summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm')
-rw-r--r--drivers/gpu/drm/amd/amdkfd/Makefile3
-rw-r--r--drivers/gpu/drm/amd/amdkfd/kfd_device.c26
-rw-r--r--drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c80
-rw-r--r--drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h1
-rw-r--r--drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c176
-rw-r--r--drivers/gpu/drm/amd/amdkfd/kfd_module.c27
-rw-r--r--drivers/gpu/drm/amd/amdkfd/kfd_pasid.c2
-rw-r--r--drivers/gpu/drm/amd/amdkfd/kfd_priv.h32
-rw-r--r--drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c18
-rw-r--r--drivers/gpu/drm/cirrus/cirrus_drv.c3
-rw-r--r--drivers/gpu/drm/cirrus/cirrus_drv.h3
-rw-r--r--drivers/gpu/drm/cirrus/cirrus_main.c2
-rw-r--r--drivers/gpu/drm/cirrus/cirrus_mode.c9
-rw-r--r--drivers/gpu/drm/drm_fb_helper.c38
-rw-r--r--drivers/gpu/drm/exynos/exynos_drm_drv.c12
-rw-r--r--drivers/gpu/drm/exynos/exynos_hdmi.c8
-rw-r--r--drivers/gpu/drm/exynos/exynos_mixer.c11
-rw-r--r--drivers/gpu/drm/i2c/tda998x_drv.c52
-rw-r--r--drivers/gpu/drm/i915/i915_drv.c14
-rw-r--r--drivers/gpu/drm/i915/i915_drv.h3
-rw-r--r--drivers/gpu/drm/i915/i915_gem.c28
-rw-r--r--drivers/gpu/drm/i915/i915_irq.c20
-rw-r--r--drivers/gpu/drm/i915/intel_display.c2
-rw-r--r--drivers/gpu/drm/i915/intel_drv.h1
-rw-r--r--drivers/gpu/drm/i915/intel_panel.c2
-rw-r--r--drivers/gpu/drm/i915/intel_pm.c14
-rw-r--r--drivers/gpu/drm/radeon/cik.c11
-rw-r--r--drivers/gpu/drm/radeon/cik_sdma.c11
-rw-r--r--drivers/gpu/drm/radeon/ni.c10
-rw-r--r--drivers/gpu/drm/radeon/ni_dma.c7
-rw-r--r--drivers/gpu/drm/radeon/nid.h24
-rw-r--r--drivers/gpu/drm/radeon/r100.c10
-rw-r--r--drivers/gpu/drm/radeon/r300.c16
-rw-r--r--drivers/gpu/drm/radeon/radeon.h9
-rw-r--r--drivers/gpu/drm/radeon/radeon_asic.c42
-rw-r--r--drivers/gpu/drm/radeon/radeon_asic.h12
-rw-r--r--drivers/gpu/drm/radeon/radeon_benchmark.c13
-rw-r--r--drivers/gpu/drm/radeon/radeon_device.c2
-rw-r--r--drivers/gpu/drm/radeon/radeon_display.c3
-rw-r--r--drivers/gpu/drm/radeon/radeon_gart.c54
-rw-r--r--drivers/gpu/drm/radeon/radeon_gem.c8
-rw-r--r--drivers/gpu/drm/radeon/radeon_kfd.c2
-rw-r--r--drivers/gpu/drm/radeon/radeon_kms.c16
-rw-r--r--drivers/gpu/drm/radeon/radeon_pm.c33
-rw-r--r--drivers/gpu/drm/radeon/radeon_test.c8
-rw-r--r--drivers/gpu/drm/radeon/radeon_vm.c12
-rw-r--r--drivers/gpu/drm/radeon/rs400.c14
-rw-r--r--drivers/gpu/drm/radeon/rs600.c14
-rw-r--r--drivers/gpu/drm/radeon/si.c10
-rw-r--r--drivers/gpu/drm/radeon/si_dma.c9
-rw-r--r--drivers/gpu/drm/radeon/si_dpm.c39
-rw-r--r--drivers/gpu/drm/radeon/sid.h18
-rw-r--r--drivers/gpu/drm/vmwgfx/vmwgfx_drv.c28
-rw-r--r--drivers/gpu/drm/vmwgfx/vmwgfx_drv.h25
-rw-r--r--drivers/gpu/drm/vmwgfx/vmwgfx_fence.c18
-rw-r--r--drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c36
-rw-r--r--drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c8
-rw-r--r--drivers/gpu/drm/vmwgfx/vmwgfx_irq.c25
-rw-r--r--drivers/gpu/drm/vmwgfx/vmwgfx_kms.c2
59 files changed, 642 insertions, 494 deletions
diff --git a/drivers/gpu/drm/amd/amdkfd/Makefile b/drivers/gpu/drm/amd/amdkfd/Makefile
index be6246d..307a309 100644
--- a/drivers/gpu/drm/amd/amdkfd/Makefile
+++ b/drivers/gpu/drm/amd/amdkfd/Makefile
@@ -8,7 +8,6 @@ amdkfd-y := kfd_module.o kfd_device.o kfd_chardev.o kfd_topology.o \
kfd_pasid.o kfd_doorbell.o kfd_flat_memory.o \
kfd_process.o kfd_queue.o kfd_mqd_manager.o \
kfd_kernel_queue.o kfd_packet_manager.o \
- kfd_process_queue_manager.o kfd_device_queue_manager.o \
- kfd_interrupt.o
+ kfd_process_queue_manager.o kfd_device_queue_manager.o
obj-$(CONFIG_HSA_AMD) += amdkfd.o
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index 43884eb..25bc47f 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -26,6 +26,7 @@
#include <linux/slab.h>
#include "kfd_priv.h"
#include "kfd_device_queue_manager.h"
+#include "kfd_pm4_headers.h"
#define MQD_SIZE_ALIGNED 768
@@ -169,9 +170,8 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd,
kfd->shared_resources = *gpu_resources;
/* calculate max size of mqds needed for queues */
- size = max_num_of_processes *
- max_num_of_queues_per_process *
- kfd->device_info->mqd_size_aligned;
+ size = max_num_of_queues_per_device *
+ kfd->device_info->mqd_size_aligned;
/* add another 512KB for all other allocations on gart */
size += 512 * 1024;
@@ -192,13 +192,6 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd,
goto kfd_topology_add_device_error;
}
- if (kfd_interrupt_init(kfd)) {
- dev_err(kfd_device,
- "Error initializing interrupts for device (%x:%x)\n",
- kfd->pdev->vendor, kfd->pdev->device);
- goto kfd_interrupt_error;
- }
-
if (!device_iommu_pasid_init(kfd)) {
dev_err(kfd_device,
"Error initializing iommuv2 for device (%x:%x)\n",
@@ -237,8 +230,6 @@ dqm_start_error:
device_queue_manager_error:
amd_iommu_free_device(kfd->pdev);
device_iommu_pasid_error:
- kfd_interrupt_exit(kfd);
-kfd_interrupt_error:
kfd_topology_remove_device(kfd);
kfd_topology_add_device_error:
kfd2kgd->fini_sa_manager(kfd->kgd);
@@ -254,7 +245,6 @@ void kgd2kfd_device_exit(struct kfd_dev *kfd)
if (kfd->init_complete) {
device_queue_manager_uninit(kfd->dqm);
amd_iommu_free_device(kfd->pdev);
- kfd_interrupt_exit(kfd);
kfd_topology_remove_device(kfd);
}
@@ -296,13 +286,5 @@ int kgd2kfd_resume(struct kfd_dev *kfd)
/* This is called directly from KGD at ISR. */
void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry)
{
- if (kfd->init_complete) {
- spin_lock(&kfd->interrupt_lock);
-
- if (kfd->interrupts_active
- && enqueue_ih_ring_entry(kfd, ih_ring_entry))
- schedule_work(&kfd->interrupt_work);
-
- spin_unlock(&kfd->interrupt_lock);
- }
+ /* Process interrupts / schedule work as necessary */
}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
index 9c8961d..0fd5927 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -183,6 +183,13 @@ static int create_queue_nocpsch(struct device_queue_manager *dqm,
mutex_lock(&dqm->lock);
+ if (dqm->total_queue_count >= max_num_of_queues_per_device) {
+ pr_warn("amdkfd: Can't create new usermode queue because %d queues were already created\n",
+ dqm->total_queue_count);
+ mutex_unlock(&dqm->lock);
+ return -EPERM;
+ }
+
if (list_empty(&qpd->queues_list)) {
retval = allocate_vmid(dqm, qpd, q);
if (retval != 0) {
@@ -207,6 +214,14 @@ static int create_queue_nocpsch(struct device_queue_manager *dqm,
list_add(&q->list, &qpd->queues_list);
dqm->queue_count++;
+ /*
+ * Unconditionally increment this counter, regardless of the queue's
+ * type or whether the queue is active.
+ */
+ dqm->total_queue_count++;
+ pr_debug("Total of %d queues are accountable so far\n",
+ dqm->total_queue_count);
+
mutex_unlock(&dqm->lock);
return 0;
}
@@ -280,7 +295,7 @@ static int create_compute_queue_nocpsch(struct device_queue_manager *dqm,
q->queue);
retval = mqd->load_mqd(mqd, q->mqd, q->pipe,
- q->queue, q->properties.write_ptr);
+ q->queue, (uint32_t __user *) q->properties.write_ptr);
if (retval != 0) {
deallocate_hqd(dqm, q);
mqd->uninit_mqd(mqd, q->mqd, q->mqd_mem_obj);
@@ -326,6 +341,15 @@ static int destroy_queue_nocpsch(struct device_queue_manager *dqm,
if (list_empty(&qpd->queues_list))
deallocate_vmid(dqm, qpd, q);
dqm->queue_count--;
+
+ /*
+ * Unconditionally decrement this counter, regardless of the queue's
+ * type
+ */
+ dqm->total_queue_count--;
+ pr_debug("Total of %d queues are accountable so far\n",
+ dqm->total_queue_count);
+
out:
mutex_unlock(&dqm->lock);
return retval;
@@ -541,10 +565,14 @@ static int init_pipelines(struct device_queue_manager *dqm,
for (i = 0; i < pipes_num; i++) {
inx = i + first_pipe;
+ /*
+ * HPD buffer on GTT is allocated by amdkfd, no need to waste
+ * space in GTT for pipelines we don't initialize
+ */
pipe_hpd_addr = dqm->pipelines_addr + i * CIK_HPD_EOP_BYTES;
pr_debug("kfd: pipeline address %llX\n", pipe_hpd_addr);
/* = log2(bytes/4)-1 */
- kfd2kgd->init_pipeline(dqm->dev->kgd, i,
+ kfd2kgd->init_pipeline(dqm->dev->kgd, inx,
CIK_HPD_EOP_BYTES_LOG2 - 3, pipe_hpd_addr);
}
@@ -560,7 +588,7 @@ static int init_scheduler(struct device_queue_manager *dqm)
pr_debug("kfd: In %s\n", __func__);
- retval = init_pipelines(dqm, get_pipes_num(dqm), KFD_DQM_FIRST_PIPE);
+ retval = init_pipelines(dqm, get_pipes_num(dqm), get_first_pipe(dqm));
if (retval != 0)
return retval;
@@ -752,6 +780,21 @@ static int create_kernel_queue_cpsch(struct device_queue_manager *dqm,
pr_debug("kfd: In func %s\n", __func__);
mutex_lock(&dqm->lock);
+ if (dqm->total_queue_count >= max_num_of_queues_per_device) {
+ pr_warn("amdkfd: Can't create new kernel queue because %d queues were already created\n",
+ dqm->total_queue_count);
+ mutex_unlock(&dqm->lock);
+ return -EPERM;
+ }
+
+ /*
+ * Unconditionally increment this counter, regardless of the queue's
+ * type or whether the queue is active.
+ */
+ dqm->total_queue_count++;
+ pr_debug("Total of %d queues are accountable so far\n",
+ dqm->total_queue_count);
+
list_add(&kq->list, &qpd->priv_queue_list);
dqm->queue_count++;
qpd->is_debug = true;
@@ -775,6 +818,13 @@ static void destroy_kernel_queue_cpsch(struct device_queue_manager *dqm,
dqm->queue_count--;
qpd->is_debug = false;
execute_queues_cpsch(dqm, false);
+ /*
+ * Unconditionally decrement this counter, regardless of the queue's
+ * type.
+ */
+ dqm->total_queue_count--;
+ pr_debug("Total of %d queues are accountable so far\n",
+ dqm->total_queue_count);
mutex_unlock(&dqm->lock);
}
@@ -793,6 +843,13 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q,
mutex_lock(&dqm->lock);
+ if (dqm->total_queue_count >= max_num_of_queues_per_device) {
+ pr_warn("amdkfd: Can't create new usermode queue because %d queues were already created\n",
+ dqm->total_queue_count);
+ retval = -EPERM;
+ goto out;
+ }
+
mqd = dqm->get_mqd_manager(dqm, KFD_MQD_TYPE_CIK_CP);
if (mqd == NULL) {
mutex_unlock(&dqm->lock);
@@ -810,6 +867,15 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q,
retval = execute_queues_cpsch(dqm, false);
}
+ /*
+ * Unconditionally increment this counter, regardless of the queue's
+ * type or whether the queue is active.
+ */
+ dqm->total_queue_count++;
+
+ pr_debug("Total of %d queues are accountable so far\n",
+ dqm->total_queue_count);
+
out:
mutex_unlock(&dqm->lock);
return retval;
@@ -930,6 +996,14 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
mqd->uninit_mqd(mqd, q->mqd, q->mqd_mem_obj);
+ /*
+ * Unconditionally decrement this counter, regardless of the queue's
+ * type
+ */
+ dqm->total_queue_count--;
+ pr_debug("Total of %d queues are accountable so far\n",
+ dqm->total_queue_count);
+
mutex_unlock(&dqm->lock);
return 0;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
index c3f189e8..52035bf 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
@@ -130,6 +130,7 @@ struct device_queue_manager {
struct list_head queues;
unsigned int processes_count;
unsigned int queue_count;
+ unsigned int total_queue_count;
unsigned int next_pipe_to_allocate;
unsigned int *allocated_queues;
unsigned int vmid_bitmap;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
deleted file mode 100644
index 5b99909..0000000
--- a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
+++ /dev/null
@@ -1,176 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- */
-
-/*
- * KFD Interrupts.
- *
- * AMD GPUs deliver interrupts by pushing an interrupt description onto the
- * interrupt ring and then sending an interrupt. KGD receives the interrupt
- * in ISR and sends us a pointer to each new entry on the interrupt ring.
- *
- * We generally can't process interrupt-signaled events from ISR, so we call
- * out to each interrupt client module (currently only the scheduler) to ask if
- * each interrupt is interesting. If they return true, then it requires further
- * processing so we copy it to an internal interrupt ring and call each
- * interrupt client again from a work-queue.
- *
- * There's no acknowledgment for the interrupts we use. The hardware simply
- * queues a new interrupt each time without waiting.
- *
- * The fixed-size internal queue means that it's possible for us to lose
- * interrupts because we have no back-pressure to the hardware.
- */
-
-#include <linux/slab.h>
-#include <linux/device.h>
-#include "kfd_priv.h"
-
-#define KFD_INTERRUPT_RING_SIZE 256
-
-static void interrupt_wq(struct work_struct *);
-
-int kfd_interrupt_init(struct kfd_dev *kfd)
-{
- void *interrupt_ring = kmalloc_array(KFD_INTERRUPT_RING_SIZE,
- kfd->device_info->ih_ring_entry_size,
- GFP_KERNEL);
- if (!interrupt_ring)
- return -ENOMEM;
-
- kfd->interrupt_ring = interrupt_ring;
- kfd->interrupt_ring_size =
- KFD_INTERRUPT_RING_SIZE * kfd->device_info->ih_ring_entry_size;
- atomic_set(&kfd->interrupt_ring_wptr, 0);
- atomic_set(&kfd->interrupt_ring_rptr, 0);
-
- spin_lock_init(&kfd->interrupt_lock);
-
- INIT_WORK(&kfd->interrupt_work, interrupt_wq);
-
- kfd->interrupts_active = true;
-
- /*
- * After this function returns, the interrupt will be enabled. This
- * barrier ensures that the interrupt running on a different processor
- * sees all the above writes.
- */
- smp_wmb();
-
- return 0;
-}
-
-void kfd_interrupt_exit(struct kfd_dev *kfd)
-{
- /*
- * Stop the interrupt handler from writing to the ring and scheduling
- * workqueue items. The spinlock ensures that any interrupt running
- * after we have unlocked sees interrupts_active = false.
- */
- unsigned long flags;
-
- spin_lock_irqsave(&kfd->interrupt_lock, flags);
- kfd->interrupts_active = false;
- spin_unlock_irqrestore(&kfd->interrupt_lock, flags);
-
- /*
- * Flush_scheduled_work ensures that there are no outstanding
- * work-queue items that will access interrupt_ring. New work items
- * can't be created because we stopped interrupt handling above.
- */
- flush_scheduled_work();
-
- kfree(kfd->interrupt_ring);
-}
-
-/*
- * This assumes that it can't be called concurrently with itself
- * but only with dequeue_ih_ring_entry.
- */
-bool enqueue_ih_ring_entry(struct kfd_dev *kfd, const void *ih_ring_entry)
-{
- unsigned int rptr = atomic_read(&kfd->interrupt_ring_rptr);
- unsigned int wptr = atomic_read(&kfd->interrupt_ring_wptr);
-
- if ((rptr - wptr) % kfd->interrupt_ring_size ==
- kfd->device_info->ih_ring_entry_size) {
- /* This is very bad, the system is likely to hang. */
- dev_err_ratelimited(kfd_chardev(),
- "Interrupt ring overflow, dropping interrupt.\n");
- return false;
- }
-
- memcpy(kfd->interrupt_ring + wptr, ih_ring_entry,
- kfd->device_info->ih_ring_entry_size);
-
- wptr = (wptr + kfd->device_info->ih_ring_entry_size) %
- kfd->interrupt_ring_size;
- smp_wmb(); /* Ensure memcpy'd data is visible before wptr update. */
- atomic_set(&kfd->interrupt_ring_wptr, wptr);
-
- return true;
-}
-
-/*
- * This assumes that it can't be called concurrently with itself
- * but only with enqueue_ih_ring_entry.
- */
-static bool dequeue_ih_ring_entry(struct kfd_dev *kfd, void *ih_ring_entry)
-{
- /*
- * Assume that wait queues have an implicit barrier, i.e. anything that
- * happened in the ISR before it queued work is visible.
- */
-
- unsigned int wptr = atomic_read(&kfd->interrupt_ring_wptr);
- unsigned int rptr = atomic_read(&kfd->interrupt_ring_rptr);
-
- if (rptr == wptr)
- return false;
-
- memcpy(ih_ring_entry, kfd->interrupt_ring + rptr,
- kfd->device_info->ih_ring_entry_size);
-
- rptr = (rptr + kfd->device_info->ih_ring_entry_size) %
- kfd->interrupt_ring_size;
-
- /*
- * Ensure the rptr write update is not visible until
- * memcpy has finished reading.
- */
- smp_mb();
- atomic_set(&kfd->interrupt_ring_rptr, rptr);
-
- return true;
-}
-
-static void interrupt_wq(struct work_struct *work)
-{
- struct kfd_dev *dev = container_of(work, struct kfd_dev,
- interrupt_work);
-
- uint32_t ih_ring_entry[DIV_ROUND_UP(
- dev->device_info->ih_ring_entry_size,
- sizeof(uint32_t))];
-
- while (dequeue_ih_ring_entry(dev, ih_ring_entry))
- ;
-}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_module.c b/drivers/gpu/drm/amd/amdkfd/kfd_module.c
index 95d5af1..1c385c2 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_module.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_module.c
@@ -50,15 +50,10 @@ module_param(sched_policy, int, 0444);
MODULE_PARM_DESC(sched_policy,
"Kernel cmdline parameter that defines the amdkfd scheduling policy");
-int max_num_of_processes = KFD_MAX_NUM_OF_PROCESSES_DEFAULT;
-module_param(max_num_of_processes, int, 0444);
-MODULE_PARM_DESC(max_num_of_processes,
- "Kernel cmdline parameter that defines the amdkfd maximum number of supported processes");
-
-int max_num_of_queues_per_process = KFD_MAX_NUM_OF_QUEUES_PER_PROCESS_DEFAULT;
-module_param(max_num_of_queues_per_process, int, 0444);
-MODULE_PARM_DESC(max_num_of_queues_per_process,
- "Kernel cmdline parameter that defines the amdkfd maximum number of supported queues per process");
+int max_num_of_queues_per_device = KFD_MAX_NUM_OF_QUEUES_PER_DEVICE_DEFAULT;
+module_param(max_num_of_queues_per_device, int, 0444);
+MODULE_PARM_DESC(max_num_of_queues_per_device,
+ "Maximum number of supported queues per device (1 = Minimum, 4096 = default)");
bool kgd2kfd_init(unsigned interface_version,
const struct kfd2kgd_calls *f2g,
@@ -100,16 +95,10 @@ static int __init kfd_module_init(void)
}
/* Verify module parameters */
- if ((max_num_of_processes < 0) ||
- (max_num_of_processes > KFD_MAX_NUM_OF_PROCESSES)) {
- pr_err("kfd: max_num_of_processes must be between 0 to KFD_MAX_NUM_OF_PROCESSES\n");
- return -1;
- }
-
- if ((max_num_of_queues_per_process < 0) ||
- (max_num_of_queues_per_process >
- KFD_MAX_NUM_OF_QUEUES_PER_PROCESS)) {
- pr_err("kfd: max_num_of_queues_per_process must be between 0 to KFD_MAX_NUM_OF_QUEUES_PER_PROCESS\n");
+ if ((max_num_of_queues_per_device < 1) ||
+ (max_num_of_queues_per_device >
+ KFD_MAX_NUM_OF_QUEUES_PER_DEVICE)) {
+ pr_err("kfd: max_num_of_queues_per_device must be between 1 to KFD_MAX_NUM_OF_QUEUES_PER_DEVICE\n");
return -1;
}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pasid.c b/drivers/gpu/drm/amd/amdkfd/kfd_pasid.c
index 4c25ef5..6cfe7f1 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_pasid.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_pasid.c
@@ -30,7 +30,7 @@ static DEFINE_MUTEX(pasid_mutex);
int kfd_pasid_init(void)
{
- pasid_limit = max_num_of_processes;
+ pasid_limit = KFD_MAX_NUM_OF_PROCESSES;
pasid_bitmap = kcalloc(BITS_TO_LONGS(pasid_limit), sizeof(long), GFP_KERNEL);
if (!pasid_bitmap)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
index a5edb29..96dc10e 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
@@ -52,20 +52,19 @@
#define kfd_alloc_struct(ptr_to_struct) \
((typeof(ptr_to_struct)) kzalloc(sizeof(*ptr_to_struct), GFP_KERNEL))
-/* Kernel module parameter to specify maximum number of supported processes */
-extern int max_num_of_processes;
-
-#define KFD_MAX_NUM_OF_PROCESSES_DEFAULT 32
#define KFD_MAX_NUM_OF_PROCESSES 512
+#define KFD_MAX_NUM_OF_QUEUES_PER_PROCESS 1024
/*
- * Kernel module parameter to specify maximum number of supported queues
- * per process
+ * Kernel module parameter to specify maximum number of supported queues per
+ * device
*/
-extern int max_num_of_queues_per_process;
+extern int max_num_of_queues_per_device;
-#define KFD_MAX_NUM_OF_QUEUES_PER_PROCESS_DEFAULT 128
-#define KFD_MAX_NUM_OF_QUEUES_PER_PROCESS 1024
+#define KFD_MAX_NUM_OF_QUEUES_PER_DEVICE_DEFAULT 4096
+#define KFD_MAX_NUM_OF_QUEUES_PER_DEVICE \
+ (KFD_MAX_NUM_OF_PROCESSES * \
+ KFD_MAX_NUM_OF_QUEUES_PER_PROCESS)
#define KFD_KERNEL_QUEUE_SIZE 2048
@@ -135,22 +134,10 @@ struct kfd_dev {
struct kgd2kfd_shared_resources shared_resources;
- void *interrupt_ring;
- size_t interrupt_ring_size;
- atomic_t interrupt_ring_rptr;
- atomic_t interrupt_ring_wptr;
- struct work_struct interrupt_work;
- spinlock_t interrupt_lock;
-
/* QCM Device instance */
struct device_queue_manager *dqm;
bool init_complete;
- /*
- * Interrupts of interest to KFD are copied
- * from the HW ring into a SW ring.
- */
- bool interrupts_active;
};
/* KGD2KFD callbacks */
@@ -531,10 +518,7 @@ struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev);
struct kfd_dev *kfd_topology_enum_kfd_devices(uint8_t idx);
/* Interrupts */
-int kfd_interrupt_init(struct kfd_dev *dev);
-void kfd_interrupt_exit(struct kfd_dev *dev);
void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry);
-bool enqueue_ih_ring_entry(struct kfd_dev *kfd, const void *ih_ring_entry);
/* Power Management */
void kgd2kfd_suspend(struct kfd_dev *kfd);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
index 4752678..2fda1927 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
@@ -54,11 +54,11 @@ static int find_available_queue_slot(struct process_queue_manager *pqm,
pr_debug("kfd: in %s\n", __func__);
found = find_first_zero_bit(pqm->queue_slot_bitmap,
- max_num_of_queues_per_process);
+ KFD_MAX_NUM_OF_QUEUES_PER_PROCESS);
pr_debug("kfd: the new slot id %lu\n", found);
- if (found >= max_num_of_queues_per_process) {
+ if (found >= KFD_MAX_NUM_OF_QUEUES_PER_PROCESS) {
pr_info("amdkfd: Can not open more queues for process with pasid %d\n",
pqm->process->pasid);
return -ENOMEM;
@@ -76,7 +76,7 @@ int pqm_init(struct process_queue_manager *pqm, struct kfd_process *p)
INIT_LIST_HEAD(&pqm->queues);
pqm->queue_slot_bitmap =
- kzalloc(DIV_ROUND_UP(max_num_of_queues_per_process,
+ kzalloc(DIV_ROUND_UP(KFD_MAX_NUM_OF_QUEUES_PER_PROCESS,
BITS_PER_BYTE), GFP_KERNEL);
if (pqm->queue_slot_bitmap == NULL)
return -ENOMEM;
@@ -203,6 +203,7 @@ int pqm_create_queue(struct process_queue_manager *pqm,
pqn->kq = NULL;
retval = dev->dqm->create_queue(dev->dqm, q, &pdd->qpd,
&q->properties.vmid);
+ pr_debug("DQM returned %d for create_queue\n", retval);
print_queue(q);
break;
case KFD_QUEUE_TYPE_DIQ:
@@ -222,7 +223,7 @@ int pqm_create_queue(struct process_queue_manager *pqm,
}
if (retval != 0) {
- pr_err("kfd: error dqm create queue\n");
+ pr_debug("Error dqm create queue\n");
goto err_create_queue;
}
@@ -241,7 +242,10 @@ int pqm_create_queue(struct process_queue_manager *pqm,
err_create_queue:
kfree(pqn);
err_allocate_pqn:
+ /* check if queues list is empty unregister process from device */
clear_bit(*qid, pqm->queue_slot_bitmap);
+ if (list_empty(&pqm->queues))
+ dev->dqm->unregister_process(dev->dqm, &pdd->qpd);
return retval;
}
@@ -311,7 +315,11 @@ int pqm_update_queue(struct process_queue_manager *pqm, unsigned int qid,
BUG_ON(!pqm);
pqn = get_queue_by_qid(pqm, qid);
- BUG_ON(!pqn);
+ if (!pqn) {
+ pr_debug("amdkfd: No queue %d exists for update operation\n",
+ qid);
+ return -EFAULT;
+ }
pqn->q->properties.queue_address = p->queue_address;
pqn->q->properties.queue_size = p->queue_size;
diff --git a/drivers/gpu/drm/cirrus/cirrus_drv.c b/drivers/gpu/drm/cirrus/cirrus_drv.c
index c2a1cba..b914003 100644
--- a/drivers/gpu/drm/cirrus/cirrus_drv.c
+++ b/drivers/gpu/drm/cirrus/cirrus_drv.c
@@ -16,9 +16,12 @@
#include "cirrus_drv.h"
int cirrus_modeset = -1;
+int cirrus_bpp = 24;
MODULE_PARM_DESC(modeset, "Disable/Enable modesetting");
module_param_named(modeset, cirrus_modeset, int, 0400);
+MODULE_PARM_DESC(bpp, "Max bits-per-pixel (default:24)");
+module_param_named(bpp, cirrus_bpp, int, 0400);
/*
* This is the generic driver code. This binds the driver to the drm core,
diff --git a/drivers/gpu/drm/cirrus/cirrus_drv.h b/drivers/gpu/drm/cirrus/cirrus_drv.h
index 693a456..7050615 100644
--- a/drivers/gpu/drm/cirrus/cirrus_drv.h
+++ b/drivers/gpu/drm/cirrus/cirrus_drv.h
@@ -262,4 +262,7 @@ static inline void cirrus_bo_unreserve(struct cirrus_bo *bo)
int cirrus_bo_push_sysram(struct cirrus_bo *bo);
int cirrus_bo_pin(struct cirrus_bo *bo, u32 pl_flag, u64 *gpu_addr);
+
+extern int cirrus_bpp;
+
#endif /* __CIRRUS_DRV_H__ */
diff --git a/drivers/gpu/drm/cirrus/cirrus_main.c b/drivers/gpu/drm/cirrus/cirrus_main.c
index 4c2d68e..e4b9766 100644
--- a/drivers/gpu/drm/cirrus/cirrus_main.c
+++ b/drivers/gpu/drm/cirrus/cirrus_main.c
@@ -320,6 +320,8 @@ bool cirrus_check_framebuffer(struct cirrus_device *cdev, int width, int height,
const int max_pitch = 0x1FF << 3; /* (4096 - 1) & ~111b bytes */
const int max_size = cdev->mc.vram_size;
+ if (bpp > cirrus_bpp)
+ return false;
if (bpp > 32)
return false;
diff --git a/drivers/gpu/drm/cirrus/cirrus_mode.c b/drivers/gpu/drm/cirrus/cirrus_mode.c
index 99d4a74..61385f2 100644
--- a/drivers/gpu/drm/cirrus/cirrus_mode.c
+++ b/drivers/gpu/drm/cirrus/cirrus_mode.c
@@ -501,8 +501,13 @@ static int cirrus_vga_get_modes(struct drm_connector *connector)
int count;
/* Just add a static list of modes */
- count = drm_add_modes_noedid(connector, 1280, 1024);
- drm_set_preferred_mode(connector, 1024, 768);
+ if (cirrus_bpp <= 24) {
+ count = drm_add_modes_noedid(connector, 1280, 1024);
+ drm_set_preferred_mode(connector, 1024, 768);
+ } else {
+ count = drm_add_modes_noedid(connector, 800, 600);
+ drm_set_preferred_mode(connector, 800, 600);
+ }
return count;
}
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 52ce26d..dc386eb 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -145,6 +145,31 @@ int drm_fb_helper_add_one_connector(struct drm_fb_helper *fb_helper, struct drm_
}
EXPORT_SYMBOL(drm_fb_helper_add_one_connector);
+static void remove_from_modeset(struct drm_mode_set *set,
+ struct drm_connector *connector)
+{
+ int i, j;
+
+ for (i = 0; i < set->num_connectors; i++) {
+ if (set->connectors[i] == connector)
+ break;
+ }
+
+ if (i == set->num_connectors)
+ return;
+
+ for (j = i + 1; j < set->num_connectors; j++) {
+ set->connectors[j - 1] = set->connectors[j];
+ }
+ set->num_connectors--;
+
+ /* because i915 is pissy about this..
+ * TODO maybe need to makes sure we set it back to !=NULL somewhere?
+ */
+ if (set->num_connectors == 0)
+ set->fb = NULL;
+}
+
int drm_fb_helper_remove_one_connector(struct drm_fb_helper *fb_helper,
struct drm_connector *connector)
{
@@ -167,6 +192,11 @@ int drm_fb_helper_remove_one_connector(struct drm_fb_helper *fb_helper,
}
fb_helper->connector_count--;
kfree(fb_helper_connector);
+
+ /* also cleanup dangling references to the connector: */
+ for (i = 0; i < fb_helper->crtc_count; i++)
+ remove_from_modeset(&fb_helper->crtc_info[i].mode_set, connector);
+
return 0;
}
EXPORT_SYMBOL(drm_fb_helper_remove_one_connector);
@@ -741,7 +771,9 @@ int drm_fb_helper_setcmap(struct fb_cmap *cmap, struct fb_info *info)
int i, j, rc = 0;
int start;
- drm_modeset_lock_all(dev);
+ if (__drm_modeset_lock_all(dev, !!oops_in_progress)) {
+ return -EBUSY;
+ }
if (!drm_fb_helper_is_bound(fb_helper)) {
drm_modeset_unlock_all(dev);
return -EBUSY;
@@ -915,7 +947,9 @@ int drm_fb_helper_pan_display(struct fb_var_screeninfo *var,
int ret = 0;
int i;
- drm_modeset_lock_all(dev);
+ if (__drm_modeset_lock_all(dev, !!oops_in_progress)) {
+ return -EBUSY;
+ }
if (!drm_fb_helper_is_bound(fb_helper)) {
drm_modeset_unlock_all(dev);
return -EBUSY;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c
index 121470a..1bcbe07 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c
@@ -645,18 +645,6 @@ static int exynos_drm_init(void)
if (!is_exynos)
return -ENODEV;
- /*
- * Register device object only in case of Exynos SoC.
- *
- * Below codes resolves temporarily infinite loop issue incurred
- * by Exynos drm driver when using multi-platform kernel.
- * So these codes will be replaced with more generic way later.
- */
- if (!of_machine_is_compatible("samsung,exynos3") &&
- !of_machine_is_compatible("samsung,exynos4") &&
- !of_machine_is_compatible("samsung,exynos5"))
- return -ENODEV;
-
exynos_drm_pdev = platform_device_register_simple("exynos-drm", -1,
NULL, 0);
if (IS_ERR(exynos_drm_pdev))
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c
index 5765a16..98051e8 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmi.c
@@ -1669,7 +1669,6 @@ static void hdmi_mode_apply(struct hdmi_context *hdata)
static void hdmiphy_conf_reset(struct hdmi_context *hdata)
{
- u8 buffer[2];
u32 reg;
clk_disable_unprepare(hdata->res.sclk_hdmi);
@@ -1677,11 +1676,8 @@ static void hdmiphy_conf_reset(struct hdmi_context *hdata)
clk_prepare_enable(hdata->res.sclk_hdmi);
/* operation mode */
- buffer[0] = 0x1f;
- buffer[1] = 0x00;
-
- if (hdata->hdmiphy_port)
- i2c_master_send(hdata->hdmiphy_port, buffer, 2);
+ hdmiphy_reg_writeb(hdata, HDMIPHY_MODE_SET_DONE,
+ HDMI_PHY_ENABLE_MODE_SET);
if (hdata->type == HDMI_TYPE13)
reg = HDMI_V13_PHY_RSTOUT;
diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
index 820b762..064ed65 100644
--- a/drivers/gpu/drm/exynos/exynos_mixer.c
+++ b/drivers/gpu/drm/exynos/exynos_mixer.c
@@ -1026,6 +1026,7 @@ static void mixer_win_disable(struct exynos_drm_manager *mgr, int zpos)
static void mixer_wait_for_vblank(struct exynos_drm_manager *mgr)
{
struct mixer_context *mixer_ctx = mgr_to_mixer(mgr);
+ int err;
mutex_lock(&mixer_ctx->mixer_mutex);
if (!mixer_ctx->powered) {
@@ -1034,7 +1035,11 @@ static void mixer_wait_for_vblank(struct exynos_drm_manager *mgr)
}
mutex_unlock(&mixer_ctx->mixer_mutex);
- drm_vblank_get(mgr->crtc->dev, mixer_ctx->pipe);
+ err = drm_vblank_get(mgr->crtc->dev, mixer_ctx->pipe);
+ if (err < 0) {
+ DRM_DEBUG_KMS("failed to acquire vblank counter\n");
+ return;
+ }
atomic_set(&mixer_ctx->wait_vsync_event, 1);
@@ -1262,8 +1267,6 @@ static int mixer_bind(struct device *dev, struct device *manager, void *data)
return ret;
}
- pm_runtime_enable(dev);
-
return 0;
}
@@ -1272,8 +1275,6 @@ static void mixer_unbind(struct device *dev, struct device *master, void *data)
struct mixer_context *ctx = dev_get_drvdata(dev);
mixer_mgr_remove(&ctx->manager);
-
- pm_runtime_disable(dev);
}
static const struct component_ops mixer_component_ops = {
diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c
index d476279..a9041d1 100644
--- a/drivers/gpu/drm/i2c/tda998x_drv.c
+++ b/drivers/gpu/drm/i2c/tda998x_drv.c
@@ -32,6 +32,8 @@
struct tda998x_priv {
struct i2c_client *cec;
struct i2c_client *hdmi;
+ struct mutex mutex;
+ struct delayed_work dwork;
uint16_t rev;
uint8_t current_page;
int dpms;
@@ -402,9 +404,10 @@ reg_read_range(struct tda998x_priv *priv, uint16_t reg, char *buf, int cnt)
uint8_t addr = REG2ADDR(reg);
int ret;
+ mutex_lock(&priv->mutex);
ret = set_page(priv, reg);
if (ret < 0)
- return ret;
+ goto out;
ret = i2c_master_send(client, &addr, sizeof(addr));
if (ret < 0)
@@ -414,10 +417,12 @@ reg_read_range(struct tda998x_priv *priv, uint16_t reg, char *buf, int cnt)
if (ret < 0)
goto fail;
- return ret;
+ goto out;
fail:
dev_err(&client->dev, "Error %d reading from 0x%x\n", ret, reg);
+out:
+ mutex_unlock(&priv->mutex);
return ret;
}
@@ -431,13 +436,16 @@ reg_write_range(struct tda998x_priv *priv, uint16_t reg, uint8_t *p, int cnt)
buf[0] = REG2ADDR(reg);
memcpy(&buf[1], p, cnt);
+ mutex_lock(&priv->mutex);
ret = set_page(priv, reg);
if (ret < 0)
- return;
+ goto out;
ret = i2c_master_send(client, buf, cnt + 1);
if (ret < 0)
dev_err(&client->dev, "Error %d writing to 0x%x\n", ret, reg);
+out:
+ mutex_unlock(&priv->mutex);
}
static int
@@ -459,13 +467,16 @@ reg_write(struct tda998x_priv *priv, uint16_t reg, uint8_t val)
uint8_t buf[] = {REG2ADDR(reg), val};
int ret;
+ mutex_lock(&priv->mutex);
ret = set_page(priv, reg);
if (ret < 0)
- return;
+ goto out;
ret = i2c_master_send(client, buf, sizeof(buf));
if (ret < 0)
dev_err(&client->dev, "Error %d writing to 0x%x\n", ret, reg);
+out:
+ mutex_unlock(&priv->mutex);
}
static void
@@ -475,13 +486,16 @@ reg_write16(struct tda998x_priv *priv, uint16_t reg, uint16_t val)
uint8_t buf[] = {REG2ADDR(reg), val >> 8, val};
int ret;
+ mutex_lock(&priv->mutex);
ret = set_page(priv, reg);
if (ret < 0)
- return;
+ goto out;
ret = i2c_master_send(client, buf, sizeof(buf));
if (ret < 0)
dev_err(&client->dev, "Error %d writing to 0x%x\n", ret, reg);
+out:
+ mutex_unlock(&priv->mutex);
}
static void
@@ -536,6 +550,17 @@ tda998x_reset(struct tda998x_priv *priv)
reg_write(priv, REG_MUX_VP_VIP_OUT, 0x24);
}
+/* handle HDMI connect/disconnect */
+static void tda998x_hpd(struct work_struct *work)
+{
+ struct delayed_work *dwork = to_delayed_work(work);
+ struct tda998x_priv *priv =
+ container_of(dwork, struct tda998x_priv, dwork);
+
+ if (priv->encoder && priv->encoder->dev)
+ drm_kms_helper_hotplug_event(priv->encoder->dev);
+}
+
/*
* only 2 interrupts may occur: screen plug/unplug and EDID read
*/
@@ -559,8 +584,7 @@ static irqreturn_t tda998x_irq_thread(int irq, void *data)
priv->wq_edid_wait = 0;
wake_up(&priv->wq_edid);
} else if (cec != 0) { /* HPD change */
- if (priv->encoder && priv->encoder->dev)
- drm_helper_hpd_irq_event(priv->encoder->dev);
+ schedule_delayed_work(&priv->dwork, HZ/10);
}
return IRQ_HANDLED;
}
@@ -1170,8 +1194,10 @@ static void tda998x_destroy(struct tda998x_priv *priv)
/* disable all IRQs and free the IRQ handler */
cec_write(priv, REG_CEC_RXSHPDINTENA, 0);
reg_clear(priv, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD);
- if (priv->hdmi->irq)
+ if (priv->hdmi->irq) {
free_irq(priv->hdmi->irq, priv);
+ cancel_delayed_work_sync(&priv->dwork);
+ }
i2c_unregister_device(priv->cec);
}
@@ -1255,6 +1281,7 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv)
struct device_node *np = client->dev.of_node;
u32 video;
int rev_lo, rev_hi, ret;
+ unsigned short cec_addr;
priv->vip_cntrl_0 = VIP_CNTRL_0_SWAP_A(2) | VIP_CNTRL_0_SWAP_B(3);
priv->vip_cntrl_1 = VIP_CNTRL_1_SWAP_C(0) | VIP_CNTRL_1_SWAP_D(1);
@@ -1262,12 +1289,16 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv)
priv->current_page = 0xff;
priv->hdmi = client;
- priv->cec = i2c_new_dummy(client->adapter, 0x34);
+ /* CEC I2C address bound to TDA998x I2C addr by configuration pins */
+ cec_addr = 0x34 + (client->addr & 0x03);
+ priv->cec = i2c_new_dummy(client->adapter, cec_addr);
if (!priv->cec)
return -ENODEV;
priv->dpms = DRM_MODE_DPMS_OFF;
+ mutex_init(&priv->mutex); /* protect the page access */
+
/* wake up the device: */
cec_write(priv, REG_CEC_ENAMODS,
CEC_ENAMODS_EN_RXSENS | CEC_ENAMODS_EN_HDMI);
@@ -1323,8 +1354,9 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv)
if (client->irq) {
int irqf_trigger;
- /* init read EDID waitqueue */
+ /* init read EDID waitqueue and HDP work */
init_waitqueue_head(&priv->wq_edid);
+ INIT_DELAYED_WORK(&priv->dwork, tda998x_hpd);
/* clear pending interrupts */
reg_read(priv, REG_INT_FLAGS_0);
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 574057c..7643300 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -462,19 +462,13 @@ void intel_detect_pch(struct drm_device *dev)
} else if (id == INTEL_PCH_LPT_DEVICE_ID_TYPE) {
dev_priv->pch_type = PCH_LPT;
DRM_DEBUG_KMS("Found LynxPoint PCH\n");
- WARN_ON(!IS_HASWELL(dev));
- WARN_ON(IS_HSW_ULT(dev));
- } else if (IS_BROADWELL(dev)) {
- dev_priv->pch_type = PCH_LPT;
- dev_priv->pch_id =
- INTEL_PCH_LPT_LP_DEVICE_ID_TYPE;
- DRM_DEBUG_KMS("This is Broadwell, assuming "
- "LynxPoint LP PCH\n");
+ WARN_ON(!IS_HASWELL(dev) && !IS_BROADWELL(dev));
+ WARN_ON(IS_HSW_ULT(dev) || IS_BDW_ULT(dev));
} else if (id == INTEL_PCH_LPT_LP_DEVICE_ID_TYPE) {
dev_priv->pch_type = PCH_LPT;
DRM_DEBUG_KMS("Found LynxPoint LP PCH\n");
- WARN_ON(!IS_HASWELL(dev));
- WARN_ON(!IS_HSW_ULT(dev));
+ WARN_ON(!IS_HASWELL(dev) && !IS_BROADWELL(dev));
+ WARN_ON(!IS_HSW_ULT(dev) && !IS_BDW_ULT(dev));
} else if (id == INTEL_PCH_SPT_DEVICE_ID_TYPE) {
dev_priv->pch_type = PCH_SPT;
DRM_DEBUG_KMS("Found SunrisePoint PCH\n");
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index e9f891c..9d7a715 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -2159,8 +2159,7 @@ struct drm_i915_cmd_table {
#define IS_HSW_EARLY_SDV(dev) (IS_HASWELL(dev) && \
(INTEL_DEVID(dev) & 0xFF00) == 0x0C00)
#define IS_BDW_ULT(dev) (IS_BROADWELL(dev) && \
- ((INTEL_DEVID(dev) & 0xf) == 0x2 || \
- (INTEL_DEVID(dev) & 0xf) == 0x6 || \
+ ((INTEL_DEVID(dev) & 0xf) == 0x6 || \
(INTEL_DEVID(dev) & 0xf) == 0xe))
#define IS_BDW_GT3(dev) (IS_BROADWELL(dev) && \
(INTEL_DEVID(dev) & 0x00F0) == 0x0020)
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index c11603b..5f61482 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3148,6 +3148,13 @@ static void i965_write_fence_reg(struct drm_device *dev, int reg,
u32 size = i915_gem_obj_ggtt_size(obj);
uint64_t val;
+ /* Adjust fence size to match tiled area */
+ if (obj->tiling_mode != I915_TILING_NONE) {
+ uint32_t row_size = obj->stride *
+ (obj->tiling_mode == I915_TILING_Y ? 32 : 8);
+ size = (size / row_size) * row_size;
+ }
+
val = (uint64_t)((i915_gem_obj_ggtt_offset(obj) + size - 4096) &
0xfffff000) << 32;
val |= i915_gem_obj_ggtt_offset(obj) & 0xfffff000;
@@ -4884,25 +4891,18 @@ i915_gem_init_hw(struct drm_device *dev)
for (i = 0; i < NUM_L3_SLICES(dev); i++)
i915_gem_l3_remap(&dev_priv->ring[RCS], i);
- /*
- * XXX: Contexts should only be initialized once. Doing a switch to the
- * default context switch however is something we'd like to do after
- * reset or thaw (the latter may not actually be necessary for HW, but
- * goes with our code better). Context switching requires rings (for
- * the do_switch), but before enabling PPGTT. So don't move this.
- */
- ret = i915_gem_context_enable(dev_priv);
+ ret = i915_ppgtt_init_hw(dev);
if (ret && ret != -EIO) {
- DRM_ERROR("Context enable failed %d\n", ret);
+ DRM_ERROR("PPGTT enable failed %d\n", ret);
i915_gem_cleanup_ringbuffer(dev);
-
- return ret;
}
- ret = i915_ppgtt_init_hw(dev);
+ ret = i915_gem_context_enable(dev_priv);
if (ret && ret != -EIO) {
- DRM_ERROR("PPGTT enable failed %d\n", ret);
+ DRM_ERROR("Context enable failed %d\n", ret);
i915_gem_cleanup_ringbuffer(dev);
+
+ return ret;
}
return ret;
@@ -5155,7 +5155,7 @@ static bool mutex_is_locked_by(struct mutex *mutex, struct task_struct *task)
if (!mutex_is_locked(mutex))
return false;
-#if defined(CONFIG_SMP) || defined(CONFIG_DEBUG_MUTEXES)
+#if defined(CONFIG_SMP) && !defined(CONFIG_DEBUG_MUTEXES)
return mutex->owner == task;
#else
/* Since UP may be pre-empted, we cannot assume that we own the lock */
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index d0d3dfb..b051a23 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -292,6 +292,23 @@ void gen6_enable_rps_interrupts(struct drm_device *dev)
spin_unlock_irq(&dev_priv->irq_lock);
}
+u32 gen6_sanitize_rps_pm_mask(struct drm_i915_private *dev_priv, u32 mask)
+{
+ /*
+ * SNB,IVB can while VLV,CHV may hard hang on looping batchbuffer
+ * if GEN6_PM_UP_EI_EXPIRED is masked.
+ *
+ * TODO: verify if this can be reproduced on VLV,CHV.
+ */
+ if (INTEL_INFO(dev_priv)->gen <= 7 && !IS_HASWELL(dev_priv))
+ mask &= ~GEN6_PM_RP_UP_EI_EXPIRED;
+
+ if (INTEL_INFO(dev_priv)->gen >= 8)
+ mask &= ~GEN8_PMINTR_REDIRECT_TO_NON_DISP;
+
+ return mask;
+}
+
void gen6_disable_rps_interrupts(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
@@ -304,8 +321,7 @@ void gen6_disable_rps_interrupts(struct drm_device *dev)
spin_lock_irq(&dev_priv->irq_lock);
- I915_WRITE(GEN6_PMINTRMSK, INTEL_INFO(dev_priv)->gen >= 8 ?
- ~GEN8_PMINTR_REDIRECT_TO_NON_DISP : ~0);
+ I915_WRITE(GEN6_PMINTRMSK, gen6_sanitize_rps_pm_mask(dev_priv, ~0));
__gen6_disable_pm_irq(dev_priv, dev_priv->pm_rps_events);
I915_WRITE(gen6_pm_ier(dev_priv), I915_READ(gen6_pm_ier(dev_priv)) &
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index e2af138..e7a16f1 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -9815,7 +9815,7 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
if (obj->tiling_mode != work->old_fb_obj->tiling_mode)
/* vlv: DISPLAY_FLIP fails to change tiling */
ring = NULL;
- } else if (IS_IVYBRIDGE(dev)) {
+ } else if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev)) {
ring = &dev_priv->ring[BCS];
} else if (INTEL_INFO(dev)->gen >= 7) {
ring = obj->ring;
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 25fdbb1..3b40a17 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -794,6 +794,7 @@ void gen6_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
void gen6_reset_rps_interrupts(struct drm_device *dev);
void gen6_enable_rps_interrupts(struct drm_device *dev);
void gen6_disable_rps_interrupts(struct drm_device *dev);
+u32 gen6_sanitize_rps_pm_mask(struct drm_i915_private *dev_priv, u32 mask);
void intel_runtime_pm_disable_interrupts(struct drm_i915_private *dev_priv);
void intel_runtime_pm_enable_interrupts(struct drm_i915_private *dev_priv);
static inline bool intel_irqs_enabled(struct drm_i915_private *dev_priv)
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index 4d63839..dfb783a 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -962,7 +962,7 @@ void intel_panel_enable_backlight(struct intel_connector *connector)
WARN_ON(panel->backlight.max == 0);
- if (panel->backlight.level == 0) {
+ if (panel->backlight.level <= panel->backlight.min) {
panel->backlight.level = panel->backlight.max;
if (panel->backlight.device)
panel->backlight.device->props.brightness =
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 964b28e..bf814a6 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -4363,16 +4363,7 @@ static u32 gen6_rps_pm_mask(struct drm_i915_private *dev_priv, u8 val)
mask |= dev_priv->pm_rps_events & (GEN6_PM_RP_DOWN_EI_EXPIRED | GEN6_PM_RP_UP_EI_EXPIRED);
mask &= dev_priv->pm_rps_events;
- /* IVB and SNB hard hangs on looping batchbuffer
- * if GEN6_PM_UP_EI_EXPIRED is masked.
- */
- if (INTEL_INFO(dev_priv->dev)->gen <= 7 && !IS_HASWELL(dev_priv->dev))
- mask |= GEN6_PM_RP_UP_EI_EXPIRED;
-
- if (IS_GEN8(dev_priv->dev))
- mask |= GEN8_PMINTR_REDIRECT_TO_NON_DISP;
-
- return ~mask;
+ return gen6_sanitize_rps_pm_mask(dev_priv, ~mask);
}
/* gen6_set_rps is called to update the frequency request, but should also be
@@ -4441,7 +4432,8 @@ static void vlv_set_rps_idle(struct drm_i915_private *dev_priv)
return;
/* Mask turbo interrupt so that they will not come in between */
- I915_WRITE(GEN6_PMINTRMSK, 0xffffffff);
+ I915_WRITE(GEN6_PMINTRMSK,
+ gen6_sanitize_rps_pm_mask(dev_priv, ~0));
vlv_force_gfx_clock(dev_priv, true);
diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c
index 6dcde37..64fdae5 100644
--- a/drivers/gpu/drm/radeon/cik.c
+++ b/drivers/gpu/drm/radeon/cik.c
@@ -6033,6 +6033,17 @@ void cik_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
radeon_ring_write(ring, 0);
radeon_ring_write(ring, 1 << vm_id);
+ /* wait for the invalidate to complete */
+ radeon_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));
+ radeon_ring_write(ring, (WAIT_REG_MEM_OPERATION(0) | /* wait */
+ WAIT_REG_MEM_FUNCTION(0) | /* always */
+ WAIT_REG_MEM_ENGINE(0))); /* me */
+ radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
+ radeon_ring_write(ring, 0);
+ radeon_ring_write(ring, 0); /* ref */
+ radeon_ring_write(ring, 0); /* mask */
+ radeon_ring_write(ring, 0x20); /* poll interval */
+
/* compute doesn't have PFP */
if (usepfp) {
/* sync PFP to ME, otherwise we might get invalid PFP reads */
diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c
index dde5c7e..42cd0cf 100644
--- a/drivers/gpu/drm/radeon/cik_sdma.c
+++ b/drivers/gpu/drm/radeon/cik_sdma.c
@@ -816,7 +816,6 @@ void cik_sdma_vm_write_pages(struct radeon_device *rdev,
for (; ndw > 0; ndw -= 2, --count, pe += 8) {
if (flags & R600_PTE_SYSTEM) {
value = radeon_vm_map_gart(rdev, addr);
- value &= 0xFFFFFFFFFFFFF000ULL;
} else if (flags & R600_PTE_VALID) {
value = addr;
} else {
@@ -903,6 +902,9 @@ void cik_sdma_vm_pad_ib(struct radeon_ib *ib)
void cik_dma_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
unsigned vm_id, uint64_t pd_addr)
{
+ u32 extra_bits = (SDMA_POLL_REG_MEM_EXTRA_OP(0) |
+ SDMA_POLL_REG_MEM_EXTRA_FUNC(0)); /* always */
+
radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
if (vm_id < 8) {
radeon_ring_write(ring, (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm_id << 2)) >> 2);
@@ -943,5 +945,12 @@ void cik_dma_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
radeon_ring_write(ring, 1 << vm_id);
+
+ radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_POLL_REG_MEM, 0, extra_bits));
+ radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
+ radeon_ring_write(ring, 0);
+ radeon_ring_write(ring, 0); /* reference */
+ radeon_ring_write(ring, 0); /* mask */
+ radeon_ring_write(ring, (0xfff << 16) | 10); /* retry count, poll interval */
}
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 360de9f..aea48c8 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -2516,6 +2516,16 @@ void cayman_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
radeon_ring_write(ring, PACKET0(VM_INVALIDATE_REQUEST, 0));
radeon_ring_write(ring, 1 << vm_id);
+ /* wait for the invalidate to complete */
+ radeon_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));
+ radeon_ring_write(ring, (WAIT_REG_MEM_FUNCTION(0) | /* always */
+ WAIT_REG_MEM_ENGINE(0))); /* me */
+ radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
+ radeon_ring_write(ring, 0);
+ radeon_ring_write(ring, 0); /* ref */
+ radeon_ring_write(ring, 0); /* mask */
+ radeon_ring_write(ring, 0x20); /* poll interval */
+
/* sync PFP to ME, otherwise we might get invalid PFP reads */
radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
radeon_ring_write(ring, 0x0);
diff --git a/drivers/gpu/drm/radeon/ni_dma.c b/drivers/gpu/drm/radeon/ni_dma.c
index 50f8861..ce787a9 100644
--- a/drivers/gpu/drm/radeon/ni_dma.c
+++ b/drivers/gpu/drm/radeon/ni_dma.c
@@ -372,7 +372,6 @@ void cayman_dma_vm_write_pages(struct radeon_device *rdev,
for (; ndw > 0; ndw -= 2, --count, pe += 8) {
if (flags & R600_PTE_SYSTEM) {
value = radeon_vm_map_gart(rdev, addr);
- value &= 0xFFFFFFFFFFFFF000ULL;
} else if (flags & R600_PTE_VALID) {
value = addr;
} else {
@@ -463,5 +462,11 @@ void cayman_dma_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0));
radeon_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST >> 2));
radeon_ring_write(ring, 1 << vm_id);
+
+ /* wait for invalidate to complete */
+ radeon_ring_write(ring, DMA_SRBM_READ_PACKET);
+ radeon_ring_write(ring, (0xff << 20) | (VM_INVALIDATE_REQUEST >> 2));
+ radeon_ring_write(ring, 0); /* mask */
+ radeon_ring_write(ring, 0); /* value */
}
diff --git a/drivers/gpu/drm/radeon/nid.h b/drivers/gpu/drm/radeon/nid.h
index 2e12e4d..ad71254 100644
--- a/drivers/gpu/drm/radeon/nid.h
+++ b/drivers/gpu/drm/radeon/nid.h
@@ -1133,6 +1133,23 @@
#define PACKET3_MEM_SEMAPHORE 0x39
#define PACKET3_MPEG_INDEX 0x3A
#define PACKET3_WAIT_REG_MEM 0x3C
+#define WAIT_REG_MEM_FUNCTION(x) ((x) << 0)
+ /* 0 - always
+ * 1 - <
+ * 2 - <=
+ * 3 - ==
+ * 4 - !=
+ * 5 - >=
+ * 6 - >
+ */
+#define WAIT_REG_MEM_MEM_SPACE(x) ((x) << 4)
+ /* 0 - reg
+ * 1 - mem
+ */
+#define WAIT_REG_MEM_ENGINE(x) ((x) << 8)
+ /* 0 - me
+ * 1 - pfp
+ */
#define PACKET3_MEM_WRITE 0x3D
#define PACKET3_PFP_SYNC_ME 0x42
#define PACKET3_SURFACE_SYNC 0x43
@@ -1272,6 +1289,13 @@
(1 << 21) | \
(((n) & 0xFFFFF) << 0))
+#define DMA_SRBM_POLL_PACKET ((9 << 28) | \
+ (1 << 27) | \
+ (1 << 26))
+
+#define DMA_SRBM_READ_PACKET ((9 << 28) | \
+ (1 << 27))
+
/* async DMA Packet types */
#define DMA_PACKET_WRITE 0x2
#define DMA_PACKET_COPY 0x3
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 74f06d5..279801c 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -644,6 +644,7 @@ int r100_pci_gart_init(struct radeon_device *rdev)
return r;
rdev->gart.table_size = rdev->gart.num_gpu_pages * 4;
rdev->asic->gart.tlb_flush = &r100_pci_gart_tlb_flush;
+ rdev->asic->gart.get_page_entry = &r100_pci_gart_get_page_entry;
rdev->asic->gart.set_page = &r100_pci_gart_set_page;
return radeon_gart_table_ram_alloc(rdev);
}
@@ -681,11 +682,16 @@ void r100_pci_gart_disable(struct radeon_device *rdev)
WREG32(RADEON_AIC_HI_ADDR, 0);
}
+uint64_t r100_pci_gart_get_page_entry(uint64_t addr, uint32_t flags)
+{
+ return addr;
+}
+
void r100_pci_gart_set_page(struct radeon_device *rdev, unsigned i,
- uint64_t addr, uint32_t flags)
+ uint64_t entry)
{
u32 *gtt = rdev->gart.ptr;
- gtt[i] = cpu_to_le32(lower_32_bits(addr));
+ gtt[i] = cpu_to_le32(lower_32_bits(entry));
}
void r100_pci_gart_fini(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 064ad55..08d68f3 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -73,11 +73,8 @@ void rv370_pcie_gart_tlb_flush(struct radeon_device *rdev)
#define R300_PTE_WRITEABLE (1 << 2)
#define R300_PTE_READABLE (1 << 3)
-void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i,
- uint64_t addr, uint32_t flags)
+uint64_t rv370_pcie_gart_get_page_entry(uint64_t addr, uint32_t flags)
{
- void __iomem *ptr = rdev->gart.ptr;
-
addr = (lower_32_bits(addr) >> 8) |
((upper_32_bits(addr) & 0xff) << 24);
if (flags & RADEON_GART_PAGE_READ)
@@ -86,10 +83,18 @@ void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i,
addr |= R300_PTE_WRITEABLE;
if (!(flags & RADEON_GART_PAGE_SNOOP))
addr |= R300_PTE_UNSNOOPED;
+ return addr;
+}
+
+void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i,
+ uint64_t entry)
+{
+ void __iomem *ptr = rdev->gart.ptr;
+
/* on x86 we want this to be CPU endian, on powerpc
* on powerpc without HW swappers, it'll get swapped on way
* into VRAM - so no need for cpu_to_le32 on VRAM tables */
- writel(addr, ((void __iomem *)ptr) + (i * 4));
+ writel(entry, ((void __iomem *)ptr) + (i * 4));
}
int rv370_pcie_gart_init(struct radeon_device *rdev)
@@ -109,6 +114,7 @@ int rv370_pcie_gart_init(struct radeon_device *rdev)
DRM_ERROR("Failed to register debugfs file for PCIE gart !\n");
rdev->gart.table_size = rdev->gart.num_gpu_pages * 4;
rdev->asic->gart.tlb_flush = &rv370_pcie_gart_tlb_flush;
+ rdev->asic->gart.get_page_entry = &rv370_pcie_gart_get_page_entry;
rdev->asic->gart.set_page = &rv370_pcie_gart_set_page;
return radeon_gart_table_vram_alloc(rdev);
}
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 54529b8..3f2a8d3 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -242,6 +242,7 @@ bool radeon_get_bios(struct radeon_device *rdev);
* Dummy page
*/
struct radeon_dummy_page {
+ uint64_t entry;
struct page *page;
dma_addr_t addr;
};
@@ -645,7 +646,7 @@ struct radeon_gart {
unsigned num_cpu_pages;
unsigned table_size;
struct page **pages;
- dma_addr_t *pages_addr;
+ uint64_t *pages_entry;
bool ready;
};
@@ -1847,8 +1848,9 @@ struct radeon_asic {
/* gart */
struct {
void (*tlb_flush)(struct radeon_device *rdev);
+ uint64_t (*get_page_entry)(uint64_t addr, uint32_t flags);
void (*set_page)(struct radeon_device *rdev, unsigned i,
- uint64_t addr, uint32_t flags);
+ uint64_t entry);
} gart;
struct {
int (*init)(struct radeon_device *rdev);
@@ -2852,7 +2854,8 @@ static inline void radeon_ring_write(struct radeon_ring *ring, uint32_t v)
#define radeon_vga_set_state(rdev, state) (rdev)->asic->vga_set_state((rdev), (state))
#define radeon_asic_reset(rdev) (rdev)->asic->asic_reset((rdev))
#define radeon_gart_tlb_flush(rdev) (rdev)->asic->gart.tlb_flush((rdev))
-#define radeon_gart_set_page(rdev, i, p, f) (rdev)->asic->gart.set_page((rdev), (i), (p), (f))
+#define radeon_gart_get_page_entry(a, f) (rdev)->asic->gart.get_page_entry((a), (f))
+#define radeon_gart_set_page(rdev, i, e) (rdev)->asic->gart.set_page((rdev), (i), (e))
#define radeon_asic_vm_init(rdev) (rdev)->asic->vm.init((rdev))
#define radeon_asic_vm_fini(rdev) (rdev)->asic->vm.fini((rdev))
#define radeon_asic_vm_copy_pages(rdev, ib, pe, src, count) ((rdev)->asic->vm.copy_pages((rdev), (ib), (pe), (src), (count)))
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index 850de57..ed0e10e 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -159,11 +159,13 @@ void radeon_agp_disable(struct radeon_device *rdev)
DRM_INFO("Forcing AGP to PCIE mode\n");
rdev->flags |= RADEON_IS_PCIE;
rdev->asic->gart.tlb_flush = &rv370_pcie_gart_tlb_flush;
+ rdev->asic->gart.get_page_entry = &rv370_pcie_gart_get_page_entry;
rdev->asic->gart.set_page = &rv370_pcie_gart_set_page;
} else {
DRM_INFO("Forcing AGP to PCI mode\n");
rdev->flags |= RADEON_IS_PCI;
rdev->asic->gart.tlb_flush = &r100_pci_gart_tlb_flush;
+ rdev->asic->gart.get_page_entry = &r100_pci_gart_get_page_entry;
rdev->asic->gart.set_page = &r100_pci_gart_set_page;
}
rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024;
@@ -199,6 +201,7 @@ static struct radeon_asic r100_asic = {
.mc_wait_for_idle = &r100_mc_wait_for_idle,
.gart = {
.tlb_flush = &r100_pci_gart_tlb_flush,
+ .get_page_entry = &r100_pci_gart_get_page_entry,
.set_page = &r100_pci_gart_set_page,
},
.ring = {
@@ -265,6 +268,7 @@ static struct radeon_asic r200_asic = {
.mc_wait_for_idle = &r100_mc_wait_for_idle,
.gart = {
.tlb_flush = &r100_pci_gart_tlb_flush,
+ .get_page_entry = &r100_pci_gart_get_page_entry,
.set_page = &r100_pci_gart_set_page,
},
.ring = {
@@ -333,6 +337,20 @@ static struct radeon_asic_ring r300_gfx_ring = {
.set_wptr = &r100_gfx_set_wptr,
};
+static struct radeon_asic_ring rv515_gfx_ring = {
+ .ib_execute = &r100_ring_ib_execute,
+ .emit_fence = &r300_fence_ring_emit,
+ .emit_semaphore = &r100_semaphore_ring_emit,
+ .cs_parse = &r300_cs_parse,
+ .ring_start = &rv515_ring_start,
+ .ring_test = &r100_ring_test,
+ .ib_test = &r100_ib_test,
+ .is_lockup = &r100_gpu_is_lockup,
+ .get_rptr = &r100_gfx_get_rptr,
+ .get_wptr = &r100_gfx_get_wptr,
+ .set_wptr = &r100_gfx_set_wptr,
+};
+
static struct radeon_asic r300_asic = {
.init = &r300_init,
.fini = &r300_fini,
@@ -345,6 +363,7 @@ static struct radeon_asic r300_asic = {
.mc_wait_for_idle = &r300_mc_wait_for_idle,
.gart = {
.tlb_flush = &r100_pci_gart_tlb_flush,
+ .get_page_entry = &r100_pci_gart_get_page_entry,
.set_page = &r100_pci_gart_set_page,
},
.ring = {
@@ -411,6 +430,7 @@ static struct radeon_asic r300_asic_pcie = {
.mc_wait_for_idle = &r300_mc_wait_for_idle,
.gart = {
.tlb_flush = &rv370_pcie_gart_tlb_flush,
+ .get_page_entry = &rv370_pcie_gart_get_page_entry,
.set_page = &rv370_pcie_gart_set_page,
},
.ring = {
@@ -477,6 +497,7 @@ static struct radeon_asic r420_asic = {
.mc_wait_for_idle = &r300_mc_wait_for_idle,
.gart = {
.tlb_flush = &rv370_pcie_gart_tlb_flush,
+ .get_page_entry = &rv370_pcie_gart_get_page_entry,
.set_page = &rv370_pcie_gart_set_page,
},
.ring = {
@@ -543,6 +564,7 @@ static struct radeon_asic rs400_asic = {
.mc_wait_for_idle = &rs400_mc_wait_for_idle,
.gart = {
.tlb_flush = &rs400_gart_tlb_flush,
+ .get_page_entry = &rs400_gart_get_page_entry,
.set_page = &rs400_gart_set_page,
},
.ring = {
@@ -609,6 +631,7 @@ static struct radeon_asic rs600_asic = {
.mc_wait_for_idle = &rs600_mc_wait_for_idle,
.gart = {
.tlb_flush = &rs600_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.ring = {
@@ -677,6 +700,7 @@ static struct radeon_asic rs690_asic = {
.mc_wait_for_idle = &rs690_mc_wait_for_idle,
.gart = {
.tlb_flush = &rs400_gart_tlb_flush,
+ .get_page_entry = &rs400_gart_get_page_entry,
.set_page = &rs400_gart_set_page,
},
.ring = {
@@ -745,10 +769,11 @@ static struct radeon_asic rv515_asic = {
.mc_wait_for_idle = &rv515_mc_wait_for_idle,
.gart = {
.tlb_flush = &rv370_pcie_gart_tlb_flush,
+ .get_page_entry = &rv370_pcie_gart_get_page_entry,
.set_page = &rv370_pcie_gart_set_page,
},
.ring = {
- [RADEON_RING_TYPE_GFX_INDEX] = &r300_gfx_ring
+ [RADEON_RING_TYPE_GFX_INDEX] = &rv515_gfx_ring
},
.irq = {
.set = &rs600_irq_set,
@@ -811,10 +836,11 @@ static struct radeon_asic r520_asic = {
.mc_wait_for_idle = &r520_mc_wait_for_idle,
.gart = {
.tlb_flush = &rv370_pcie_gart_tlb_flush,
+ .get_page_entry = &rv370_pcie_gart_get_page_entry,
.set_page = &rv370_pcie_gart_set_page,
},
.ring = {
- [RADEON_RING_TYPE_GFX_INDEX] = &r300_gfx_ring
+ [RADEON_RING_TYPE_GFX_INDEX] = &rv515_gfx_ring
},
.irq = {
.set = &rs600_irq_set,
@@ -905,6 +931,7 @@ static struct radeon_asic r600_asic = {
.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
.gart = {
.tlb_flush = &r600_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.ring = {
@@ -990,6 +1017,7 @@ static struct radeon_asic rv6xx_asic = {
.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
.gart = {
.tlb_flush = &r600_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.ring = {
@@ -1081,6 +1109,7 @@ static struct radeon_asic rs780_asic = {
.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
.gart = {
.tlb_flush = &r600_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.ring = {
@@ -1185,6 +1214,7 @@ static struct radeon_asic rv770_asic = {
.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
.gart = {
.tlb_flush = &r600_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.ring = {
@@ -1303,6 +1333,7 @@ static struct radeon_asic evergreen_asic = {
.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
.gart = {
.tlb_flush = &evergreen_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.ring = {
@@ -1395,6 +1426,7 @@ static struct radeon_asic sumo_asic = {
.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
.gart = {
.tlb_flush = &evergreen_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.ring = {
@@ -1486,6 +1518,7 @@ static struct radeon_asic btc_asic = {
.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
.gart = {
.tlb_flush = &evergreen_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.ring = {
@@ -1621,6 +1654,7 @@ static struct radeon_asic cayman_asic = {
.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
.gart = {
.tlb_flush = &cayman_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.vm = {
@@ -1724,6 +1758,7 @@ static struct radeon_asic trinity_asic = {
.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
.gart = {
.tlb_flush = &cayman_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.vm = {
@@ -1857,6 +1892,7 @@ static struct radeon_asic si_asic = {
.get_gpu_clock_counter = &si_get_gpu_clock_counter,
.gart = {
.tlb_flush = &si_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.vm = {
@@ -2018,6 +2054,7 @@ static struct radeon_asic ci_asic = {
.get_gpu_clock_counter = &cik_get_gpu_clock_counter,
.gart = {
.tlb_flush = &cik_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.vm = {
@@ -2125,6 +2162,7 @@ static struct radeon_asic kv_asic = {
.get_gpu_clock_counter = &cik_get_gpu_clock_counter,
.gart = {
.tlb_flush = &cik_pcie_gart_tlb_flush,
+ .get_page_entry = &rs600_gart_get_page_entry,
.set_page = &rs600_gart_set_page,
},
.vm = {
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 2a45d54..8d787d1 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -67,8 +67,9 @@ bool r100_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
int r100_asic_reset(struct radeon_device *rdev);
u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc);
void r100_pci_gart_tlb_flush(struct radeon_device *rdev);
+uint64_t r100_pci_gart_get_page_entry(uint64_t addr, uint32_t flags);
void r100_pci_gart_set_page(struct radeon_device *rdev, unsigned i,
- uint64_t addr, uint32_t flags);
+ uint64_t entry);
void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring);
int r100_irq_set(struct radeon_device *rdev);
int r100_irq_process(struct radeon_device *rdev);
@@ -172,8 +173,9 @@ extern void r300_fence_ring_emit(struct radeon_device *rdev,
struct radeon_fence *fence);
extern int r300_cs_parse(struct radeon_cs_parser *p);
extern void rv370_pcie_gart_tlb_flush(struct radeon_device *rdev);
+extern uint64_t rv370_pcie_gart_get_page_entry(uint64_t addr, uint32_t flags);
extern void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i,
- uint64_t addr, uint32_t flags);
+ uint64_t entry);
extern void rv370_set_pcie_lanes(struct radeon_device *rdev, int lanes);
extern int rv370_get_pcie_lanes(struct radeon_device *rdev);
extern void r300_set_reg_safe(struct radeon_device *rdev);
@@ -208,8 +210,9 @@ extern void rs400_fini(struct radeon_device *rdev);
extern int rs400_suspend(struct radeon_device *rdev);
extern int rs400_resume(struct radeon_device *rdev);
void rs400_gart_tlb_flush(struct radeon_device *rdev);
+uint64_t rs400_gart_get_page_entry(uint64_t addr, uint32_t flags);
void rs400_gart_set_page(struct radeon_device *rdev, unsigned i,
- uint64_t addr, uint32_t flags);
+ uint64_t entry);
uint32_t rs400_mc_rreg(struct radeon_device *rdev, uint32_t reg);
void rs400_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
int rs400_gart_init(struct radeon_device *rdev);
@@ -232,8 +235,9 @@ int rs600_irq_process(struct radeon_device *rdev);
void rs600_irq_disable(struct radeon_device *rdev);
u32 rs600_get_vblank_counter(struct radeon_device *rdev, int crtc);
void rs600_gart_tlb_flush(struct radeon_device *rdev);
+uint64_t rs600_gart_get_page_entry(uint64_t addr, uint32_t flags);
void rs600_gart_set_page(struct radeon_device *rdev, unsigned i,
- uint64_t addr, uint32_t flags);
+ uint64_t entry);
uint32_t rs600_mc_rreg(struct radeon_device *rdev, uint32_t reg);
void rs600_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
void rs600_bandwidth_update(struct radeon_device *rdev);
diff --git a/drivers/gpu/drm/radeon/radeon_benchmark.c b/drivers/gpu/drm/radeon/radeon_benchmark.c
index 9e7f23d..87d5fb2 100644
--- a/drivers/gpu/drm/radeon/radeon_benchmark.c
+++ b/drivers/gpu/drm/radeon/radeon_benchmark.c
@@ -34,7 +34,8 @@
static int radeon_benchmark_do_move(struct radeon_device *rdev, unsigned size,
uint64_t saddr, uint64_t daddr,
- int flag, int n)
+ int flag, int n,
+ struct reservation_object *resv)
{
unsigned long start_jiffies;
unsigned long end_jiffies;
@@ -47,12 +48,12 @@ static int radeon_benchmark_do_move(struct radeon_device *rdev, unsigned size,
case RADEON_BENCHMARK_COPY_DMA:
fence = radeon_copy_dma(rdev, saddr, daddr,
size / RADEON_GPU_PAGE_SIZE,
- NULL);
+ resv);
break;
case RADEON_BENCHMARK_COPY_BLIT:
fence = radeon_copy_blit(rdev, saddr, daddr,
size / RADEON_GPU_PAGE_SIZE,
- NULL);
+ resv);
break;
default:
DRM_ERROR("Unknown copy method\n");
@@ -120,7 +121,8 @@ static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size,
if (rdev->asic->copy.dma) {
time = radeon_benchmark_do_move(rdev, size, saddr, daddr,
- RADEON_BENCHMARK_COPY_DMA, n);
+ RADEON_BENCHMARK_COPY_DMA, n,
+ dobj->tbo.resv);
if (time < 0)
goto out_cleanup;
if (time > 0)
@@ -130,7 +132,8 @@ static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size,
if (rdev->asic->copy.blit) {
time = radeon_benchmark_do_move(rdev, size, saddr, daddr,
- RADEON_BENCHMARK_COPY_BLIT, n);
+ RADEON_BENCHMARK_COPY_BLIT, n,
+ dobj->tbo.resv);
if (time < 0)
goto out_cleanup;
if (time > 0)
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 0ec6516..bd7519f 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -774,6 +774,8 @@ int radeon_dummy_page_init(struct radeon_device *rdev)
rdev->dummy_page.page = NULL;
return -ENOMEM;
}
+ rdev->dummy_page.entry = radeon_gart_get_page_entry(rdev->dummy_page.addr,
+ RADEON_GART_PAGE_DUMMY);
return 0;
}
diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index 1021169..913fafa 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -960,6 +960,9 @@ void radeon_compute_pll_avivo(struct radeon_pll *pll,
if (pll->flags & RADEON_PLL_USE_FRAC_FB_DIV &&
pll->flags & RADEON_PLL_USE_REF_DIV)
ref_div_max = pll->reference_div;
+ else if (pll->flags & RADEON_PLL_PREFER_MINM_OVER_MAXP)
+ /* fix for problems on RS880 */
+ ref_div_max = min(pll->max_ref_div, 7u);
else
ref_div_max = pll->max_ref_div;
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index 84146d5..5450fa9 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -165,6 +165,19 @@ int radeon_gart_table_vram_pin(struct radeon_device *rdev)
radeon_bo_unpin(rdev->gart.robj);
radeon_bo_unreserve(rdev->gart.robj);
rdev->gart.table_addr = gpu_addr;
+
+ if (!r) {
+ int i;
+
+ /* We might have dropped some GART table updates while it wasn't
+ * mapped, restore all entries
+ */
+ for (i = 0; i < rdev->gart.num_gpu_pages; i++)
+ radeon_gart_set_page(rdev, i, rdev->gart.pages_entry[i]);
+ mb();
+ radeon_gart_tlb_flush(rdev);
+ }
+
return r;
}
@@ -228,7 +241,6 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
unsigned t;
unsigned p;
int i, j;
- u64 page_base;
if (!rdev->gart.ready) {
WARN(1, "trying to unbind memory from uninitialized GART !\n");
@@ -239,14 +251,12 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
for (i = 0; i < pages; i++, p++) {
if (rdev->gart.pages[p]) {
rdev->gart.pages[p] = NULL;
- rdev->gart.pages_addr[p] = rdev->dummy_page.addr;
- page_base = rdev->gart.pages_addr[p];
for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
+ rdev->gart.pages_entry[t] = rdev->dummy_page.entry;
if (rdev->gart.ptr) {
- radeon_gart_set_page(rdev, t, page_base,
- RADEON_GART_PAGE_DUMMY);
+ radeon_gart_set_page(rdev, t,
+ rdev->dummy_page.entry);
}
- page_base += RADEON_GPU_PAGE_SIZE;
}
}
}
@@ -274,7 +284,7 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
{
unsigned t;
unsigned p;
- uint64_t page_base;
+ uint64_t page_base, page_entry;
int i, j;
if (!rdev->gart.ready) {
@@ -285,14 +295,15 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
p = t / (PAGE_SIZE / RADEON_GPU_PAGE_SIZE);
for (i = 0; i < pages; i++, p++) {
- rdev->gart.pages_addr[p] = dma_addr[i];
rdev->gart.pages[p] = pagelist[i];
- if (rdev->gart.ptr) {
- page_base = rdev->gart.pages_addr[p];
- for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
- radeon_gart_set_page(rdev, t, page_base, flags);
- page_base += RADEON_GPU_PAGE_SIZE;
+ page_base = dma_addr[i];
+ for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
+ page_entry = radeon_gart_get_page_entry(page_base, flags);
+ rdev->gart.pages_entry[t] = page_entry;
+ if (rdev->gart.ptr) {
+ radeon_gart_set_page(rdev, t, page_entry);
}
+ page_base += RADEON_GPU_PAGE_SIZE;
}
}
mb();
@@ -334,16 +345,15 @@ int radeon_gart_init(struct radeon_device *rdev)
radeon_gart_fini(rdev);
return -ENOMEM;
}
- rdev->gart.pages_addr = vzalloc(sizeof(dma_addr_t) *
- rdev->gart.num_cpu_pages);
- if (rdev->gart.pages_addr == NULL) {
+ rdev->gart.pages_entry = vmalloc(sizeof(uint64_t) *
+ rdev->gart.num_gpu_pages);
+ if (rdev->gart.pages_entry == NULL) {
radeon_gart_fini(rdev);
return -ENOMEM;
}
/* set GART entry to point to the dummy page by default */
- for (i = 0; i < rdev->gart.num_cpu_pages; i++) {
- rdev->gart.pages_addr[i] = rdev->dummy_page.addr;
- }
+ for (i = 0; i < rdev->gart.num_gpu_pages; i++)
+ rdev->gart.pages_entry[i] = rdev->dummy_page.entry;
return 0;
}
@@ -356,15 +366,15 @@ int radeon_gart_init(struct radeon_device *rdev)
*/
void radeon_gart_fini(struct radeon_device *rdev)
{
- if (rdev->gart.pages && rdev->gart.pages_addr && rdev->gart.ready) {
+ if (rdev->gart.ready) {
/* unbind pages */
radeon_gart_unbind(rdev, 0, rdev->gart.num_cpu_pages);
}
rdev->gart.ready = false;
vfree(rdev->gart.pages);
- vfree(rdev->gart.pages_addr);
+ vfree(rdev->gart.pages_entry);
rdev->gart.pages = NULL;
- rdev->gart.pages_addr = NULL;
+ rdev->gart.pages_entry = NULL;
radeon_dummy_page_fini(rdev);
}
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index a46f737..ac3c131 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -146,7 +146,8 @@ int radeon_gem_object_open(struct drm_gem_object *obj, struct drm_file *file_pri
struct radeon_bo_va *bo_va;
int r;
- if (rdev->family < CHIP_CAYMAN) {
+ if ((rdev->family < CHIP_CAYMAN) ||
+ (!rdev->accel_working)) {
return 0;
}
@@ -176,7 +177,8 @@ void radeon_gem_object_close(struct drm_gem_object *obj,
struct radeon_bo_va *bo_va;
int r;
- if (rdev->family < CHIP_CAYMAN) {
+ if ((rdev->family < CHIP_CAYMAN) ||
+ (!rdev->accel_working)) {
return;
}
@@ -576,7 +578,7 @@ error_unreserve:
error_free:
drm_free_large(vm_bos);
- if (r)
+ if (r && r != -ERESTARTSYS)
DRM_ERROR("Couldn't update BO_VA (%d)\n", r);
}
diff --git a/drivers/gpu/drm/radeon/radeon_kfd.c b/drivers/gpu/drm/radeon/radeon_kfd.c
index 8bf87f1..bef9a09 100644
--- a/drivers/gpu/drm/radeon/radeon_kfd.c
+++ b/drivers/gpu/drm/radeon/radeon_kfd.c
@@ -436,7 +436,7 @@ static int kgd_init_memory(struct kgd_dev *kgd)
static int kgd_init_pipeline(struct kgd_dev *kgd, uint32_t pipe_id,
uint32_t hpd_size, uint64_t hpd_gpu_addr)
{
- uint32_t mec = (++pipe_id / CIK_PIPE_PER_MEC) + 1;
+ uint32_t mec = (pipe_id / CIK_PIPE_PER_MEC) + 1;
uint32_t pipe = (pipe_id % CIK_PIPE_PER_MEC);
lock_srbm(kgd, mec, pipe, 0, 0);
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 3cf9c1f..686411e 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -605,14 +605,14 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
return -ENOMEM;
}
- vm = &fpriv->vm;
- r = radeon_vm_init(rdev, vm);
- if (r) {
- kfree(fpriv);
- return r;
- }
-
if (rdev->accel_working) {
+ vm = &fpriv->vm;
+ r = radeon_vm_init(rdev, vm);
+ if (r) {
+ kfree(fpriv);
+ return r;
+ }
+
r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
if (r) {
radeon_vm_fini(rdev, vm);
@@ -668,9 +668,9 @@ void radeon_driver_postclose_kms(struct drm_device *dev,
radeon_vm_bo_rmv(rdev, vm->ib_bo_va);
radeon_bo_unreserve(rdev->ring_tmp_bo.bo);
}
+ radeon_vm_fini(rdev, vm);
}
- radeon_vm_fini(rdev, vm);
kfree(fpriv);
file_priv->driver_priv = NULL;
}
diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c
index 32522cc..f7da8fe 100644
--- a/drivers/gpu/drm/radeon/radeon_pm.c
+++ b/drivers/gpu/drm/radeon/radeon_pm.c
@@ -1287,8 +1287,39 @@ dpm_failed:
return ret;
}
+struct radeon_dpm_quirk {
+ u32 chip_vendor;
+ u32 chip_device;
+ u32 subsys_vendor;
+ u32 subsys_device;
+};
+
+/* cards with dpm stability problems */
+static struct radeon_dpm_quirk radeon_dpm_quirk_list[] = {
+ /* TURKS - https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1386534 */
+ { PCI_VENDOR_ID_ATI, 0x6759, 0x1682, 0x3195 },
+ /* TURKS - https://bugzilla.kernel.org/show_bug.cgi?id=83731 */
+ { PCI_VENDOR_ID_ATI, 0x6840, 0x1179, 0xfb81 },
+ { 0, 0, 0, 0 },
+};
+
int radeon_pm_init(struct radeon_device *rdev)
{
+ struct radeon_dpm_quirk *p = radeon_dpm_quirk_list;
+ bool disable_dpm = false;
+
+ /* Apply dpm quirks */
+ while (p && p->chip_device != 0) {
+ if (rdev->pdev->vendor == p->chip_vendor &&
+ rdev->pdev->device == p->chip_device &&
+ rdev->pdev->subsystem_vendor == p->subsys_vendor &&
+ rdev->pdev->subsystem_device == p->subsys_device) {
+ disable_dpm = true;
+ break;
+ }
+ ++p;
+ }
+
/* enable dpm on rv6xx+ */
switch (rdev->family) {
case CHIP_RV610:
@@ -1344,6 +1375,8 @@ int radeon_pm_init(struct radeon_device *rdev)
(!(rdev->flags & RADEON_IS_IGP)) &&
(!rdev->smc_fw))
rdev->pm.pm_method = PM_METHOD_PROFILE;
+ else if (disable_dpm && (radeon_dpm == -1))
+ rdev->pm.pm_method = PM_METHOD_PROFILE;
else if (radeon_dpm == 0)
rdev->pm.pm_method = PM_METHOD_PROFILE;
else
diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c
index 07b506b..79181816 100644
--- a/drivers/gpu/drm/radeon/radeon_test.c
+++ b/drivers/gpu/drm/radeon/radeon_test.c
@@ -119,11 +119,11 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
if (ring == R600_RING_TYPE_DMA_INDEX)
fence = radeon_copy_dma(rdev, gtt_addr, vram_addr,
size / RADEON_GPU_PAGE_SIZE,
- NULL);
+ vram_obj->tbo.resv);
else
fence = radeon_copy_blit(rdev, gtt_addr, vram_addr,
size / RADEON_GPU_PAGE_SIZE,
- NULL);
+ vram_obj->tbo.resv);
if (IS_ERR(fence)) {
DRM_ERROR("Failed GTT->VRAM copy %d\n", i);
r = PTR_ERR(fence);
@@ -170,11 +170,11 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
if (ring == R600_RING_TYPE_DMA_INDEX)
fence = radeon_copy_dma(rdev, vram_addr, gtt_addr,
size / RADEON_GPU_PAGE_SIZE,
- NULL);
+ vram_obj->tbo.resv);
else
fence = radeon_copy_blit(rdev, vram_addr, gtt_addr,
size / RADEON_GPU_PAGE_SIZE,
- NULL);
+ vram_obj->tbo.resv);
if (IS_ERR(fence)) {
DRM_ERROR("Failed VRAM->GTT copy %d\n", i);
r = PTR_ERR(fence);
diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c
index cde48c4..2a5a4a9 100644
--- a/drivers/gpu/drm/radeon/radeon_vm.c
+++ b/drivers/gpu/drm/radeon/radeon_vm.c
@@ -587,10 +587,8 @@ uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr)
uint64_t result;
/* page table offset */
- result = rdev->gart.pages_addr[addr >> PAGE_SHIFT];
-
- /* in case cpu page size != gpu page size*/
- result |= addr & (~PAGE_MASK);
+ result = rdev->gart.pages_entry[addr >> RADEON_GPU_PAGE_SHIFT];
+ result &= ~RADEON_GPU_PAGE_MASK;
return result;
}
@@ -745,9 +743,11 @@ static void radeon_vm_frag_ptes(struct radeon_device *rdev,
*/
/* NI is optimized for 256KB fragments, SI and newer for 64KB */
- uint64_t frag_flags = rdev->family == CHIP_CAYMAN ?
+ uint64_t frag_flags = ((rdev->family == CHIP_CAYMAN) ||
+ (rdev->family == CHIP_ARUBA)) ?
R600_PTE_FRAG_256KB : R600_PTE_FRAG_64KB;
- uint64_t frag_align = rdev->family == CHIP_CAYMAN ? 0x200 : 0x80;
+ uint64_t frag_align = ((rdev->family == CHIP_CAYMAN) ||
+ (rdev->family == CHIP_ARUBA)) ? 0x200 : 0x80;
uint64_t frag_start = ALIGN(pe_start, frag_align);
uint64_t frag_end = pe_end & ~(frag_align - 1);
diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c
index c5799f16..34e3235 100644
--- a/drivers/gpu/drm/radeon/rs400.c
+++ b/drivers/gpu/drm/radeon/rs400.c
@@ -212,11 +212,9 @@ void rs400_gart_fini(struct radeon_device *rdev)
#define RS400_PTE_WRITEABLE (1 << 2)
#define RS400_PTE_READABLE (1 << 3)
-void rs400_gart_set_page(struct radeon_device *rdev, unsigned i,
- uint64_t addr, uint32_t flags)
+uint64_t rs400_gart_get_page_entry(uint64_t addr, uint32_t flags)
{
uint32_t entry;
- u32 *gtt = rdev->gart.ptr;
entry = (lower_32_bits(addr) & PAGE_MASK) |
((upper_32_bits(addr) & 0xff) << 4);
@@ -226,8 +224,14 @@ void rs400_gart_set_page(struct radeon_device *rdev, unsigned i,
entry |= RS400_PTE_WRITEABLE;
if (!(flags & RADEON_GART_PAGE_SNOOP))
entry |= RS400_PTE_UNSNOOPED;
- entry = cpu_to_le32(entry);
- gtt[i] = entry;
+ return entry;
+}
+
+void rs400_gart_set_page(struct radeon_device *rdev, unsigned i,
+ uint64_t entry)
+{
+ u32 *gtt = rdev->gart.ptr;
+ gtt[i] = cpu_to_le32(lower_32_bits(entry));
}
int rs400_mc_wait_for_idle(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c
index 9acb1c3..74bce91 100644
--- a/drivers/gpu/drm/radeon/rs600.c
+++ b/drivers/gpu/drm/radeon/rs600.c
@@ -625,11 +625,8 @@ static void rs600_gart_fini(struct radeon_device *rdev)
radeon_gart_table_vram_free(rdev);
}
-void rs600_gart_set_page(struct radeon_device *rdev, unsigned i,
- uint64_t addr, uint32_t flags)
+uint64_t rs600_gart_get_page_entry(uint64_t addr, uint32_t flags)
{
- void __iomem *ptr = (void *)rdev->gart.ptr;
-
addr = addr & 0xFFFFFFFFFFFFF000ULL;
addr |= R600_PTE_SYSTEM;
if (flags & RADEON_GART_PAGE_VALID)
@@ -640,7 +637,14 @@ void rs600_gart_set_page(struct radeon_device *rdev, unsigned i,
addr |= R600_PTE_WRITEABLE;
if (flags & RADEON_GART_PAGE_SNOOP)
addr |= R600_PTE_SNOOPED;
- writeq(addr, ptr + (i * 8));
+ return addr;
+}
+
+void rs600_gart_set_page(struct radeon_device *rdev, unsigned i,
+ uint64_t entry)
+{
+ void __iomem *ptr = (void *)rdev->gart.ptr;
+ writeq(entry, ptr + (i * 8));
}
int rs600_irq_set(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index 60df444..5d89b87 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -5057,6 +5057,16 @@ void si_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
radeon_ring_write(ring, 0);
radeon_ring_write(ring, 1 << vm_id);
+ /* wait for the invalidate to complete */
+ radeon_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));
+ radeon_ring_write(ring, (WAIT_REG_MEM_FUNCTION(0) | /* always */
+ WAIT_REG_MEM_ENGINE(0))); /* me */
+ radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
+ radeon_ring_write(ring, 0);
+ radeon_ring_write(ring, 0); /* ref */
+ radeon_ring_write(ring, 0); /* mask */
+ radeon_ring_write(ring, 0x20); /* poll interval */
+
/* sync PFP to ME, otherwise we might get invalid PFP reads */
radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
radeon_ring_write(ring, 0x0);
diff --git a/drivers/gpu/drm/radeon/si_dma.c b/drivers/gpu/drm/radeon/si_dma.c
index f5cc777..8320792 100644
--- a/drivers/gpu/drm/radeon/si_dma.c
+++ b/drivers/gpu/drm/radeon/si_dma.c
@@ -123,7 +123,6 @@ void si_dma_vm_write_pages(struct radeon_device *rdev,
for (; ndw > 0; ndw -= 2, --count, pe += 8) {
if (flags & R600_PTE_SYSTEM) {
value = radeon_vm_map_gart(rdev, addr);
- value &= 0xFFFFFFFFFFFFF000ULL;
} else if (flags & R600_PTE_VALID) {
value = addr;
} else {
@@ -206,6 +205,14 @@ void si_dma_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
radeon_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST >> 2));
radeon_ring_write(ring, 1 << vm_id);
+
+ /* wait for invalidate to complete */
+ radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_POLL_REG_MEM, 0, 0, 0, 0));
+ radeon_ring_write(ring, VM_INVALIDATE_REQUEST);
+ radeon_ring_write(ring, 0xff << 16); /* retry */
+ radeon_ring_write(ring, 1 << vm_id); /* mask */
+ radeon_ring_write(ring, 0); /* value */
+ radeon_ring_write(ring, (0 << 28) | 0x20); /* func(always) | poll interval */
}
/**
diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c
index 32e354b..eff8a64 100644
--- a/drivers/gpu/drm/radeon/si_dpm.c
+++ b/drivers/gpu/drm/radeon/si_dpm.c
@@ -2908,6 +2908,22 @@ static int si_init_smc_spll_table(struct radeon_device *rdev)
return ret;
}
+struct si_dpm_quirk {
+ u32 chip_vendor;
+ u32 chip_device;
+ u32 subsys_vendor;
+ u32 subsys_device;
+ u32 max_sclk;
+ u32 max_mclk;
+};
+
+/* cards with dpm stability problems */
+static struct si_dpm_quirk si_dpm_quirk_list[] = {
+ /* PITCAIRN - https://bugs.freedesktop.org/show_bug.cgi?id=76490 */
+ { PCI_VENDOR_ID_ATI, 0x6810, 0x1462, 0x3036, 0, 120000 },
+ { 0, 0, 0, 0 },
+};
+
static void si_apply_state_adjust_rules(struct radeon_device *rdev,
struct radeon_ps *rps)
{
@@ -2918,7 +2934,22 @@ static void si_apply_state_adjust_rules(struct radeon_device *rdev,
u32 mclk, sclk;
u16 vddc, vddci;
u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc;
+ u32 max_sclk = 0, max_mclk = 0;
int i;
+ struct si_dpm_quirk *p = si_dpm_quirk_list;
+
+ /* Apply dpm quirks */
+ while (p && p->chip_device != 0) {
+ if (rdev->pdev->vendor == p->chip_vendor &&
+ rdev->pdev->device == p->chip_device &&
+ rdev->pdev->subsystem_vendor == p->subsys_vendor &&
+ rdev->pdev->subsystem_device == p->subsys_device) {
+ max_sclk = p->max_sclk;
+ max_mclk = p->max_mclk;
+ break;
+ }
+ ++p;
+ }
if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
ni_dpm_vblank_too_short(rdev))
@@ -2972,6 +3003,14 @@ static void si_apply_state_adjust_rules(struct radeon_device *rdev,
if (ps->performance_levels[i].mclk > max_mclk_vddc)
ps->performance_levels[i].mclk = max_mclk_vddc;
}
+ if (max_mclk) {
+ if (ps->performance_levels[i].mclk > max_mclk)
+ ps->performance_levels[i].mclk = max_mclk;
+ }
+ if (max_sclk) {
+ if (ps->performance_levels[i].sclk > max_sclk)
+ ps->performance_levels[i].sclk = max_sclk;
+ }
}
/* XXX validate the min clocks required for display */
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
index 4069be89..8499924 100644
--- a/drivers/gpu/drm/radeon/sid.h
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -1632,6 +1632,23 @@
#define PACKET3_MPEG_INDEX 0x3A
#define PACKET3_COPY_DW 0x3B
#define PACKET3_WAIT_REG_MEM 0x3C
+#define WAIT_REG_MEM_FUNCTION(x) ((x) << 0)
+ /* 0 - always
+ * 1 - <
+ * 2 - <=
+ * 3 - ==
+ * 4 - !=
+ * 5 - >=
+ * 6 - >
+ */
+#define WAIT_REG_MEM_MEM_SPACE(x) ((x) << 4)
+ /* 0 - reg
+ * 1 - mem
+ */
+#define WAIT_REG_MEM_ENGINE(x) ((x) << 8)
+ /* 0 - me
+ * 1 - pfp
+ */
#define PACKET3_MEM_WRITE 0x3D
#define PACKET3_COPY_DATA 0x40
#define PACKET3_CP_DMA 0x41
@@ -1835,6 +1852,7 @@
#define DMA_PACKET_TRAP 0x7
#define DMA_PACKET_SRBM_WRITE 0x9
#define DMA_PACKET_CONSTANT_FILL 0xd
+#define DMA_PACKET_POLL_REG_MEM 0xe
#define DMA_PACKET_NOP 0xf
#define VCE_STATUS 0x20004
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 7b5d221..6c6b655 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -406,11 +406,9 @@ int vmw_3d_resource_inc(struct vmw_private *dev_priv,
if (unlikely(ret != 0))
--dev_priv->num_3d_resources;
} else if (unhide_svga) {
- mutex_lock(&dev_priv->hw_mutex);
vmw_write(dev_priv, SVGA_REG_ENABLE,
vmw_read(dev_priv, SVGA_REG_ENABLE) &
~SVGA_REG_ENABLE_HIDE);
- mutex_unlock(&dev_priv->hw_mutex);
}
mutex_unlock(&dev_priv->release_mutex);
@@ -433,13 +431,10 @@ void vmw_3d_resource_dec(struct vmw_private *dev_priv,
mutex_lock(&dev_priv->release_mutex);
if (unlikely(--dev_priv->num_3d_resources == 0))
vmw_release_device(dev_priv);
- else if (hide_svga) {
- mutex_lock(&dev_priv->hw_mutex);
+ else if (hide_svga)
vmw_write(dev_priv, SVGA_REG_ENABLE,
vmw_read(dev_priv, SVGA_REG_ENABLE) |
SVGA_REG_ENABLE_HIDE);
- mutex_unlock(&dev_priv->hw_mutex);
- }
n3d = (int32_t) dev_priv->num_3d_resources;
mutex_unlock(&dev_priv->release_mutex);
@@ -600,12 +595,14 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
dev_priv->dev = dev;
dev_priv->vmw_chipset = chipset;
dev_priv->last_read_seqno = (uint32_t) -100;
- mutex_init(&dev_priv->hw_mutex);
mutex_init(&dev_priv->cmdbuf_mutex);
mutex_init(&dev_priv->release_mutex);
mutex_init(&dev_priv->binding_mutex);
rwlock_init(&dev_priv->resource_lock);
ttm_lock_init(&dev_priv->reservation_sem);
+ spin_lock_init(&dev_priv->hw_lock);
+ spin_lock_init(&dev_priv->waiter_lock);
+ spin_lock_init(&dev_priv->cap_lock);
for (i = vmw_res_context; i < vmw_res_max; ++i) {
idr_init(&dev_priv->res_idr[i]);
@@ -626,14 +623,11 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
dev_priv->enable_fb = enable_fbdev;
- mutex_lock(&dev_priv->hw_mutex);
-
vmw_write(dev_priv, SVGA_REG_ID, SVGA_ID_2);
svga_id = vmw_read(dev_priv, SVGA_REG_ID);
if (svga_id != SVGA_ID_2) {
ret = -ENOSYS;
DRM_ERROR("Unsupported SVGA ID 0x%x\n", svga_id);
- mutex_unlock(&dev_priv->hw_mutex);
goto out_err0;
}
@@ -683,10 +677,8 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
dev_priv->prim_bb_mem = dev_priv->vram_size;
ret = vmw_dma_masks(dev_priv);
- if (unlikely(ret != 0)) {
- mutex_unlock(&dev_priv->hw_mutex);
+ if (unlikely(ret != 0))
goto out_err0;
- }
/*
* Limit back buffer size to VRAM size. Remove this once
@@ -695,8 +687,6 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
if (dev_priv->prim_bb_mem > dev_priv->vram_size)
dev_priv->prim_bb_mem = dev_priv->vram_size;
- mutex_unlock(&dev_priv->hw_mutex);
-
vmw_print_capabilities(dev_priv->capabilities);
if (dev_priv->capabilities & SVGA_CAP_GMR2) {
@@ -1160,9 +1150,7 @@ static int vmw_master_set(struct drm_device *dev,
if (unlikely(ret != 0))
return ret;
vmw_kms_save_vga(dev_priv);
- mutex_lock(&dev_priv->hw_mutex);
vmw_write(dev_priv, SVGA_REG_TRACES, 0);
- mutex_unlock(&dev_priv->hw_mutex);
}
if (active) {
@@ -1196,9 +1184,7 @@ out_no_active_lock:
if (!dev_priv->enable_fb) {
vmw_kms_restore_vga(dev_priv);
vmw_3d_resource_dec(dev_priv, true);
- mutex_lock(&dev_priv->hw_mutex);
vmw_write(dev_priv, SVGA_REG_TRACES, 1);
- mutex_unlock(&dev_priv->hw_mutex);
}
return ret;
}
@@ -1233,9 +1219,7 @@ static void vmw_master_drop(struct drm_device *dev,
DRM_ERROR("Unable to clean VRAM on master drop.\n");
vmw_kms_restore_vga(dev_priv);
vmw_3d_resource_dec(dev_priv, true);
- mutex_lock(&dev_priv->hw_mutex);
vmw_write(dev_priv, SVGA_REG_TRACES, 1);
- mutex_unlock(&dev_priv->hw_mutex);
}
dev_priv->active_master = &dev_priv->fbdev_master;
@@ -1367,10 +1351,8 @@ static void vmw_pm_complete(struct device *kdev)
struct drm_device *dev = pci_get_drvdata(pdev);
struct vmw_private *dev_priv = vmw_priv(dev);
- mutex_lock(&dev_priv->hw_mutex);
vmw_write(dev_priv, SVGA_REG_ID, SVGA_ID_2);
(void) vmw_read(dev_priv, SVGA_REG_ID);
- mutex_unlock(&dev_priv->hw_mutex);
/**
* Reclaim 3d reference held by fbdev and potentially
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index 4ee799b..d26a6da 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -399,7 +399,8 @@ struct vmw_private {
uint32_t memory_size;
bool has_gmr;
bool has_mob;
- struct mutex hw_mutex;
+ spinlock_t hw_lock;
+ spinlock_t cap_lock;
/*
* VGA registers.
@@ -449,8 +450,9 @@ struct vmw_private {
atomic_t marker_seq;
wait_queue_head_t fence_queue;
wait_queue_head_t fifo_queue;
- int fence_queue_waiters; /* Protected by hw_mutex */
- int goal_queue_waiters; /* Protected by hw_mutex */
+ spinlock_t waiter_lock;
+ int fence_queue_waiters; /* Protected by waiter_lock */
+ int goal_queue_waiters; /* Protected by waiter_lock */
atomic_t fifo_queue_waiters;
uint32_t last_read_seqno;
spinlock_t irq_lock;
@@ -553,20 +555,35 @@ static inline struct vmw_master *vmw_master(struct drm_master *master)
return (struct vmw_master *) master->driver_priv;
}
+/*
+ * The locking here is fine-grained, so that it is performed once
+ * for every read- and write operation. This is of course costly, but we
+ * don't perform much register access in the timing critical paths anyway.
+ * Instead we have the extra benefit of being sure that we don't forget
+ * the hw lock around register accesses.
+ */
static inline void vmw_write(struct vmw_private *dev_priv,
unsigned int offset, uint32_t value)
{
+ unsigned long irq_flags;
+
+ spin_lock_irqsave(&dev_priv->hw_lock, irq_flags);
outl(offset, dev_priv->io_start + VMWGFX_INDEX_PORT);
outl(value, dev_priv->io_start + VMWGFX_VALUE_PORT);
+ spin_unlock_irqrestore(&dev_priv->hw_lock, irq_flags);
}
static inline uint32_t vmw_read(struct vmw_private *dev_priv,
unsigned int offset)
{
- uint32_t val;
+ unsigned long irq_flags;
+ u32 val;
+ spin_lock_irqsave(&dev_priv->hw_lock, irq_flags);
outl(offset, dev_priv->io_start + VMWGFX_INDEX_PORT);
val = inl(dev_priv->io_start + VMWGFX_VALUE_PORT);
+ spin_unlock_irqrestore(&dev_priv->hw_lock, irq_flags);
+
return val;
}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
index b7594cb..945f1e0 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
@@ -35,7 +35,7 @@ struct vmw_fence_manager {
struct vmw_private *dev_priv;
spinlock_t lock;
struct list_head fence_list;
- struct work_struct work, ping_work;
+ struct work_struct work;
u32 user_fence_size;
u32 fence_size;
u32 event_fence_action_size;
@@ -134,14 +134,6 @@ static const char *vmw_fence_get_timeline_name(struct fence *f)
return "svga";
}
-static void vmw_fence_ping_func(struct work_struct *work)
-{
- struct vmw_fence_manager *fman =
- container_of(work, struct vmw_fence_manager, ping_work);
-
- vmw_fifo_ping_host(fman->dev_priv, SVGA_SYNC_GENERIC);
-}
-
static bool vmw_fence_enable_signaling(struct fence *f)
{
struct vmw_fence_obj *fence =
@@ -155,11 +147,7 @@ static bool vmw_fence_enable_signaling(struct fence *f)
if (seqno - fence->base.seqno < VMW_FENCE_WRAP)
return false;
- if (mutex_trylock(&dev_priv->hw_mutex)) {
- vmw_fifo_ping_host_locked(dev_priv, SVGA_SYNC_GENERIC);
- mutex_unlock(&dev_priv->hw_mutex);
- } else
- schedule_work(&fman->ping_work);
+ vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
return true;
}
@@ -305,7 +293,6 @@ struct vmw_fence_manager *vmw_fence_manager_init(struct vmw_private *dev_priv)
INIT_LIST_HEAD(&fman->fence_list);
INIT_LIST_HEAD(&fman->cleanup_list);
INIT_WORK(&fman->work, &vmw_fence_work_func);
- INIT_WORK(&fman->ping_work, &vmw_fence_ping_func);
fman->fifo_down = true;
fman->user_fence_size = ttm_round_pot(sizeof(struct vmw_user_fence));
fman->fence_size = ttm_round_pot(sizeof(struct vmw_fence_obj));
@@ -323,7 +310,6 @@ void vmw_fence_manager_takedown(struct vmw_fence_manager *fman)
bool lists_empty;
(void) cancel_work_sync(&fman->work);
- (void) cancel_work_sync(&fman->ping_work);
spin_lock_irqsave(&fman->lock, irq_flags);
lists_empty = list_empty(&fman->fence_list) &&
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
index 09e10ae..39f2b03 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
@@ -44,10 +44,10 @@ bool vmw_fifo_have_3d(struct vmw_private *dev_priv)
if (!dev_priv->has_mob)
return false;
- mutex_lock(&dev_priv->hw_mutex);
+ spin_lock(&dev_priv->cap_lock);
vmw_write(dev_priv, SVGA_REG_DEV_CAP, SVGA3D_DEVCAP_3D);
result = vmw_read(dev_priv, SVGA_REG_DEV_CAP);
- mutex_unlock(&dev_priv->hw_mutex);
+ spin_unlock(&dev_priv->cap_lock);
return (result != 0);
}
@@ -120,7 +120,6 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
DRM_INFO("height %d\n", vmw_read(dev_priv, SVGA_REG_HEIGHT));
DRM_INFO("bpp %d\n", vmw_read(dev_priv, SVGA_REG_BITS_PER_PIXEL));
- mutex_lock(&dev_priv->hw_mutex);
dev_priv->enable_state = vmw_read(dev_priv, SVGA_REG_ENABLE);
dev_priv->config_done_state = vmw_read(dev_priv, SVGA_REG_CONFIG_DONE);
dev_priv->traces_state = vmw_read(dev_priv, SVGA_REG_TRACES);
@@ -143,7 +142,6 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
mb();
vmw_write(dev_priv, SVGA_REG_CONFIG_DONE, 1);
- mutex_unlock(&dev_priv->hw_mutex);
max = ioread32(fifo_mem + SVGA_FIFO_MAX);
min = ioread32(fifo_mem + SVGA_FIFO_MIN);
@@ -160,31 +158,28 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
return vmw_fifo_send_fence(dev_priv, &dummy);
}
-void vmw_fifo_ping_host_locked(struct vmw_private *dev_priv, uint32_t reason)
+void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason)
{
__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
+ static DEFINE_SPINLOCK(ping_lock);
+ unsigned long irq_flags;
+ /*
+ * The ping_lock is needed because we don't have an atomic
+ * test-and-set of the SVGA_FIFO_BUSY register.
+ */
+ spin_lock_irqsave(&ping_lock, irq_flags);
if (unlikely(ioread32(fifo_mem + SVGA_FIFO_BUSY) == 0)) {
iowrite32(1, fifo_mem + SVGA_FIFO_BUSY);
vmw_write(dev_priv, SVGA_REG_SYNC, reason);
}
-}
-
-void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason)
-{
- mutex_lock(&dev_priv->hw_mutex);
-
- vmw_fifo_ping_host_locked(dev_priv, reason);
-
- mutex_unlock(&dev_priv->hw_mutex);
+ spin_unlock_irqrestore(&ping_lock, irq_flags);
}
void vmw_fifo_release(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
{
__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
- mutex_lock(&dev_priv->hw_mutex);
-
vmw_write(dev_priv, SVGA_REG_SYNC, SVGA_SYNC_GENERIC);
while (vmw_read(dev_priv, SVGA_REG_BUSY) != 0)
;
@@ -198,7 +193,6 @@ void vmw_fifo_release(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
vmw_write(dev_priv, SVGA_REG_TRACES,
dev_priv->traces_state);
- mutex_unlock(&dev_priv->hw_mutex);
vmw_marker_queue_takedown(&fifo->marker_queue);
if (likely(fifo->static_buffer != NULL)) {
@@ -271,7 +265,7 @@ static int vmw_fifo_wait(struct vmw_private *dev_priv,
return vmw_fifo_wait_noirq(dev_priv, bytes,
interruptible, timeout);
- mutex_lock(&dev_priv->hw_mutex);
+ spin_lock(&dev_priv->waiter_lock);
if (atomic_add_return(1, &dev_priv->fifo_queue_waiters) > 0) {
spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
outl(SVGA_IRQFLAG_FIFO_PROGRESS,
@@ -280,7 +274,7 @@ static int vmw_fifo_wait(struct vmw_private *dev_priv,
vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
}
- mutex_unlock(&dev_priv->hw_mutex);
+ spin_unlock(&dev_priv->waiter_lock);
if (interruptible)
ret = wait_event_interruptible_timeout
@@ -296,14 +290,14 @@ static int vmw_fifo_wait(struct vmw_private *dev_priv,
else if (likely(ret > 0))
ret = 0;
- mutex_lock(&dev_priv->hw_mutex);
+ spin_lock(&dev_priv->waiter_lock);
if (atomic_dec_and_test(&dev_priv->fifo_queue_waiters)) {
spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
dev_priv->irq_mask &= ~SVGA_IRQFLAG_FIFO_PROGRESS;
vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
}
- mutex_unlock(&dev_priv->hw_mutex);
+ spin_unlock(&dev_priv->waiter_lock);
return ret;
}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
index 37881ec..69c8ce23 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
@@ -135,13 +135,13 @@ static int vmw_fill_compat_cap(struct vmw_private *dev_priv, void *bounce,
(pair_offset + max_size * sizeof(SVGA3dCapPair)) / sizeof(u32);
compat_cap->header.type = SVGA3DCAPS_RECORD_DEVCAPS;
- mutex_lock(&dev_priv->hw_mutex);
+ spin_lock(&dev_priv->cap_lock);
for (i = 0; i < max_size; ++i) {
vmw_write(dev_priv, SVGA_REG_DEV_CAP, i);
compat_cap->pairs[i][0] = i;
compat_cap->pairs[i][1] = vmw_read(dev_priv, SVGA_REG_DEV_CAP);
}
- mutex_unlock(&dev_priv->hw_mutex);
+ spin_unlock(&dev_priv->cap_lock);
return 0;
}
@@ -191,12 +191,12 @@ int vmw_get_cap_3d_ioctl(struct drm_device *dev, void *data,
if (num > SVGA3D_DEVCAP_MAX)
num = SVGA3D_DEVCAP_MAX;
- mutex_lock(&dev_priv->hw_mutex);
+ spin_lock(&dev_priv->cap_lock);
for (i = 0; i < num; ++i) {
vmw_write(dev_priv, SVGA_REG_DEV_CAP, i);
*bounce32++ = vmw_read(dev_priv, SVGA_REG_DEV_CAP);
}
- mutex_unlock(&dev_priv->hw_mutex);
+ spin_unlock(&dev_priv->cap_lock);
} else if (gb_objects) {
ret = vmw_fill_compat_cap(dev_priv, bounce, size);
if (unlikely(ret != 0))
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
index 0c42376..9fe9827 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
@@ -62,13 +62,8 @@ irqreturn_t vmw_irq_handler(int irq, void *arg)
static bool vmw_fifo_idle(struct vmw_private *dev_priv, uint32_t seqno)
{
- uint32_t busy;
- mutex_lock(&dev_priv->hw_mutex);
- busy = vmw_read(dev_priv, SVGA_REG_BUSY);
- mutex_unlock(&dev_priv->hw_mutex);
-
- return (busy == 0);
+ return (vmw_read(dev_priv, SVGA_REG_BUSY) == 0);
}
void vmw_update_seqno(struct vmw_private *dev_priv,
@@ -184,7 +179,7 @@ int vmw_fallback_wait(struct vmw_private *dev_priv,
void vmw_seqno_waiter_add(struct vmw_private *dev_priv)
{
- mutex_lock(&dev_priv->hw_mutex);
+ spin_lock(&dev_priv->waiter_lock);
if (dev_priv->fence_queue_waiters++ == 0) {
unsigned long irq_flags;
@@ -195,12 +190,12 @@ void vmw_seqno_waiter_add(struct vmw_private *dev_priv)
vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
}
- mutex_unlock(&dev_priv->hw_mutex);
+ spin_unlock(&dev_priv->waiter_lock);
}
void vmw_seqno_waiter_remove(struct vmw_private *dev_priv)
{
- mutex_lock(&dev_priv->hw_mutex);
+ spin_lock(&dev_priv->waiter_lock);
if (--dev_priv->fence_queue_waiters == 0) {
unsigned long irq_flags;
@@ -209,13 +204,13 @@ void vmw_seqno_waiter_remove(struct vmw_private *dev_priv)
vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
}
- mutex_unlock(&dev_priv->hw_mutex);
+ spin_unlock(&dev_priv->waiter_lock);
}
void vmw_goal_waiter_add(struct vmw_private *dev_priv)
{
- mutex_lock(&dev_priv->hw_mutex);
+ spin_lock(&dev_priv->waiter_lock);
if (dev_priv->goal_queue_waiters++ == 0) {
unsigned long irq_flags;
@@ -226,12 +221,12 @@ void vmw_goal_waiter_add(struct vmw_private *dev_priv)
vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
}
- mutex_unlock(&dev_priv->hw_mutex);
+ spin_unlock(&dev_priv->waiter_lock);
}
void vmw_goal_waiter_remove(struct vmw_private *dev_priv)
{
- mutex_lock(&dev_priv->hw_mutex);
+ spin_lock(&dev_priv->waiter_lock);
if (--dev_priv->goal_queue_waiters == 0) {
unsigned long irq_flags;
@@ -240,7 +235,7 @@ void vmw_goal_waiter_remove(struct vmw_private *dev_priv)
vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
}
- mutex_unlock(&dev_priv->hw_mutex);
+ spin_unlock(&dev_priv->waiter_lock);
}
int vmw_wait_seqno(struct vmw_private *dev_priv,
@@ -315,9 +310,7 @@ void vmw_irq_uninstall(struct drm_device *dev)
if (!(dev_priv->capabilities & SVGA_CAP_IRQMASK))
return;
- mutex_lock(&dev_priv->hw_mutex);
vmw_write(dev_priv, SVGA_REG_IRQMASK, 0);
- mutex_unlock(&dev_priv->hw_mutex);
status = inl(dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
outl(status, dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index 3725b52..8725b79 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -1828,9 +1828,7 @@ vmw_du_connector_detect(struct drm_connector *connector, bool force)
struct vmw_private *dev_priv = vmw_priv(dev);
struct vmw_display_unit *du = vmw_connector_to_du(connector);
- mutex_lock(&dev_priv->hw_mutex);
num_displays = vmw_read(dev_priv, SVGA_REG_NUM_DISPLAYS);
- mutex_unlock(&dev_priv->hw_mutex);
return ((vmw_connector_to_du(connector)->unit < num_displays &&
du->pref_active) ?