Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Matthew Auld <matthew.auld@intel.com>
To: intel-xe@lists.freedesktop.org
Cc: Andrzej Hajda <andrzej.hajda@intel.com>,
	Rodrigo Vivi <rodrigo.vivi@intel.com>,
	Jagmeet Randhawa <jagmeet.randhawa@intel.com>
Subject: [CI v2 03/18] drm/xe/ggtt: use drm_dev_enter to mark device section
Date: Mon, 20 May 2024 16:42:52 +0100	[thread overview]
Message-ID: <20240520154249.52888-21-matthew.auld@intel.com> (raw)
In-Reply-To: <20240520154249.52888-19-matthew.auld@intel.com>

Device can be hotunplugged before we start destroying gem objects. In
such a case don't touch the GGTT entries, trigger any invalidations or
mess around with rpm.  This should already be taken care of when
removing the device, we just need to take care of dealing with the
software state, like removing the mm node.

v2: (Andrzej)
  - Avoid some duplication by tracking the bound status and checking
    that instead.

References: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/1717
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Andrzej Hajda <andrzej.hajda@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Jagmeet Randhawa <jagmeet.randhawa@intel.com>
---
 drivers/gpu/drm/xe/xe_ggtt.c | 18 +++++++++++++++---
 1 file changed, 15 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_ggtt.c b/drivers/gpu/drm/xe/xe_ggtt.c
index 0d541f55b4fc..17e5066763db 100644
--- a/drivers/gpu/drm/xe/xe_ggtt.c
+++ b/drivers/gpu/drm/xe/xe_ggtt.c
@@ -8,6 +8,7 @@
 #include <linux/io-64-nonatomic-lo-hi.h>
 #include <linux/sizes.h>
 
+#include <drm/drm_drv.h>
 #include <drm/drm_managed.h>
 #include <drm/i915_drm.h>
 
@@ -433,18 +434,29 @@ int xe_ggtt_insert_bo(struct xe_ggtt *ggtt, struct xe_bo *bo)
 void xe_ggtt_remove_node(struct xe_ggtt *ggtt, struct drm_mm_node *node,
 			 bool invalidate)
 {
-	xe_pm_runtime_get_noresume(tile_to_xe(ggtt->tile));
+	struct xe_device *xe = tile_to_xe(ggtt->tile);
+	bool bound;
+	int idx;
+
+	bound = drm_dev_enter(&xe->drm, &idx);
+	if (bound)
+		xe_pm_runtime_get_noresume(xe);
 
 	mutex_lock(&ggtt->lock);
-	xe_ggtt_clear(ggtt, node->start, node->size);
+	if (bound)
+		xe_ggtt_clear(ggtt, node->start, node->size);
 	drm_mm_remove_node(node);
 	node->size = 0;
 	mutex_unlock(&ggtt->lock);
 
+	if (!bound)
+		return;
+
 	if (invalidate)
 		xe_ggtt_invalidate(ggtt);
 
-	xe_pm_runtime_put(tile_to_xe(ggtt->tile));
+	xe_pm_runtime_put(xe);
+	drm_dev_exit(idx);
 }
 
 void xe_ggtt_remove_bo(struct xe_ggtt *ggtt, struct xe_bo *bo)
-- 
2.45.1


  parent reply	other threads:[~2024-05-20 15:45 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-05-20 15:42 [CI v2 01/18] drm/xe/pci: remove broken driver_release Matthew Auld
2024-05-20 15:42 ` [CI v2 02/18] drm/xe: covert sysfs over to devm Matthew Auld
2024-05-20 15:42 ` Matthew Auld [this message]
2024-05-20 15:42 ` [CI v2 04/18] drm/xe/guc: move guc_fini " Matthew Auld
2024-05-20 15:42 ` [CI v2 05/18] drm/xe/guc: s/guc_fini/guc_fini_hw/ Matthew Auld
2024-05-20 15:42 ` [CI v2 06/18] drm/xe/guc_pc: move pc_fini to devm Matthew Auld
2024-05-20 15:42 ` [CI v2 07/18] drm/xe/guc_pc: s/pc_fini/pc_fini_hw/ Matthew Auld
2024-05-20 15:42 ` [CI v2 08/18] drm/xe/irq: move irq_uninstall over to devm Matthew Auld
2024-05-20 15:42 ` [CI v2 09/18] drm/xe/device: move flr " Matthew Auld
2024-05-20 15:42 ` [CI v2 10/18] drm/xe/device: move xe_device_sanitize over " Matthew Auld
2024-05-20 15:43 ` [CI v2 11/18] drm/xe/coredump: move " Matthew Auld
2024-05-20 15:43 ` [CI v2 12/18] drm/xe/gt: break out gt_fini into sw vs hw state Matthew Auld
2024-05-20 15:43 ` [CI v2 13/18] drm/xe: make gt_remove use devm Matthew Auld
2024-05-20 15:43 ` [CI v2 14/18] drm/xe/mmio: move mmio_fini over to devm Matthew Auld
2024-05-20 15:43 ` [CI v2 15/18] drm/xe: reset mmio mappings with devm Matthew Auld
2024-05-20 15:43 ` [CI v2 16/18] drm/xe/display: move display fini stuff to devm Matthew Auld
2024-05-20 15:43 ` [CI v2 17/18] drm/xe/display: stop calling domains_driver_remove twice Matthew Auld
2024-05-20 15:43 ` [CI v2 18/18] drm/xe/display: move device_remove over to drmm Matthew Auld
2024-05-20 15:51 ` ✓ CI.Patch_applied: success for series starting with [CI,v2,01/18] drm/xe/pci: remove broken driver_release Patchwork
2024-05-20 15:51 ` ✗ CI.checkpatch: warning " Patchwork
2024-05-20 15:52 ` ✓ CI.KUnit: success " Patchwork
2024-05-20 16:04 ` ✓ CI.Build: " Patchwork
2024-05-20 16:06 ` ✗ CI.Hooks: failure " Patchwork
2024-05-20 16:08 ` ✓ CI.checksparse: success " Patchwork
2024-05-20 16:30 ` ✓ CI.BAT: " Patchwork
2024-05-20 18:18 ` ✓ CI.FULL: " Patchwork

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=20240520154249.52888-21-matthew.auld@intel.com \
    --to=matthew.auld@intel.com \
    --cc=andrzej.hajda@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=jagmeet.randhawa@intel.com \
    --cc=rodrigo.vivi@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