linux/drivers/gpu/drm/vc4/vc4_irq.c
Varad Gautam ca26d28bba drm/vc4: improve throughput by pipelining binning and rendering jobs
The hardware provides us with separate threads for binning and
rendering, and the existing model waits for them both to complete
before submitting the next job.

Splitting the binning and rendering submissions reduces idle time and
gives us approx 20-30% speedup with some x11perf tests such as -line10
and -tilerect1.  Improves openarena performance by 1.01897% +/-
0.247857% (n=16).

Thanks to anholt for suggesting this.

v2: Rebase on the spurious resets fix (change by anholt).

Signed-off-by: Varad Gautam <varadgautam@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Eric Anholt <eric@anholt.net>
2016-03-13 17:05:05 -07:00

251 lines
7.0 KiB
C

/*
* Copyright © 2014 Broadcom
*
* 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 (including the next
* paragraph) 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 AUTHORS OR COPYRIGHT HOLDERS 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.
*/
/** DOC: Interrupt management for the V3D engine.
*
* We have an interrupt status register (V3D_INTCTL) which reports
* interrupts, and where writing 1 bits clears those interrupts.
* There are also a pair of interrupt registers
* (V3D_INTENA/V3D_INTDIS) where writing a 1 to their bits enables or
* disables that specific interrupt, and 0s written are ignored
* (reading either one returns the set of enabled interrupts).
*
* When we take a binning flush done interrupt, we need to submit the
* next frame for binning and move the finished frame to the render
* thread.
*
* When we take a render frame interrupt, we need to wake the
* processes waiting for some frame to be done, and get the next frame
* submitted ASAP (so the hardware doesn't sit idle when there's work
* to do).
*
* When we take the binner out of memory interrupt, we need to
* allocate some new memory and pass it to the binner so that the
* current job can make progress.
*/
#include "vc4_drv.h"
#include "vc4_regs.h"
#define V3D_DRIVER_IRQS (V3D_INT_OUTOMEM | \
V3D_INT_FLDONE | \
V3D_INT_FRDONE)
DECLARE_WAIT_QUEUE_HEAD(render_wait);
static void
vc4_overflow_mem_work(struct work_struct *work)
{
struct vc4_dev *vc4 =
container_of(work, struct vc4_dev, overflow_mem_work);
struct drm_device *dev = vc4->dev;
struct vc4_bo *bo;
bo = vc4_bo_create(dev, 256 * 1024, true);
if (IS_ERR(bo)) {
DRM_ERROR("Couldn't allocate binner overflow mem\n");
return;
}
/* If there's a job executing currently, then our previous
* overflow allocation is getting used in that job and we need
* to queue it to be released when the job is done. But if no
* job is executing at all, then we can free the old overflow
* object direcctly.
*
* No lock necessary for this pointer since we're the only
* ones that update the pointer, and our workqueue won't
* reenter.
*/
if (vc4->overflow_mem) {
struct vc4_exec_info *current_exec;
unsigned long irqflags;
spin_lock_irqsave(&vc4->job_lock, irqflags);
current_exec = vc4_first_bin_job(vc4);
if (current_exec) {
vc4->overflow_mem->seqno = vc4->finished_seqno + 1;
list_add_tail(&vc4->overflow_mem->unref_head,
&current_exec->unref_list);
vc4->overflow_mem = NULL;
}
spin_unlock_irqrestore(&vc4->job_lock, irqflags);
}
if (vc4->overflow_mem)
drm_gem_object_unreference_unlocked(&vc4->overflow_mem->base.base);
vc4->overflow_mem = bo;
V3D_WRITE(V3D_BPOA, bo->base.paddr);
V3D_WRITE(V3D_BPOS, bo->base.base.size);
V3D_WRITE(V3D_INTCTL, V3D_INT_OUTOMEM);
V3D_WRITE(V3D_INTENA, V3D_INT_OUTOMEM);
}
static void
vc4_irq_finish_bin_job(struct drm_device *dev)
{
struct vc4_dev *vc4 = to_vc4_dev(dev);
struct vc4_exec_info *exec = vc4_first_bin_job(vc4);
if (!exec)
return;
vc4_move_job_to_render(dev, exec);
vc4_submit_next_bin_job(dev);
}
static void
vc4_cancel_bin_job(struct drm_device *dev)
{
struct vc4_dev *vc4 = to_vc4_dev(dev);
struct vc4_exec_info *exec = vc4_first_bin_job(vc4);
if (!exec)
return;
list_move_tail(&exec->head, &vc4->bin_job_list);
vc4_submit_next_bin_job(dev);
}
static void
vc4_irq_finish_render_job(struct drm_device *dev)
{
struct vc4_dev *vc4 = to_vc4_dev(dev);
struct vc4_exec_info *exec = vc4_first_render_job(vc4);
if (!exec)
return;
vc4->finished_seqno++;
list_move_tail(&exec->head, &vc4->job_done_list);
vc4_submit_next_render_job(dev);
wake_up_all(&vc4->job_wait_queue);
schedule_work(&vc4->job_done_work);
}
irqreturn_t
vc4_irq(int irq, void *arg)
{
struct drm_device *dev = arg;
struct vc4_dev *vc4 = to_vc4_dev(dev);
uint32_t intctl;
irqreturn_t status = IRQ_NONE;
barrier();
intctl = V3D_READ(V3D_INTCTL);
/* Acknowledge the interrupts we're handling here. The binner
* last flush / render frame done interrupt will be cleared,
* while OUTOMEM will stay high until the underlying cause is
* cleared.
*/
V3D_WRITE(V3D_INTCTL, intctl);
if (intctl & V3D_INT_OUTOMEM) {
/* Disable OUTOMEM until the work is done. */
V3D_WRITE(V3D_INTDIS, V3D_INT_OUTOMEM);
schedule_work(&vc4->overflow_mem_work);
status = IRQ_HANDLED;
}
if (intctl & V3D_INT_FLDONE) {
spin_lock(&vc4->job_lock);
vc4_irq_finish_bin_job(dev);
spin_unlock(&vc4->job_lock);
status = IRQ_HANDLED;
}
if (intctl & V3D_INT_FRDONE) {
spin_lock(&vc4->job_lock);
vc4_irq_finish_render_job(dev);
spin_unlock(&vc4->job_lock);
status = IRQ_HANDLED;
}
return status;
}
void
vc4_irq_preinstall(struct drm_device *dev)
{
struct vc4_dev *vc4 = to_vc4_dev(dev);
init_waitqueue_head(&vc4->job_wait_queue);
INIT_WORK(&vc4->overflow_mem_work, vc4_overflow_mem_work);
/* Clear any pending interrupts someone might have left around
* for us.
*/
V3D_WRITE(V3D_INTCTL, V3D_DRIVER_IRQS);
}
int
vc4_irq_postinstall(struct drm_device *dev)
{
struct vc4_dev *vc4 = to_vc4_dev(dev);
/* Enable both the render done and out of memory interrupts. */
V3D_WRITE(V3D_INTENA, V3D_DRIVER_IRQS);
return 0;
}
void
vc4_irq_uninstall(struct drm_device *dev)
{
struct vc4_dev *vc4 = to_vc4_dev(dev);
/* Disable sending interrupts for our driver's IRQs. */
V3D_WRITE(V3D_INTDIS, V3D_DRIVER_IRQS);
/* Clear any pending interrupts we might have left. */
V3D_WRITE(V3D_INTCTL, V3D_DRIVER_IRQS);
cancel_work_sync(&vc4->overflow_mem_work);
}
/** Reinitializes interrupt registers when a GPU reset is performed. */
void vc4_irq_reset(struct drm_device *dev)
{
struct vc4_dev *vc4 = to_vc4_dev(dev);
unsigned long irqflags;
/* Acknowledge any stale IRQs. */
V3D_WRITE(V3D_INTCTL, V3D_DRIVER_IRQS);
/*
* Turn all our interrupts on. Binner out of memory is the
* only one we expect to trigger at this point, since we've
* just come from poweron and haven't supplied any overflow
* memory yet.
*/
V3D_WRITE(V3D_INTENA, V3D_DRIVER_IRQS);
spin_lock_irqsave(&vc4->job_lock, irqflags);
vc4_cancel_bin_job(dev);
vc4_irq_finish_render_job(dev);
spin_unlock_irqrestore(&vc4->job_lock, irqflags);
}