linux_dsm_epyc7002/drivers/gpu/drm/i915/i915_sw_fence.c

625 lines
14 KiB
C
Raw Normal View History

/*
* SPDX-License-Identifier: MIT
*
* (C) Copyright 2016 Intel Corporation
*/
#include <linux/slab.h>
dma-buf: Rename struct fence to dma_fence I plan to usurp the short name of struct fence for a core kernel struct, and so I need to rename the specialised fence/timeline for DMA operations to make room. A consensus was reached in https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html that making clear this fence applies to DMA operations was a good thing. Since then the patch has grown a bit as usage increases, so hopefully it remains a good thing! (v2...: rebase, rerun spatch) v3: Compile on msm, spotted a manual fixup that I broke. v4: Try again for msm, sorry Daniel coccinelle script: @@ @@ - struct fence + struct dma_fence @@ @@ - struct fence_ops + struct dma_fence_ops @@ @@ - struct fence_cb + struct dma_fence_cb @@ @@ - struct fence_array + struct dma_fence_array @@ @@ - enum fence_flag_bits + enum dma_fence_flag_bits @@ @@ ( - fence_init + dma_fence_init | - fence_release + dma_fence_release | - fence_free + dma_fence_free | - fence_get + dma_fence_get | - fence_get_rcu + dma_fence_get_rcu | - fence_put + dma_fence_put | - fence_signal + dma_fence_signal | - fence_signal_locked + dma_fence_signal_locked | - fence_default_wait + dma_fence_default_wait | - fence_add_callback + dma_fence_add_callback | - fence_remove_callback + dma_fence_remove_callback | - fence_enable_sw_signaling + dma_fence_enable_sw_signaling | - fence_is_signaled_locked + dma_fence_is_signaled_locked | - fence_is_signaled + dma_fence_is_signaled | - fence_is_later + dma_fence_is_later | - fence_later + dma_fence_later | - fence_wait_timeout + dma_fence_wait_timeout | - fence_wait_any_timeout + dma_fence_wait_any_timeout | - fence_wait + dma_fence_wait | - fence_context_alloc + dma_fence_context_alloc | - fence_array_create + dma_fence_array_create | - to_fence_array + to_dma_fence_array | - fence_is_array + dma_fence_is_array | - trace_fence_emit + trace_dma_fence_emit | - FENCE_TRACE + DMA_FENCE_TRACE | - FENCE_WARN + DMA_FENCE_WARN | - FENCE_ERR + DMA_FENCE_ERR ) ( ... ) Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Acked-by: Sumit Semwal <sumit.semwal@linaro.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk
2016-10-25 19:00:45 +07:00
#include <linux/dma-fence.h>
#include <linux/irq_work.h>
#include <linux/dma-resv.h>
#include "i915_sw_fence.h"
#include "i915_selftest.h"
drm/i915: Use a ctor for TYPESAFE_BY_RCU i915_request As we start peeking into requests for longer and longer, e.g. incorporating use of spinlocks when only protected by an rcu_read_lock(), we need to be careful in how we reset the request when recycling and need to preserve any barriers that may still be in use as the request is reset for reuse. Quoting Linus Torvalds: > If there is refcounting going on then why use SLAB_TYPESAFE_BY_RCU? .. because the object can be accessed (by RCU) after the refcount has gone down to zero, and the thing has been released. That's the whole and only point of SLAB_TYPESAFE_BY_RCU. That flag basically says: "I may end up accessing this object *after* it has been free'd, because there may be RCU lookups in flight" This has nothing to do with constructors. It's ok if the object gets reused as an object of the same type and does *not* get re-initialized, because we're perfectly fine seeing old stale data. What it guarantees is that the slab isn't shared with any other kind of object, _and_ that the underlying pages are free'd after an RCU quiescent period (so the pages aren't shared with another kind of object either during an RCU walk). And it doesn't necessarily have to have a constructor, because the thing that a RCU walk will care about is (a) guaranteed to be an object that *has* been on some RCU list (so it's not a "new" object) (b) the RCU walk needs to have logic to verify that it's still the *same* object and hasn't been re-used as something else. In contrast, a SLAB_TYPESAFE_BY_RCU memory gets free'd and re-used immediately, but because it gets reused as the same kind of object, the RCU walker can "know" what parts have meaning for re-use, in a way it couidn't if the re-use was random. That said, it *is* subtle, and people should be careful. > So the re-use might initialize the fields lazily, not necessarily using a ctor. If you have a well-defined refcount, and use "atomic_inc_not_zero()" to guard the speculative RCU access section, and use "atomic_dec_and_test()" in the freeing section, then you should be safe wrt new allocations. If you have a completely new allocation that has "random stale content", you know that it cannot be on the RCU list, so there is no speculative access that can ever see that random content. So the only case you need to worry about is a re-use allocation, and you know that the refcount will start out as zero even if you don't have a constructor. So you can think of the refcount itself as always having a zero constructor, *BUT* you need to be careful with ordering. In particular, whoever does the allocation needs to then set the refcount to a non-zero value *after* it has initialized all the other fields. And in particular, it needs to make sure that it uses the proper memory ordering to do so. NOTE! One thing to be very worried about is that re-initializing whatever RCU lists means that now the RCU walker may be walking on the wrong list so the walker may do the right thing for this particular entry, but it may miss walking *other* entries. So then you can get spurious lookup failures, because the RCU walker never walked all the way to the end of the right list. That ends up being a much more subtle bug. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191122094924.629690-1-chris@chris-wilson.co.uk
2019-11-22 16:49:24 +07:00
#if IS_ENABLED(CONFIG_DRM_I915_DEBUG)
#define I915_SW_FENCE_BUG_ON(expr) BUG_ON(expr)
#else
#define I915_SW_FENCE_BUG_ON(expr) BUILD_BUG_ON_INVALID(expr)
#endif
#define I915_SW_FENCE_FLAG_ALLOC BIT(3) /* after WQ_FLAG_* for safety */
static DEFINE_SPINLOCK(i915_sw_fence_lock);
enum {
DEBUG_FENCE_IDLE = 0,
DEBUG_FENCE_NOTIFY,
};
static void *i915_sw_fence_debug_hint(void *addr)
{
return (void *)(((struct i915_sw_fence *)addr)->flags & I915_SW_FENCE_MASK);
}
#ifdef CONFIG_DRM_I915_SW_FENCE_DEBUG_OBJECTS
static struct debug_obj_descr i915_sw_fence_debug_descr = {
.name = "i915_sw_fence",
.debug_hint = i915_sw_fence_debug_hint,
};
static inline void debug_fence_init(struct i915_sw_fence *fence)
{
debug_object_init(fence, &i915_sw_fence_debug_descr);
}
static inline void debug_fence_init_onstack(struct i915_sw_fence *fence)
{
debug_object_init_on_stack(fence, &i915_sw_fence_debug_descr);
}
static inline void debug_fence_activate(struct i915_sw_fence *fence)
{
debug_object_activate(fence, &i915_sw_fence_debug_descr);
}
static inline void debug_fence_set_state(struct i915_sw_fence *fence,
int old, int new)
{
debug_object_active_state(fence, &i915_sw_fence_debug_descr, old, new);
}
static inline void debug_fence_deactivate(struct i915_sw_fence *fence)
{
debug_object_deactivate(fence, &i915_sw_fence_debug_descr);
}
static inline void debug_fence_destroy(struct i915_sw_fence *fence)
{
debug_object_destroy(fence, &i915_sw_fence_debug_descr);
}
static inline void debug_fence_free(struct i915_sw_fence *fence)
{
debug_object_free(fence, &i915_sw_fence_debug_descr);
drm/i915: Flush the change in debugobject before reallocation When marking the debugobject as freed, be sure that write is flushed before another CPU may see it on a reallocation path. Only seen once in CI: [ 159.240873] WARNING: CPU: 3 PID: 6735 at lib/debugobjects.c:263 debug_print_object+0x87/0xb0 [ 159.240897] ODEBUG: init destroyed (active state 0) object type: i915_sw_fence hint: submit_notify+0x0/0x4c [i915] [ 159.240902] Modules linked in: snd_hda_intel i915 x86_pkg_temp_thermal intel_powerclamp coretemp crct10dif_pclmul snd_hda_codec_realtek crc32_pclmul snd_hda_codec_generic snd_hda_codec_hdmi ghash_clmulni_intel snd_hda_codec snd_hwdep snd_hda_core snd_pcm mei_me lpc_ich mei e1000e ptp pps_core [last unloaded: i915] [ 159.240913] CPU: 3 PID: 6735 Comm: gem_exec_nop Tainted: G U 4.10.0-rc3-CI-Trybot_479+ #1 [ 159.240913] Hardware name: LENOVO 10AGS00601/SHARKBAY, BIOS FBKT34AUS 04/24/2013 [ 159.240914] Call Trace: [ 159.240916] dump_stack+0x67/0x92 [ 159.240919] __warn+0xc6/0xe0 [ 159.240920] warn_slowpath_fmt+0x4a/0x50 [ 159.240921] debug_print_object+0x87/0xb0 [ 159.240935] ? __i915_request_wait_for_execute+0x1d0/0x1d0 [i915] [ 159.240936] __debug_object_init+0xb2/0x410 [ 159.240950] ? __i915_request_wait_for_execute+0x1d0/0x1d0 [i915] [ 159.240951] debug_object_init+0x16/0x20 [ 159.240962] __i915_sw_fence_init+0x29/0x60 [i915] [ 159.240975] i915_gem_request_alloc+0x1fb/0x450 [i915] [ 159.240987] i915_gem_do_execbuffer.isra.15+0x798/0x1b20 [i915] [ 159.241000] i915_gem_execbuffer2+0xc0/0x250 [i915] [ 159.241003] drm_ioctl+0x200/0x450 [ 159.241016] ? i915_gem_execbuffer+0x330/0x330 [i915] [ 159.241018] do_vfs_ioctl+0x90/0x6e0 [ 159.241020] ? trace_hardirqs_on_caller+0x122/0x1b0 [ 159.241021] SyS_ioctl+0x3c/0x70 [ 159.241023] entry_SYSCALL_64_fastpath+0x1c/0xb1 [ 159.241024] RIP: 0033:0x7f9bc4f41357 [ 159.241025] RSP: 002b:00007ffc6cd5c568 EFLAGS: 00000246 ORIG_RAX: 0000000000000010 [ 159.241026] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f9bc4f41357 [ 159.241026] RDX: 00007ffc6cd5c640 RSI: 0000000040406469 RDI: 0000000000000003 [ 159.241027] RBP: 00007ffc6cd5c640 R08: 0000000000047508 R09: 0000000000000001 [ 159.241027] R10: 000b58552d323c3d R11: 0000000000000246 R12: 0000000040406469 [ 159.241028] R13: 0000000000000003 R14: 0000000000000004 R15: 0000000000000001 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/20170113214335.5829-1-chris@chris-wilson.co.uk Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
2017-01-14 04:43:35 +07:00
smp_wmb(); /* flush the change in state before reallocation */
}
static inline void debug_fence_assert(struct i915_sw_fence *fence)
{
debug_object_assert_init(fence, &i915_sw_fence_debug_descr);
}
#else
static inline void debug_fence_init(struct i915_sw_fence *fence)
{
}
static inline void debug_fence_init_onstack(struct i915_sw_fence *fence)
{
}
static inline void debug_fence_activate(struct i915_sw_fence *fence)
{
}
static inline void debug_fence_set_state(struct i915_sw_fence *fence,
int old, int new)
{
}
static inline void debug_fence_deactivate(struct i915_sw_fence *fence)
{
}
static inline void debug_fence_destroy(struct i915_sw_fence *fence)
{
}
static inline void debug_fence_free(struct i915_sw_fence *fence)
{
}
static inline void debug_fence_assert(struct i915_sw_fence *fence)
{
}
#endif
static int __i915_sw_fence_notify(struct i915_sw_fence *fence,
enum i915_sw_fence_notify state)
{
i915_sw_fence_notify_t fn;
fn = (i915_sw_fence_notify_t)(fence->flags & I915_SW_FENCE_MASK);
return fn(fence, state);
}
#ifdef CONFIG_DRM_I915_SW_FENCE_DEBUG_OBJECTS
void i915_sw_fence_fini(struct i915_sw_fence *fence)
{
debug_fence_free(fence);
}
#endif
static void __i915_sw_fence_wake_up_all(struct i915_sw_fence *fence,
struct list_head *continuation)
{
wait_queue_head_t *x = &fence->wait;
wait_queue_entry_t *pos, *next;
unsigned long flags;
debug_fence_deactivate(fence);
atomic_set_release(&fence->pending, -1); /* 0 -> -1 [done] */
/*
* To prevent unbounded recursion as we traverse the graph of
sched/wait: Disambiguate wq_entry->task_list and wq_head->task_list naming So I've noticed a number of instances where it was not obvious from the code whether ->task_list was for a wait-queue head or a wait-queue entry. Furthermore, there's a number of wait-queue users where the lists are not for 'tasks' but other entities (poll tables, etc.), in which case the 'task_list' name is actively confusing. To clear this all up, name the wait-queue head and entry list structure fields unambiguously: struct wait_queue_head::task_list => ::head struct wait_queue_entry::task_list => ::entry For example, this code: rqw->wait.task_list.next != &wait->task_list ... is was pretty unclear (to me) what it's doing, while now it's written this way: rqw->wait.head.next != &wait->entry ... which makes it pretty clear that we are iterating a list until we see the head. Other examples are: list_for_each_entry_safe(pos, next, &x->task_list, task_list) { list_for_each_entry(wq, &fence->wait.task_list, task_list) { ... where it's unclear (to me) what we are iterating, and during review it's hard to tell whether it's trying to walk a wait-queue entry (which would be a bug), while now it's written as: list_for_each_entry_safe(pos, next, &x->head, entry) { list_for_each_entry(wq, &fence->wait.head, entry) { Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-06-20 17:06:46 +07:00
* i915_sw_fences, we move the entry list from this, the next ready
* fence, to the tail of the original fence's entry list
* (and so added to the list to be woken).
*/
spin_lock_irqsave_nested(&x->lock, flags, 1 + !!continuation);
if (continuation) {
sched/wait: Disambiguate wq_entry->task_list and wq_head->task_list naming So I've noticed a number of instances where it was not obvious from the code whether ->task_list was for a wait-queue head or a wait-queue entry. Furthermore, there's a number of wait-queue users where the lists are not for 'tasks' but other entities (poll tables, etc.), in which case the 'task_list' name is actively confusing. To clear this all up, name the wait-queue head and entry list structure fields unambiguously: struct wait_queue_head::task_list => ::head struct wait_queue_entry::task_list => ::entry For example, this code: rqw->wait.task_list.next != &wait->task_list ... is was pretty unclear (to me) what it's doing, while now it's written this way: rqw->wait.head.next != &wait->entry ... which makes it pretty clear that we are iterating a list until we see the head. Other examples are: list_for_each_entry_safe(pos, next, &x->task_list, task_list) { list_for_each_entry(wq, &fence->wait.task_list, task_list) { ... where it's unclear (to me) what we are iterating, and during review it's hard to tell whether it's trying to walk a wait-queue entry (which would be a bug), while now it's written as: list_for_each_entry_safe(pos, next, &x->head, entry) { list_for_each_entry(wq, &fence->wait.head, entry) { Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-06-20 17:06:46 +07:00
list_for_each_entry_safe(pos, next, &x->head, entry) {
if (pos->func == autoremove_wake_function)
pos->func(pos, TASK_NORMAL, 0, continuation);
else
sched/wait: Disambiguate wq_entry->task_list and wq_head->task_list naming So I've noticed a number of instances where it was not obvious from the code whether ->task_list was for a wait-queue head or a wait-queue entry. Furthermore, there's a number of wait-queue users where the lists are not for 'tasks' but other entities (poll tables, etc.), in which case the 'task_list' name is actively confusing. To clear this all up, name the wait-queue head and entry list structure fields unambiguously: struct wait_queue_head::task_list => ::head struct wait_queue_entry::task_list => ::entry For example, this code: rqw->wait.task_list.next != &wait->task_list ... is was pretty unclear (to me) what it's doing, while now it's written this way: rqw->wait.head.next != &wait->entry ... which makes it pretty clear that we are iterating a list until we see the head. Other examples are: list_for_each_entry_safe(pos, next, &x->task_list, task_list) { list_for_each_entry(wq, &fence->wait.task_list, task_list) { ... where it's unclear (to me) what we are iterating, and during review it's hard to tell whether it's trying to walk a wait-queue entry (which would be a bug), while now it's written as: list_for_each_entry_safe(pos, next, &x->head, entry) { list_for_each_entry(wq, &fence->wait.head, entry) { Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-06-20 17:06:46 +07:00
list_move_tail(&pos->entry, continuation);
}
} else {
LIST_HEAD(extra);
do {
list_for_each_entry_safe(pos, next, &x->head, entry) {
pos->func(pos,
TASK_NORMAL, fence->error,
&extra);
}
if (list_empty(&extra))
break;
sched/wait: Disambiguate wq_entry->task_list and wq_head->task_list naming So I've noticed a number of instances where it was not obvious from the code whether ->task_list was for a wait-queue head or a wait-queue entry. Furthermore, there's a number of wait-queue users where the lists are not for 'tasks' but other entities (poll tables, etc.), in which case the 'task_list' name is actively confusing. To clear this all up, name the wait-queue head and entry list structure fields unambiguously: struct wait_queue_head::task_list => ::head struct wait_queue_entry::task_list => ::entry For example, this code: rqw->wait.task_list.next != &wait->task_list ... is was pretty unclear (to me) what it's doing, while now it's written this way: rqw->wait.head.next != &wait->entry ... which makes it pretty clear that we are iterating a list until we see the head. Other examples are: list_for_each_entry_safe(pos, next, &x->task_list, task_list) { list_for_each_entry(wq, &fence->wait.task_list, task_list) { ... where it's unclear (to me) what we are iterating, and during review it's hard to tell whether it's trying to walk a wait-queue entry (which would be a bug), while now it's written as: list_for_each_entry_safe(pos, next, &x->head, entry) { list_for_each_entry(wq, &fence->wait.head, entry) { Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-06-20 17:06:46 +07:00
list_splice_tail_init(&extra, &x->head);
} while (1);
}
spin_unlock_irqrestore(&x->lock, flags);
debug_fence_assert(fence);
}
static void __i915_sw_fence_complete(struct i915_sw_fence *fence,
struct list_head *continuation)
{
debug_fence_assert(fence);
if (!atomic_dec_and_test(&fence->pending))
return;
debug_fence_set_state(fence, DEBUG_FENCE_IDLE, DEBUG_FENCE_NOTIFY);
if (__i915_sw_fence_notify(fence, FENCE_COMPLETE) != NOTIFY_DONE)
return;
debug_fence_set_state(fence, DEBUG_FENCE_NOTIFY, DEBUG_FENCE_IDLE);
__i915_sw_fence_wake_up_all(fence, continuation);
debug_fence_destroy(fence);
__i915_sw_fence_notify(fence, FENCE_FREE);
}
drm/i915: Use HW semaphores for inter-engine synchronisation on gen8+ Having introduced per-context seqno, we now have a means to identity progress across the system without feel of rollback as befell the global_seqno. That is we can program a MI_SEMAPHORE_WAIT operation in advance of submission safe in the knowledge that our target seqno and address is stable. However, since we are telling the GPU to busy-spin on the target address until it matches the signaling seqno, we only want to do so when we are sure that busy-spin will be completed quickly. To achieve this we only submit the request to HW once the signaler is itself executing (modulo preemption causing us to wait longer), and we only do so for default and above priority requests (so that idle priority tasks never themselves hog the GPU waiting for others). As might be reasonably expected, HW semaphores excel in inter-engine synchronisation microbenchmarks (where the 3x reduced latency / increased throughput more than offset the power cost of spinning on a second ring) and have significant improvement (can be up to ~10%, most see no change) for single clients that utilize multiple engines (typically media players and transcoders), without regressing multiple clients that can saturate the system or changing the power envelope dramatically. v3: Drop the older NEQ branch, now we pin the signaler's HWSP anyway. v4: Tell the world and include it as part of scheduler caps. Testcase: igt/gem_exec_whisper Testcase: igt/benchmarks/gem_wsim Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20190301170901.8340-3-chris@chris-wilson.co.uk
2019-03-02 00:09:00 +07:00
void i915_sw_fence_complete(struct i915_sw_fence *fence)
{
debug_fence_assert(fence);
if (WARN_ON(i915_sw_fence_done(fence)))
return;
__i915_sw_fence_complete(fence, NULL);
}
bool i915_sw_fence_await(struct i915_sw_fence *fence)
{
int pending;
/*
* It is only safe to add a new await to the fence while it has
* not yet been signaled (i.e. there are still existing signalers).
*/
pending = atomic_read(&fence->pending);
do {
if (pending < 1)
return false;
} while (!atomic_try_cmpxchg(&fence->pending, &pending, pending + 1));
return true;
}
void __i915_sw_fence_init(struct i915_sw_fence *fence,
i915_sw_fence_notify_t fn,
const char *name,
struct lock_class_key *key)
{
BUG_ON(!fn || (unsigned long)fn & ~I915_SW_FENCE_MASK);
drm/i915: Use a ctor for TYPESAFE_BY_RCU i915_request As we start peeking into requests for longer and longer, e.g. incorporating use of spinlocks when only protected by an rcu_read_lock(), we need to be careful in how we reset the request when recycling and need to preserve any barriers that may still be in use as the request is reset for reuse. Quoting Linus Torvalds: > If there is refcounting going on then why use SLAB_TYPESAFE_BY_RCU? .. because the object can be accessed (by RCU) after the refcount has gone down to zero, and the thing has been released. That's the whole and only point of SLAB_TYPESAFE_BY_RCU. That flag basically says: "I may end up accessing this object *after* it has been free'd, because there may be RCU lookups in flight" This has nothing to do with constructors. It's ok if the object gets reused as an object of the same type and does *not* get re-initialized, because we're perfectly fine seeing old stale data. What it guarantees is that the slab isn't shared with any other kind of object, _and_ that the underlying pages are free'd after an RCU quiescent period (so the pages aren't shared with another kind of object either during an RCU walk). And it doesn't necessarily have to have a constructor, because the thing that a RCU walk will care about is (a) guaranteed to be an object that *has* been on some RCU list (so it's not a "new" object) (b) the RCU walk needs to have logic to verify that it's still the *same* object and hasn't been re-used as something else. In contrast, a SLAB_TYPESAFE_BY_RCU memory gets free'd and re-used immediately, but because it gets reused as the same kind of object, the RCU walker can "know" what parts have meaning for re-use, in a way it couidn't if the re-use was random. That said, it *is* subtle, and people should be careful. > So the re-use might initialize the fields lazily, not necessarily using a ctor. If you have a well-defined refcount, and use "atomic_inc_not_zero()" to guard the speculative RCU access section, and use "atomic_dec_and_test()" in the freeing section, then you should be safe wrt new allocations. If you have a completely new allocation that has "random stale content", you know that it cannot be on the RCU list, so there is no speculative access that can ever see that random content. So the only case you need to worry about is a re-use allocation, and you know that the refcount will start out as zero even if you don't have a constructor. So you can think of the refcount itself as always having a zero constructor, *BUT* you need to be careful with ordering. In particular, whoever does the allocation needs to then set the refcount to a non-zero value *after* it has initialized all the other fields. And in particular, it needs to make sure that it uses the proper memory ordering to do so. NOTE! One thing to be very worried about is that re-initializing whatever RCU lists means that now the RCU walker may be walking on the wrong list so the walker may do the right thing for this particular entry, but it may miss walking *other* entries. So then you can get spurious lookup failures, because the RCU walker never walked all the way to the end of the right list. That ends up being a much more subtle bug. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191122094924.629690-1-chris@chris-wilson.co.uk
2019-11-22 16:49:24 +07:00
__init_waitqueue_head(&fence->wait, name, key);
fence->flags = (unsigned long)fn;
i915_sw_fence_reinit(fence);
}
void i915_sw_fence_reinit(struct i915_sw_fence *fence)
{
debug_fence_init(fence);
atomic_set(&fence->pending, 1);
fence->error = 0;
drm/i915: Use a ctor for TYPESAFE_BY_RCU i915_request As we start peeking into requests for longer and longer, e.g. incorporating use of spinlocks when only protected by an rcu_read_lock(), we need to be careful in how we reset the request when recycling and need to preserve any barriers that may still be in use as the request is reset for reuse. Quoting Linus Torvalds: > If there is refcounting going on then why use SLAB_TYPESAFE_BY_RCU? .. because the object can be accessed (by RCU) after the refcount has gone down to zero, and the thing has been released. That's the whole and only point of SLAB_TYPESAFE_BY_RCU. That flag basically says: "I may end up accessing this object *after* it has been free'd, because there may be RCU lookups in flight" This has nothing to do with constructors. It's ok if the object gets reused as an object of the same type and does *not* get re-initialized, because we're perfectly fine seeing old stale data. What it guarantees is that the slab isn't shared with any other kind of object, _and_ that the underlying pages are free'd after an RCU quiescent period (so the pages aren't shared with another kind of object either during an RCU walk). And it doesn't necessarily have to have a constructor, because the thing that a RCU walk will care about is (a) guaranteed to be an object that *has* been on some RCU list (so it's not a "new" object) (b) the RCU walk needs to have logic to verify that it's still the *same* object and hasn't been re-used as something else. In contrast, a SLAB_TYPESAFE_BY_RCU memory gets free'd and re-used immediately, but because it gets reused as the same kind of object, the RCU walker can "know" what parts have meaning for re-use, in a way it couidn't if the re-use was random. That said, it *is* subtle, and people should be careful. > So the re-use might initialize the fields lazily, not necessarily using a ctor. If you have a well-defined refcount, and use "atomic_inc_not_zero()" to guard the speculative RCU access section, and use "atomic_dec_and_test()" in the freeing section, then you should be safe wrt new allocations. If you have a completely new allocation that has "random stale content", you know that it cannot be on the RCU list, so there is no speculative access that can ever see that random content. So the only case you need to worry about is a re-use allocation, and you know that the refcount will start out as zero even if you don't have a constructor. So you can think of the refcount itself as always having a zero constructor, *BUT* you need to be careful with ordering. In particular, whoever does the allocation needs to then set the refcount to a non-zero value *after* it has initialized all the other fields. And in particular, it needs to make sure that it uses the proper memory ordering to do so. NOTE! One thing to be very worried about is that re-initializing whatever RCU lists means that now the RCU walker may be walking on the wrong list so the walker may do the right thing for this particular entry, but it may miss walking *other* entries. So then you can get spurious lookup failures, because the RCU walker never walked all the way to the end of the right list. That ends up being a much more subtle bug. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20191122094924.629690-1-chris@chris-wilson.co.uk
2019-11-22 16:49:24 +07:00
I915_SW_FENCE_BUG_ON(!fence->flags);
I915_SW_FENCE_BUG_ON(!list_empty(&fence->wait.head));
}
void i915_sw_fence_commit(struct i915_sw_fence *fence)
{
debug_fence_activate(fence);
i915_sw_fence_complete(fence);
}
static int i915_sw_fence_wake(wait_queue_entry_t *wq, unsigned mode, int flags, void *key)
{
i915_sw_fence_set_error_once(wq->private, flags);
sched/wait: Disambiguate wq_entry->task_list and wq_head->task_list naming So I've noticed a number of instances where it was not obvious from the code whether ->task_list was for a wait-queue head or a wait-queue entry. Furthermore, there's a number of wait-queue users where the lists are not for 'tasks' but other entities (poll tables, etc.), in which case the 'task_list' name is actively confusing. To clear this all up, name the wait-queue head and entry list structure fields unambiguously: struct wait_queue_head::task_list => ::head struct wait_queue_entry::task_list => ::entry For example, this code: rqw->wait.task_list.next != &wait->task_list ... is was pretty unclear (to me) what it's doing, while now it's written this way: rqw->wait.head.next != &wait->entry ... which makes it pretty clear that we are iterating a list until we see the head. Other examples are: list_for_each_entry_safe(pos, next, &x->task_list, task_list) { list_for_each_entry(wq, &fence->wait.task_list, task_list) { ... where it's unclear (to me) what we are iterating, and during review it's hard to tell whether it's trying to walk a wait-queue entry (which would be a bug), while now it's written as: list_for_each_entry_safe(pos, next, &x->head, entry) { list_for_each_entry(wq, &fence->wait.head, entry) { Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-06-20 17:06:46 +07:00
list_del(&wq->entry);
__i915_sw_fence_complete(wq->private, key);
if (wq->flags & I915_SW_FENCE_FLAG_ALLOC)
kfree(wq);
return 0;
}
static bool __i915_sw_fence_check_if_after(struct i915_sw_fence *fence,
const struct i915_sw_fence * const signaler)
{
wait_queue_entry_t *wq;
if (__test_and_set_bit(I915_SW_FENCE_CHECKED_BIT, &fence->flags))
return false;
if (fence == signaler)
return true;
sched/wait: Disambiguate wq_entry->task_list and wq_head->task_list naming So I've noticed a number of instances where it was not obvious from the code whether ->task_list was for a wait-queue head or a wait-queue entry. Furthermore, there's a number of wait-queue users where the lists are not for 'tasks' but other entities (poll tables, etc.), in which case the 'task_list' name is actively confusing. To clear this all up, name the wait-queue head and entry list structure fields unambiguously: struct wait_queue_head::task_list => ::head struct wait_queue_entry::task_list => ::entry For example, this code: rqw->wait.task_list.next != &wait->task_list ... is was pretty unclear (to me) what it's doing, while now it's written this way: rqw->wait.head.next != &wait->entry ... which makes it pretty clear that we are iterating a list until we see the head. Other examples are: list_for_each_entry_safe(pos, next, &x->task_list, task_list) { list_for_each_entry(wq, &fence->wait.task_list, task_list) { ... where it's unclear (to me) what we are iterating, and during review it's hard to tell whether it's trying to walk a wait-queue entry (which would be a bug), while now it's written as: list_for_each_entry_safe(pos, next, &x->head, entry) { list_for_each_entry(wq, &fence->wait.head, entry) { Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-06-20 17:06:46 +07:00
list_for_each_entry(wq, &fence->wait.head, entry) {
if (wq->func != i915_sw_fence_wake)
continue;
if (__i915_sw_fence_check_if_after(wq->private, signaler))
return true;
}
return false;
}
static void __i915_sw_fence_clear_checked_bit(struct i915_sw_fence *fence)
{
wait_queue_entry_t *wq;
if (!__test_and_clear_bit(I915_SW_FENCE_CHECKED_BIT, &fence->flags))
return;
sched/wait: Disambiguate wq_entry->task_list and wq_head->task_list naming So I've noticed a number of instances where it was not obvious from the code whether ->task_list was for a wait-queue head or a wait-queue entry. Furthermore, there's a number of wait-queue users where the lists are not for 'tasks' but other entities (poll tables, etc.), in which case the 'task_list' name is actively confusing. To clear this all up, name the wait-queue head and entry list structure fields unambiguously: struct wait_queue_head::task_list => ::head struct wait_queue_entry::task_list => ::entry For example, this code: rqw->wait.task_list.next != &wait->task_list ... is was pretty unclear (to me) what it's doing, while now it's written this way: rqw->wait.head.next != &wait->entry ... which makes it pretty clear that we are iterating a list until we see the head. Other examples are: list_for_each_entry_safe(pos, next, &x->task_list, task_list) { list_for_each_entry(wq, &fence->wait.task_list, task_list) { ... where it's unclear (to me) what we are iterating, and during review it's hard to tell whether it's trying to walk a wait-queue entry (which would be a bug), while now it's written as: list_for_each_entry_safe(pos, next, &x->head, entry) { list_for_each_entry(wq, &fence->wait.head, entry) { Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-06-20 17:06:46 +07:00
list_for_each_entry(wq, &fence->wait.head, entry) {
if (wq->func != i915_sw_fence_wake)
continue;
__i915_sw_fence_clear_checked_bit(wq->private);
}
}
static bool i915_sw_fence_check_if_after(struct i915_sw_fence *fence,
const struct i915_sw_fence * const signaler)
{
unsigned long flags;
bool err;
if (!IS_ENABLED(CONFIG_DRM_I915_SW_FENCE_CHECK_DAG))
return false;
spin_lock_irqsave(&i915_sw_fence_lock, flags);
err = __i915_sw_fence_check_if_after(fence, signaler);
__i915_sw_fence_clear_checked_bit(fence);
spin_unlock_irqrestore(&i915_sw_fence_lock, flags);
return err;
}
static int __i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
struct i915_sw_fence *signaler,
wait_queue_entry_t *wq, gfp_t gfp)
{
unsigned long flags;
int pending;
debug_fence_assert(fence);
might_sleep_if(gfpflags_allow_blocking(gfp));
if (i915_sw_fence_done(signaler)) {
i915_sw_fence_set_error_once(fence, signaler->error);
return 0;
}
debug_fence_assert(signaler);
/* The dependency graph must be acyclic. */
if (unlikely(i915_sw_fence_check_if_after(fence, signaler)))
return -EINVAL;
pending = 0;
if (!wq) {
wq = kmalloc(sizeof(*wq), gfp);
if (!wq) {
if (!gfpflags_allow_blocking(gfp))
return -ENOMEM;
i915_sw_fence_wait(signaler);
i915_sw_fence_set_error_once(fence, signaler->error);
return 0;
}
pending |= I915_SW_FENCE_FLAG_ALLOC;
}
sched/wait: Disambiguate wq_entry->task_list and wq_head->task_list naming So I've noticed a number of instances where it was not obvious from the code whether ->task_list was for a wait-queue head or a wait-queue entry. Furthermore, there's a number of wait-queue users where the lists are not for 'tasks' but other entities (poll tables, etc.), in which case the 'task_list' name is actively confusing. To clear this all up, name the wait-queue head and entry list structure fields unambiguously: struct wait_queue_head::task_list => ::head struct wait_queue_entry::task_list => ::entry For example, this code: rqw->wait.task_list.next != &wait->task_list ... is was pretty unclear (to me) what it's doing, while now it's written this way: rqw->wait.head.next != &wait->entry ... which makes it pretty clear that we are iterating a list until we see the head. Other examples are: list_for_each_entry_safe(pos, next, &x->task_list, task_list) { list_for_each_entry(wq, &fence->wait.task_list, task_list) { ... where it's unclear (to me) what we are iterating, and during review it's hard to tell whether it's trying to walk a wait-queue entry (which would be a bug), while now it's written as: list_for_each_entry_safe(pos, next, &x->head, entry) { list_for_each_entry(wq, &fence->wait.head, entry) { Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-06-20 17:06:46 +07:00
INIT_LIST_HEAD(&wq->entry);
wq->flags = pending;
wq->func = i915_sw_fence_wake;
wq->private = fence;
i915_sw_fence_await(fence);
spin_lock_irqsave(&signaler->wait.lock, flags);
if (likely(!i915_sw_fence_done(signaler))) {
__add_wait_queue_entry_tail(&signaler->wait, wq);
pending = 1;
} else {
i915_sw_fence_wake(wq, 0, signaler->error, NULL);
pending = 0;
}
spin_unlock_irqrestore(&signaler->wait.lock, flags);
return pending;
}
int i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
struct i915_sw_fence *signaler,
wait_queue_entry_t *wq)
{
return __i915_sw_fence_await_sw_fence(fence, signaler, wq, 0);
}
int i915_sw_fence_await_sw_fence_gfp(struct i915_sw_fence *fence,
struct i915_sw_fence *signaler,
gfp_t gfp)
{
return __i915_sw_fence_await_sw_fence(fence, signaler, NULL, gfp);
}
struct i915_sw_dma_fence_cb_timer {
struct i915_sw_dma_fence_cb base;
dma-buf: Rename struct fence to dma_fence I plan to usurp the short name of struct fence for a core kernel struct, and so I need to rename the specialised fence/timeline for DMA operations to make room. A consensus was reached in https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html that making clear this fence applies to DMA operations was a good thing. Since then the patch has grown a bit as usage increases, so hopefully it remains a good thing! (v2...: rebase, rerun spatch) v3: Compile on msm, spotted a manual fixup that I broke. v4: Try again for msm, sorry Daniel coccinelle script: @@ @@ - struct fence + struct dma_fence @@ @@ - struct fence_ops + struct dma_fence_ops @@ @@ - struct fence_cb + struct dma_fence_cb @@ @@ - struct fence_array + struct dma_fence_array @@ @@ - enum fence_flag_bits + enum dma_fence_flag_bits @@ @@ ( - fence_init + dma_fence_init | - fence_release + dma_fence_release | - fence_free + dma_fence_free | - fence_get + dma_fence_get | - fence_get_rcu + dma_fence_get_rcu | - fence_put + dma_fence_put | - fence_signal + dma_fence_signal | - fence_signal_locked + dma_fence_signal_locked | - fence_default_wait + dma_fence_default_wait | - fence_add_callback + dma_fence_add_callback | - fence_remove_callback + dma_fence_remove_callback | - fence_enable_sw_signaling + dma_fence_enable_sw_signaling | - fence_is_signaled_locked + dma_fence_is_signaled_locked | - fence_is_signaled + dma_fence_is_signaled | - fence_is_later + dma_fence_is_later | - fence_later + dma_fence_later | - fence_wait_timeout + dma_fence_wait_timeout | - fence_wait_any_timeout + dma_fence_wait_any_timeout | - fence_wait + dma_fence_wait | - fence_context_alloc + dma_fence_context_alloc | - fence_array_create + dma_fence_array_create | - to_fence_array + to_dma_fence_array | - fence_is_array + dma_fence_is_array | - trace_fence_emit + trace_dma_fence_emit | - FENCE_TRACE + DMA_FENCE_TRACE | - FENCE_WARN + DMA_FENCE_WARN | - FENCE_ERR + DMA_FENCE_ERR ) ( ... ) Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Acked-by: Sumit Semwal <sumit.semwal@linaro.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk
2016-10-25 19:00:45 +07:00
struct dma_fence *dma;
struct timer_list timer;
struct irq_work work;
struct rcu_head rcu;
};
static void dma_i915_sw_fence_wake(struct dma_fence *dma,
struct dma_fence_cb *data)
{
struct i915_sw_dma_fence_cb *cb = container_of(data, typeof(*cb), base);
i915_sw_fence_set_error_once(cb->fence, dma->error);
i915_sw_fence_complete(cb->fence);
kfree(cb);
}
static void timer_i915_sw_fence_wake(struct timer_list *t)
{
struct i915_sw_dma_fence_cb_timer *cb = from_timer(cb, t, timer);
struct i915_sw_fence *fence;
fence = xchg(&cb->base.fence, NULL);
if (!fence)
return;
pr_notice("Asynchronous wait on fence %s:%s:%llx timed out (hint:%ps)\n",
cb->dma->ops->get_driver_name(cb->dma),
cb->dma->ops->get_timeline_name(cb->dma),
cb->dma->seqno,
i915_sw_fence_debug_hint(fence));
i915_sw_fence_set_error_once(fence, -ETIMEDOUT);
i915_sw_fence_complete(fence);
}
static void dma_i915_sw_fence_wake_timer(struct dma_fence *dma,
struct dma_fence_cb *data)
{
struct i915_sw_dma_fence_cb_timer *cb =
container_of(data, typeof(*cb), base.base);
struct i915_sw_fence *fence;
fence = xchg(&cb->base.fence, NULL);
if (fence) {
i915_sw_fence_set_error_once(fence, dma->error);
i915_sw_fence_complete(fence);
}
irq_work_queue(&cb->work);
}
static void irq_i915_sw_fence_work(struct irq_work *wrk)
{
struct i915_sw_dma_fence_cb_timer *cb =
container_of(wrk, typeof(*cb), work);
del_timer_sync(&cb->timer);
dma-buf: Rename struct fence to dma_fence I plan to usurp the short name of struct fence for a core kernel struct, and so I need to rename the specialised fence/timeline for DMA operations to make room. A consensus was reached in https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html that making clear this fence applies to DMA operations was a good thing. Since then the patch has grown a bit as usage increases, so hopefully it remains a good thing! (v2...: rebase, rerun spatch) v3: Compile on msm, spotted a manual fixup that I broke. v4: Try again for msm, sorry Daniel coccinelle script: @@ @@ - struct fence + struct dma_fence @@ @@ - struct fence_ops + struct dma_fence_ops @@ @@ - struct fence_cb + struct dma_fence_cb @@ @@ - struct fence_array + struct dma_fence_array @@ @@ - enum fence_flag_bits + enum dma_fence_flag_bits @@ @@ ( - fence_init + dma_fence_init | - fence_release + dma_fence_release | - fence_free + dma_fence_free | - fence_get + dma_fence_get | - fence_get_rcu + dma_fence_get_rcu | - fence_put + dma_fence_put | - fence_signal + dma_fence_signal | - fence_signal_locked + dma_fence_signal_locked | - fence_default_wait + dma_fence_default_wait | - fence_add_callback + dma_fence_add_callback | - fence_remove_callback + dma_fence_remove_callback | - fence_enable_sw_signaling + dma_fence_enable_sw_signaling | - fence_is_signaled_locked + dma_fence_is_signaled_locked | - fence_is_signaled + dma_fence_is_signaled | - fence_is_later + dma_fence_is_later | - fence_later + dma_fence_later | - fence_wait_timeout + dma_fence_wait_timeout | - fence_wait_any_timeout + dma_fence_wait_any_timeout | - fence_wait + dma_fence_wait | - fence_context_alloc + dma_fence_context_alloc | - fence_array_create + dma_fence_array_create | - to_fence_array + to_dma_fence_array | - fence_is_array + dma_fence_is_array | - trace_fence_emit + trace_dma_fence_emit | - FENCE_TRACE + DMA_FENCE_TRACE | - FENCE_WARN + DMA_FENCE_WARN | - FENCE_ERR + DMA_FENCE_ERR ) ( ... ) Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Acked-by: Sumit Semwal <sumit.semwal@linaro.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk
2016-10-25 19:00:45 +07:00
dma_fence_put(cb->dma);
kfree_rcu(cb, rcu);
}
int i915_sw_fence_await_dma_fence(struct i915_sw_fence *fence,
dma-buf: Rename struct fence to dma_fence I plan to usurp the short name of struct fence for a core kernel struct, and so I need to rename the specialised fence/timeline for DMA operations to make room. A consensus was reached in https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html that making clear this fence applies to DMA operations was a good thing. Since then the patch has grown a bit as usage increases, so hopefully it remains a good thing! (v2...: rebase, rerun spatch) v3: Compile on msm, spotted a manual fixup that I broke. v4: Try again for msm, sorry Daniel coccinelle script: @@ @@ - struct fence + struct dma_fence @@ @@ - struct fence_ops + struct dma_fence_ops @@ @@ - struct fence_cb + struct dma_fence_cb @@ @@ - struct fence_array + struct dma_fence_array @@ @@ - enum fence_flag_bits + enum dma_fence_flag_bits @@ @@ ( - fence_init + dma_fence_init | - fence_release + dma_fence_release | - fence_free + dma_fence_free | - fence_get + dma_fence_get | - fence_get_rcu + dma_fence_get_rcu | - fence_put + dma_fence_put | - fence_signal + dma_fence_signal | - fence_signal_locked + dma_fence_signal_locked | - fence_default_wait + dma_fence_default_wait | - fence_add_callback + dma_fence_add_callback | - fence_remove_callback + dma_fence_remove_callback | - fence_enable_sw_signaling + dma_fence_enable_sw_signaling | - fence_is_signaled_locked + dma_fence_is_signaled_locked | - fence_is_signaled + dma_fence_is_signaled | - fence_is_later + dma_fence_is_later | - fence_later + dma_fence_later | - fence_wait_timeout + dma_fence_wait_timeout | - fence_wait_any_timeout + dma_fence_wait_any_timeout | - fence_wait + dma_fence_wait | - fence_context_alloc + dma_fence_context_alloc | - fence_array_create + dma_fence_array_create | - to_fence_array + to_dma_fence_array | - fence_is_array + dma_fence_is_array | - trace_fence_emit + trace_dma_fence_emit | - FENCE_TRACE + DMA_FENCE_TRACE | - FENCE_WARN + DMA_FENCE_WARN | - FENCE_ERR + DMA_FENCE_ERR ) ( ... ) Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Acked-by: Sumit Semwal <sumit.semwal@linaro.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk
2016-10-25 19:00:45 +07:00
struct dma_fence *dma,
unsigned long timeout,
gfp_t gfp)
{
dma-buf: Rename struct fence to dma_fence I plan to usurp the short name of struct fence for a core kernel struct, and so I need to rename the specialised fence/timeline for DMA operations to make room. A consensus was reached in https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html that making clear this fence applies to DMA operations was a good thing. Since then the patch has grown a bit as usage increases, so hopefully it remains a good thing! (v2...: rebase, rerun spatch) v3: Compile on msm, spotted a manual fixup that I broke. v4: Try again for msm, sorry Daniel coccinelle script: @@ @@ - struct fence + struct dma_fence @@ @@ - struct fence_ops + struct dma_fence_ops @@ @@ - struct fence_cb + struct dma_fence_cb @@ @@ - struct fence_array + struct dma_fence_array @@ @@ - enum fence_flag_bits + enum dma_fence_flag_bits @@ @@ ( - fence_init + dma_fence_init | - fence_release + dma_fence_release | - fence_free + dma_fence_free | - fence_get + dma_fence_get | - fence_get_rcu + dma_fence_get_rcu | - fence_put + dma_fence_put | - fence_signal + dma_fence_signal | - fence_signal_locked + dma_fence_signal_locked | - fence_default_wait + dma_fence_default_wait | - fence_add_callback + dma_fence_add_callback | - fence_remove_callback + dma_fence_remove_callback | - fence_enable_sw_signaling + dma_fence_enable_sw_signaling | - fence_is_signaled_locked + dma_fence_is_signaled_locked | - fence_is_signaled + dma_fence_is_signaled | - fence_is_later + dma_fence_is_later | - fence_later + dma_fence_later | - fence_wait_timeout + dma_fence_wait_timeout | - fence_wait_any_timeout + dma_fence_wait_any_timeout | - fence_wait + dma_fence_wait | - fence_context_alloc + dma_fence_context_alloc | - fence_array_create + dma_fence_array_create | - to_fence_array + to_dma_fence_array | - fence_is_array + dma_fence_is_array | - trace_fence_emit + trace_dma_fence_emit | - FENCE_TRACE + DMA_FENCE_TRACE | - FENCE_WARN + DMA_FENCE_WARN | - FENCE_ERR + DMA_FENCE_ERR ) ( ... ) Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Acked-by: Sumit Semwal <sumit.semwal@linaro.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk
2016-10-25 19:00:45 +07:00
struct i915_sw_dma_fence_cb *cb;
dma_fence_func_t func;
int ret;
debug_fence_assert(fence);
might_sleep_if(gfpflags_allow_blocking(gfp));
if (dma_fence_is_signaled(dma)) {
i915_sw_fence_set_error_once(fence, dma->error);
return 0;
}
cb = kmalloc(timeout ?
sizeof(struct i915_sw_dma_fence_cb_timer) :
sizeof(struct i915_sw_dma_fence_cb),
gfp);
if (!cb) {
if (!gfpflags_allow_blocking(gfp))
return -ENOMEM;
ret = dma_fence_wait(dma, false);
if (ret)
return ret;
i915_sw_fence_set_error_once(fence, dma->error);
return 0;
}
cb->fence = fence;
i915_sw_fence_await(fence);
func = dma_i915_sw_fence_wake;
if (timeout) {
struct i915_sw_dma_fence_cb_timer *timer =
container_of(cb, typeof(*timer), base);
timer->dma = dma_fence_get(dma);
init_irq_work(&timer->work, irq_i915_sw_fence_work);
timer_setup(&timer->timer,
timer_i915_sw_fence_wake, TIMER_IRQSAFE);
mod_timer(&timer->timer, round_jiffies_up(jiffies + timeout));
func = dma_i915_sw_fence_wake_timer;
}
ret = dma_fence_add_callback(dma, &cb->base, func);
if (ret == 0) {
ret = 1;
} else {
func(dma, &cb->base);
if (ret == -ENOENT) /* fence already signaled */
ret = 0;
}
return ret;
}
static void __dma_i915_sw_fence_wake(struct dma_fence *dma,
struct dma_fence_cb *data)
{
struct i915_sw_dma_fence_cb *cb = container_of(data, typeof(*cb), base);
i915_sw_fence_set_error_once(cb->fence, dma->error);
i915_sw_fence_complete(cb->fence);
}
int __i915_sw_fence_await_dma_fence(struct i915_sw_fence *fence,
struct dma_fence *dma,
struct i915_sw_dma_fence_cb *cb)
{
int ret;
debug_fence_assert(fence);
if (dma_fence_is_signaled(dma)) {
i915_sw_fence_set_error_once(fence, dma->error);
return 0;
}
cb->fence = fence;
i915_sw_fence_await(fence);
ret = dma_fence_add_callback(dma, &cb->base, __dma_i915_sw_fence_wake);
if (ret == 0) {
ret = 1;
} else {
__dma_i915_sw_fence_wake(dma, &cb->base);
if (ret == -ENOENT) /* fence already signaled */
ret = 0;
}
return ret;
}
int i915_sw_fence_await_reservation(struct i915_sw_fence *fence,
struct dma_resv *resv,
dma-buf: Rename struct fence to dma_fence I plan to usurp the short name of struct fence for a core kernel struct, and so I need to rename the specialised fence/timeline for DMA operations to make room. A consensus was reached in https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html that making clear this fence applies to DMA operations was a good thing. Since then the patch has grown a bit as usage increases, so hopefully it remains a good thing! (v2...: rebase, rerun spatch) v3: Compile on msm, spotted a manual fixup that I broke. v4: Try again for msm, sorry Daniel coccinelle script: @@ @@ - struct fence + struct dma_fence @@ @@ - struct fence_ops + struct dma_fence_ops @@ @@ - struct fence_cb + struct dma_fence_cb @@ @@ - struct fence_array + struct dma_fence_array @@ @@ - enum fence_flag_bits + enum dma_fence_flag_bits @@ @@ ( - fence_init + dma_fence_init | - fence_release + dma_fence_release | - fence_free + dma_fence_free | - fence_get + dma_fence_get | - fence_get_rcu + dma_fence_get_rcu | - fence_put + dma_fence_put | - fence_signal + dma_fence_signal | - fence_signal_locked + dma_fence_signal_locked | - fence_default_wait + dma_fence_default_wait | - fence_add_callback + dma_fence_add_callback | - fence_remove_callback + dma_fence_remove_callback | - fence_enable_sw_signaling + dma_fence_enable_sw_signaling | - fence_is_signaled_locked + dma_fence_is_signaled_locked | - fence_is_signaled + dma_fence_is_signaled | - fence_is_later + dma_fence_is_later | - fence_later + dma_fence_later | - fence_wait_timeout + dma_fence_wait_timeout | - fence_wait_any_timeout + dma_fence_wait_any_timeout | - fence_wait + dma_fence_wait | - fence_context_alloc + dma_fence_context_alloc | - fence_array_create + dma_fence_array_create | - to_fence_array + to_dma_fence_array | - fence_is_array + dma_fence_is_array | - trace_fence_emit + trace_dma_fence_emit | - FENCE_TRACE + DMA_FENCE_TRACE | - FENCE_WARN + DMA_FENCE_WARN | - FENCE_ERR + DMA_FENCE_ERR ) ( ... ) Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Acked-by: Sumit Semwal <sumit.semwal@linaro.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk
2016-10-25 19:00:45 +07:00
const struct dma_fence_ops *exclude,
bool write,
unsigned long timeout,
gfp_t gfp)
{
dma-buf: Rename struct fence to dma_fence I plan to usurp the short name of struct fence for a core kernel struct, and so I need to rename the specialised fence/timeline for DMA operations to make room. A consensus was reached in https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html that making clear this fence applies to DMA operations was a good thing. Since then the patch has grown a bit as usage increases, so hopefully it remains a good thing! (v2...: rebase, rerun spatch) v3: Compile on msm, spotted a manual fixup that I broke. v4: Try again for msm, sorry Daniel coccinelle script: @@ @@ - struct fence + struct dma_fence @@ @@ - struct fence_ops + struct dma_fence_ops @@ @@ - struct fence_cb + struct dma_fence_cb @@ @@ - struct fence_array + struct dma_fence_array @@ @@ - enum fence_flag_bits + enum dma_fence_flag_bits @@ @@ ( - fence_init + dma_fence_init | - fence_release + dma_fence_release | - fence_free + dma_fence_free | - fence_get + dma_fence_get | - fence_get_rcu + dma_fence_get_rcu | - fence_put + dma_fence_put | - fence_signal + dma_fence_signal | - fence_signal_locked + dma_fence_signal_locked | - fence_default_wait + dma_fence_default_wait | - fence_add_callback + dma_fence_add_callback | - fence_remove_callback + dma_fence_remove_callback | - fence_enable_sw_signaling + dma_fence_enable_sw_signaling | - fence_is_signaled_locked + dma_fence_is_signaled_locked | - fence_is_signaled + dma_fence_is_signaled | - fence_is_later + dma_fence_is_later | - fence_later + dma_fence_later | - fence_wait_timeout + dma_fence_wait_timeout | - fence_wait_any_timeout + dma_fence_wait_any_timeout | - fence_wait + dma_fence_wait | - fence_context_alloc + dma_fence_context_alloc | - fence_array_create + dma_fence_array_create | - to_fence_array + to_dma_fence_array | - fence_is_array + dma_fence_is_array | - trace_fence_emit + trace_dma_fence_emit | - FENCE_TRACE + DMA_FENCE_TRACE | - FENCE_WARN + DMA_FENCE_WARN | - FENCE_ERR + DMA_FENCE_ERR ) ( ... ) Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Acked-by: Sumit Semwal <sumit.semwal@linaro.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk
2016-10-25 19:00:45 +07:00
struct dma_fence *excl;
int ret = 0, pending;
debug_fence_assert(fence);
might_sleep_if(gfpflags_allow_blocking(gfp));
if (write) {
dma-buf: Rename struct fence to dma_fence I plan to usurp the short name of struct fence for a core kernel struct, and so I need to rename the specialised fence/timeline for DMA operations to make room. A consensus was reached in https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html that making clear this fence applies to DMA operations was a good thing. Since then the patch has grown a bit as usage increases, so hopefully it remains a good thing! (v2...: rebase, rerun spatch) v3: Compile on msm, spotted a manual fixup that I broke. v4: Try again for msm, sorry Daniel coccinelle script: @@ @@ - struct fence + struct dma_fence @@ @@ - struct fence_ops + struct dma_fence_ops @@ @@ - struct fence_cb + struct dma_fence_cb @@ @@ - struct fence_array + struct dma_fence_array @@ @@ - enum fence_flag_bits + enum dma_fence_flag_bits @@ @@ ( - fence_init + dma_fence_init | - fence_release + dma_fence_release | - fence_free + dma_fence_free | - fence_get + dma_fence_get | - fence_get_rcu + dma_fence_get_rcu | - fence_put + dma_fence_put | - fence_signal + dma_fence_signal | - fence_signal_locked + dma_fence_signal_locked | - fence_default_wait + dma_fence_default_wait | - fence_add_callback + dma_fence_add_callback | - fence_remove_callback + dma_fence_remove_callback | - fence_enable_sw_signaling + dma_fence_enable_sw_signaling | - fence_is_signaled_locked + dma_fence_is_signaled_locked | - fence_is_signaled + dma_fence_is_signaled | - fence_is_later + dma_fence_is_later | - fence_later + dma_fence_later | - fence_wait_timeout + dma_fence_wait_timeout | - fence_wait_any_timeout + dma_fence_wait_any_timeout | - fence_wait + dma_fence_wait | - fence_context_alloc + dma_fence_context_alloc | - fence_array_create + dma_fence_array_create | - to_fence_array + to_dma_fence_array | - fence_is_array + dma_fence_is_array | - trace_fence_emit + trace_dma_fence_emit | - FENCE_TRACE + DMA_FENCE_TRACE | - FENCE_WARN + DMA_FENCE_WARN | - FENCE_ERR + DMA_FENCE_ERR ) ( ... ) Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Acked-by: Sumit Semwal <sumit.semwal@linaro.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk
2016-10-25 19:00:45 +07:00
struct dma_fence **shared;
unsigned int count, i;
ret = dma_resv_get_fences_rcu(resv, &excl, &count, &shared);
if (ret)
return ret;
for (i = 0; i < count; i++) {
if (shared[i]->ops == exclude)
continue;
pending = i915_sw_fence_await_dma_fence(fence,
shared[i],
timeout,
gfp);
if (pending < 0) {
ret = pending;
break;
}
ret |= pending;
}
for (i = 0; i < count; i++)
dma-buf: Rename struct fence to dma_fence I plan to usurp the short name of struct fence for a core kernel struct, and so I need to rename the specialised fence/timeline for DMA operations to make room. A consensus was reached in https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html that making clear this fence applies to DMA operations was a good thing. Since then the patch has grown a bit as usage increases, so hopefully it remains a good thing! (v2...: rebase, rerun spatch) v3: Compile on msm, spotted a manual fixup that I broke. v4: Try again for msm, sorry Daniel coccinelle script: @@ @@ - struct fence + struct dma_fence @@ @@ - struct fence_ops + struct dma_fence_ops @@ @@ - struct fence_cb + struct dma_fence_cb @@ @@ - struct fence_array + struct dma_fence_array @@ @@ - enum fence_flag_bits + enum dma_fence_flag_bits @@ @@ ( - fence_init + dma_fence_init | - fence_release + dma_fence_release | - fence_free + dma_fence_free | - fence_get + dma_fence_get | - fence_get_rcu + dma_fence_get_rcu | - fence_put + dma_fence_put | - fence_signal + dma_fence_signal | - fence_signal_locked + dma_fence_signal_locked | - fence_default_wait + dma_fence_default_wait | - fence_add_callback + dma_fence_add_callback | - fence_remove_callback + dma_fence_remove_callback | - fence_enable_sw_signaling + dma_fence_enable_sw_signaling | - fence_is_signaled_locked + dma_fence_is_signaled_locked | - fence_is_signaled + dma_fence_is_signaled | - fence_is_later + dma_fence_is_later | - fence_later + dma_fence_later | - fence_wait_timeout + dma_fence_wait_timeout | - fence_wait_any_timeout + dma_fence_wait_any_timeout | - fence_wait + dma_fence_wait | - fence_context_alloc + dma_fence_context_alloc | - fence_array_create + dma_fence_array_create | - to_fence_array + to_dma_fence_array | - fence_is_array + dma_fence_is_array | - trace_fence_emit + trace_dma_fence_emit | - FENCE_TRACE + DMA_FENCE_TRACE | - FENCE_WARN + DMA_FENCE_WARN | - FENCE_ERR + DMA_FENCE_ERR ) ( ... ) Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Acked-by: Sumit Semwal <sumit.semwal@linaro.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk
2016-10-25 19:00:45 +07:00
dma_fence_put(shared[i]);
kfree(shared);
} else {
excl = dma_resv_get_excl_rcu(resv);
}
if (ret >= 0 && excl && excl->ops != exclude) {
pending = i915_sw_fence_await_dma_fence(fence,
excl,
timeout,
gfp);
if (pending < 0)
ret = pending;
else
ret |= pending;
}
dma-buf: Rename struct fence to dma_fence I plan to usurp the short name of struct fence for a core kernel struct, and so I need to rename the specialised fence/timeline for DMA operations to make room. A consensus was reached in https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html that making clear this fence applies to DMA operations was a good thing. Since then the patch has grown a bit as usage increases, so hopefully it remains a good thing! (v2...: rebase, rerun spatch) v3: Compile on msm, spotted a manual fixup that I broke. v4: Try again for msm, sorry Daniel coccinelle script: @@ @@ - struct fence + struct dma_fence @@ @@ - struct fence_ops + struct dma_fence_ops @@ @@ - struct fence_cb + struct dma_fence_cb @@ @@ - struct fence_array + struct dma_fence_array @@ @@ - enum fence_flag_bits + enum dma_fence_flag_bits @@ @@ ( - fence_init + dma_fence_init | - fence_release + dma_fence_release | - fence_free + dma_fence_free | - fence_get + dma_fence_get | - fence_get_rcu + dma_fence_get_rcu | - fence_put + dma_fence_put | - fence_signal + dma_fence_signal | - fence_signal_locked + dma_fence_signal_locked | - fence_default_wait + dma_fence_default_wait | - fence_add_callback + dma_fence_add_callback | - fence_remove_callback + dma_fence_remove_callback | - fence_enable_sw_signaling + dma_fence_enable_sw_signaling | - fence_is_signaled_locked + dma_fence_is_signaled_locked | - fence_is_signaled + dma_fence_is_signaled | - fence_is_later + dma_fence_is_later | - fence_later + dma_fence_later | - fence_wait_timeout + dma_fence_wait_timeout | - fence_wait_any_timeout + dma_fence_wait_any_timeout | - fence_wait + dma_fence_wait | - fence_context_alloc + dma_fence_context_alloc | - fence_array_create + dma_fence_array_create | - to_fence_array + to_dma_fence_array | - fence_is_array + dma_fence_is_array | - trace_fence_emit + trace_dma_fence_emit | - FENCE_TRACE + DMA_FENCE_TRACE | - FENCE_WARN + DMA_FENCE_WARN | - FENCE_ERR + DMA_FENCE_ERR ) ( ... ) Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Acked-by: Sumit Semwal <sumit.semwal@linaro.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk
2016-10-25 19:00:45 +07:00
dma_fence_put(excl);
return ret;
}
#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
#include "selftests/lib_sw_fence.c"
#include "selftests/i915_sw_fence.c"
#endif