93479db0f89f6 100644
--- a/drivers/gpu/drm/i915/display/intel_crt.c
+++ b/drivers/gpu/drm/i915/display/intel_crt.c
@@ -933,6 +933,9 @@ static int intel_crt_get_modes(struct drm_connector *connector)
 	struct i2c_adapter *ddc;
 	int ret;
 
+	if (!intel_display_driver_check_access(dev_priv))
+		return drm_edid_connector_add_modes(connector);
+
 	wakeref = intel_display_power_get(dev_priv,
 					  intel_encoder->power_domain);
 
diff --git a/drivers/gpu/drm/i915/display/intel_dp_mst.c b/drivers/gpu/drm/i915/display/intel_dp_mst.c
index 5fa25a5a36b55..5307ddd4edcf5 100644
--- a/drivers/gpu/drm/i915/display/intel_dp_mst.c
+++ b/drivers/gpu/drm/i915/display/intel_dp_mst.c
@@ -1197,6 +1197,7 @@ static bool intel_dp_mst_initial_fastset_check(struct intel_encoder *encoder,
 static int intel_dp_mst_get_ddc_modes(struct drm_connector *connector)
 {
 	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct drm_i915_private *i915 = to_i915(intel_connector->base.dev);
 	struct intel_dp *intel_dp = intel_connector->mst_port;
 	const struct drm_edid *drm_edid;
 	int ret;
@@ -1204,6 +1205,9 @@ static int intel_dp_mst_get_ddc_modes(struct drm_connector *connector)
 	if (drm_connector_is_unregistered(connector))
 		return intel_connector_update_modes(connector, NULL);
 
+	if (!intel_display_driver_check_access(i915))
+		return drm_edid_connector_add_modes(connector);
+
 	drm_edid = drm_dp_mst_edid_read(connector, &intel_dp->mst_mgr, intel_connector->port);
 
 	ret = intel_connector_update_modes(connector, drm_edid);
diff --git a/drivers/gpu/drm/i915/display/intel_dvo.c b/drivers/gpu/drm/i915/display/intel_dvo.c
index 8ca9ae4798a89..c076da75b066e 100644
--- a/drivers/gpu/drm/i915/display/intel_dvo.c
+++ b/drivers/gpu/drm/i915/display/intel_dvo.c
@@ -30,6 +30,7 @@
 
 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_crtc.h>
+#include <drm/drm_edid.h>
 
 #include "i915_drv.h"
 #include "i915_reg.h"
@@ -338,8 +339,12 @@ intel_dvo_detect(struct drm_connector *_connector, bool force)
 static int intel_dvo_get_modes(struct drm_connector *_connector)
 {
 	struct intel_connector *connector = to_intel_connector(_connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
 	int num_modes;
 
+	if (!intel_display_driver_check_access(i915))
+		return drm_edid_connector_add_modes(&connector->base);
+
 	/*
 	 * We should probably have an i2c driver get_modes function for those
 	 * devices which will have a fixed set of modes determined by the chip
diff --git a/drivers/gpu/drm/i915/display/intel_sdvo.c b/drivers/gpu/drm/i915/display/intel_sdvo.c
index c67605059aa3e..825638702ac1b 100644
--- a/drivers/gpu/drm/i915/display/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/display/intel_sdvo.c
@@ -2202,12 +2202,16 @@ intel_sdvo_detect(struct drm_connector *connector, bool force)
 
 static int intel_sdvo_get_ddc_modes(struct drm_connector *connector)
 {
+	struct drm_i915_private *i915 = to_i915(connector->dev);
 	int num_modes = 0;
 	const struct drm_edid *drm_edid;
 
 	drm_dbg_kms(connector->dev, "[CONNECTOR:%d:%s]\n",
 		    connector->base.id, connector->name);
 
+	if (!intel_display_driver_check_access(i915))
+		return drm_edid_connector_add_modes(connector);
+
 	/* set the bus switch and get the modes */
 	drm_edid = intel_sdvo_get_edid(connector);
 
-- 
2.39.2


  reply	other threads:[~2024-02-12 17:52 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-02-12 17:52 [CI 1/2] drm/i915: Prevent HW access during init from SDVO TV get_modes hook Imre Deak
2024-02-12 17:52 ` Imre Deak [this message]
2024-02-12 18:53 ` ✗ Fi.CI.CHECKPATCH: warning for series starting with [CI,1/2] " Patchwork
2024-02-12 19:03 ` ✓ Fi.CI.BAT: success " Patchwork
2024-02-12 21:54 ` ✗ Fi.CI.IGT: failure " Patchwork
2024-02-13 11:43   ` Imre Deak

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=20240212175237.2625812-2-imre.deak@intel.com \
    --to=imre.deak@intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=jouni.hogander@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