Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1575674

[PATCH 4.9 66/66] drm/i915/execlists: Reset RING registers upon resume

From Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Newsgroups linux.kernel
Subject [PATCH 4.9 66/66] drm/i915/execlists: Reset RING registers upon resume
Date 2017-02-07 14:10 +0100
Message-ID <t8dGW-3hn-29@gated-at.bofh.it> (permalink)
References <t8dxf-2Y8-5@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw

4.9-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Chris Wilson <chris@chris-wilson.co.uk>

commit bafb2f7d4755bf1571bd5e9a03b97f3fc4fe69ae upstream.

There is a disparity in the context image saved to disk and our own
bookkeeping - that is we presume the RING_HEAD and RING_TAIL match our
stored ce->ring->tail value. However, as we emit WA_TAIL_DWORDS into the
ring but may not tell the GPU about them, the GPU may be lagging behind
our bookkeeping. Upon hibernation we do not save stolen pages, presuming
that their contents are volatile. This means that although we start
writing into the ring at tail, the GPU starts executing from its HEAD
and there may be some garbage in between and so the GPU promptly hangs
upon resume.

Testcase: igt/gem_exec_suspend/basic-S4
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96526
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160921135108.29574-3-chris@chris-wilson.co.uk
Cc: Eric Blau <eblau1@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/gpu/drm/i915/intel_lrc.c |   58 +++++++++++++++++++++++----------------
 1 file changed, 35 insertions(+), 23 deletions(-)

--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -2152,30 +2152,42 @@ error_deref_obj:
 
 void intel_lr_context_resume(struct drm_i915_private *dev_priv)
 {
-	struct i915_gem_context *ctx = dev_priv->kernel_context;
 	struct intel_engine_cs *engine;
+	struct i915_gem_context *ctx;
 
-	for_each_engine(engine, dev_priv) {
-		struct intel_context *ce = &ctx->engine[engine->id];
-		void *vaddr;
-		uint32_t *reg_state;
-
-		if (!ce->state)
-			continue;
-
-		vaddr = i915_gem_object_pin_map(ce->state->obj, I915_MAP_WB);
-		if (WARN_ON(IS_ERR(vaddr)))
-			continue;
-
-		reg_state = vaddr + LRC_STATE_PN * PAGE_SIZE;
-
-		reg_state[CTX_RING_HEAD+1] = 0;
-		reg_state[CTX_RING_TAIL+1] = 0;
-
-		ce->state->obj->dirty = true;
-		i915_gem_object_unpin_map(ce->state->obj);
-
-		ce->ring->head = 0;
-		ce->ring->tail = 0;
+	/* Because we emit WA_TAIL_DWORDS there may be a disparity
+	 * between our bookkeeping in ce->ring->head and ce->ring->tail and
+	 * that stored in context. As we only write new commands from
+	 * ce->ring->tail onwards, everything before that is junk. If the GPU
+	 * starts reading from its RING_HEAD from the context, it may try to
+	 * execute that junk and die.
+	 *
+	 * So to avoid that we reset the context images upon resume. For
+	 * simplicity, we just zero everything out.
+	 */
+	list_for_each_entry(ctx, &dev_priv->context_list, link) {
+		for_each_engine(engine, dev_priv) {
+			struct intel_context *ce = &ctx->engine[engine->id];
+			u32 *reg;
+
+			if (!ce->state)
+				continue;
+
+			reg = i915_gem_object_pin_map(ce->state->obj,
+						      I915_MAP_WB);
+			if (WARN_ON(IS_ERR(reg)))
+				continue;
+
+			reg += LRC_STATE_PN * PAGE_SIZE / sizeof(*reg);
+			reg[CTX_RING_HEAD+1] = 0;
+			reg[CTX_RING_TAIL+1] = 0;
+
+			ce->state->obj->dirty = true;
+			i915_gem_object_unpin_map(ce->state->obj);
+
+			ce->ring->head = ce->ring->tail = 0;
+			ce->ring->last_retired_head = -1;
+			intel_ring_update_space(ce->ring);
+		}
 	}
 }

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[PATCH 4.9 00/66] 4.9.9-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:00 +0100
  [PATCH 4.9 01/66] PCI/ASPM: Handle PCI-to-PCIe bridges as roots of PCIe hierarchies Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:00 +0100
  [PATCH 4.9 50/66] USB: serial: qcserial: add Dell DW5570 QDL Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 45/66] HID: wacom: Fix poor prox handling in wacom_pl_irq Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 58/66] iio: health: afe4404: retrieve a valid iio_dev in suspend/resume Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 40/66] pinctrl: intel: merrifield: Add missed check in mrfld_config_set() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 59/66] iio: health: afe4403: retrieve a valid iio_dev in suspend/resume Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 25/66] zswap: disable changing params if init fails Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 05/66] xtensa: fix noMMU build on cores with MMU Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 24/66] svcrpc: fix oops in absence of krb5 module Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 63/66] x86/irq: Make irq activate operations symmetric Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 47/66] dmaengine: cppi41: Fix runtime PM timeouts with USB mass storage Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 07/66] HID: cp2112: fix gpio-callback error handling Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 62/66] irqdomain: Avoid activating interrupts more than once Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 48/66] dmaengine: cppi41: Fix oops in cppi41_runtime_resume Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 66/66] drm/i915/execlists: Reset RING registers upon resume Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 20/66] powerpc: Add missing error check to prom_find_boot_cpu() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 21/66] powerpc: Fix build failure with clang due to BUILD_BUG_ON() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 65/66] fs: break out of iomap_file_buffered_write on fatal signals Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 42/66] iwlwifi: mvm: avoid crash on restart w/o reserved queues Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 28/66] base/memory, hotplug: fix a kernel oops in show_valid_zones() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 61/66] iio: health: max30100: fixed parenthesis around FIFO count check Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 64/66] iw_cxgb4: set correct FetchBurstMax for QPs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 57/66] iio: adc: palmas_gpadc: retrieve a valid iio_dev in suspend/resume Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 54/66] usb: gadget: f_fs: Assorted buffer overflow checks. Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 60/66] iio: dht11: Use usleep_range instead of msleep for start signal Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 32/66] cgroup: dont online subsystems before cgroup_name/path() are operational Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 12/66] crypto: api - Clear CRYPTO_ALG_DEAD bit before registering an alg Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 10/66] drm/nouveau/disp/gt215: Fix HDA ELD handling (thus, HDMI audio) on gt215 Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 30/66] tracing: Fix hwlat kthread migration Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 39/66] pinctrl: baytrail: Debounce register is one per community Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 18/66] libata: Fix ATA request sense Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 15/66] perf/core: Fix PERF_RECORD_MMAP2 prot/flags for anonymous memory Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 23/66] NFSD: Fix a null reference case in find_or_create_lock_stateid() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 56/66] staging: greybus: timesync: validate platform state callback Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 31/66] can: bcm: fix hrtimer/tasklet termination in bcm op removal Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 13/66] crypto: arm64/aes-blk - honour iv_out requirement in CBC and CTR modes Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 29/66] mm, fs: check for fatal signals in do_generic_file_read() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 26/66] cifs: initialize file_info_lock Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 44/66] HID: hid-lg: Fix immediate disconnection of Logitech Rumblepad 2 Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 16/66] ata: sata_mv:- Handle return value of devm_ioremap. Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 08/66] pinctrl: baytrail: Add missing spinlock usage in byt_gpio_irq_handler Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 19/66] powerpc/eeh: Fix wrong flag passed to eeh_unfreeze_pe() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 35/66] regulator: axp20x: AXP806: Fix dcdcb being set instead of dcdce Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 27/66] mm/memory_hotplug.c: check start_pfn in test_pages_in_a_zone() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 38/66] Revert "vring: Force use of DMA API for ARM-based systems with legacy devices" Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 51/66] USB: serial: pl2303: add ATEN device ID Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 53/66] usb: musb: Fix host mode error -71 regression Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:10 +0100
  [PATCH 4.9 33/66] mmc: sdhci: Ignore unexpected CARD_INT interrupts Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:20 +0100
  [PATCH 4.9 06/66] HID: cp2112: fix sleep-while-atomic Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:20 +0100
  [PATCH 4.9 22/66] powerpc/mm: Use the correct pointer when setting a 2MB pte Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:20 +0100
  [PATCH 4.9 04/66] efi/fdt: Avoid FDT manipulation after ExitBootServices() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:20 +0100
  [PATCH 4.9 36/66] percpu-refcount: fix reference leak during percpu-atomic transition Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:20 +0100
  [PATCH 4.9 03/66] x86/efi: Always map the first physical page into the EFI pagetables Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:20 +0100
  [PATCH 4.9 43/66] HID: usbhid: Quirk a AMI virtual mouse and keyboard with ALWAYS_POLL Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:20 +0100
  [PATCH 4.9 34/66] vhost: fix initialization for vq->is_le Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:20 +0100
  [PATCH 4.9 46/66] perf/x86/intel/uncore: Clean up hotplug conversion fallout Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:20 +0100
  [PATCH 4.9 02/66] ext4: validate s_first_meta_bg at mount time Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:20 +0100
  [PATCH 4.9 14/66] perf/core: Fix use-after-free bug Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:30 +0100
  [PATCH 4.9 11/66] drm/nouveau/nv1a,nv1f/disp: fix memory clock rate retrieval Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 14:30 +0100
  Re: [PATCH 4.9 00/66] 4.9.9-stable review Shuah Khan <shuahkh@osg.samsung.com> - 2017-02-07 17:10 +0100
    Re: [PATCH 4.9 00/66] 4.9.9-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-07 17:20 +0100
  Re: [PATCH 4.9 00/66] 4.9.9-stable review Guenter Roeck <linux@roeck-us.net> - 2017-02-07 22:50 +0100
    Re: [PATCH 4.9 00/66] 4.9.9-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-08 07:40 +0100
  Re: [PATCH 4.9 00/66] 4.9.9-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-08 07:40 +0100

csiph-web