From: Steven Rostedt (VMware) rostedt@goodmis.org
Commit e941759c74 ("fence: dma-buf cross-device synchronization") added trace event fence_annotate_wait_on, but never used it. It was renamed to dma_fence_annotate_wait_on by commit f54d186700 ("dma-buf: Rename struct fence to dma_fence") but still not used. As defined trace events have data structures and functions created for them, it is a waste of memory if they are not used. Remove the unused trace event.
Signed-off-by: Steven Rostedt (VMware) rostedt@goodmis.org --- Index: linux-trace.git/drivers/dma-buf/dma-fence.c =================================================================== --- linux-trace.git.orig/drivers/dma-buf/dma-fence.c +++ linux-trace.git/drivers/dma-buf/dma-fence.c @@ -27,7 +27,6 @@ #define CREATE_TRACE_POINTS #include <trace/events/dma_fence.h>
-EXPORT_TRACEPOINT_SYMBOL(dma_fence_annotate_wait_on); EXPORT_TRACEPOINT_SYMBOL(dma_fence_emit); EXPORT_TRACEPOINT_SYMBOL(dma_fence_enable_signal);
Index: linux-trace.git/include/trace/events/dma_fence.h =================================================================== --- linux-trace.git.orig/include/trace/events/dma_fence.h +++ linux-trace.git/include/trace/events/dma_fence.h @@ -8,46 +8,6 @@
struct dma_fence;
-TRACE_EVENT(dma_fence_annotate_wait_on, - - /* fence: the fence waiting on f1, f1: the fence to be waited on. */ - TP_PROTO(struct dma_fence *fence, struct dma_fence *f1), - - TP_ARGS(fence, f1), - - TP_STRUCT__entry( - __string(driver, fence->ops->get_driver_name(fence)) - __string(timeline, fence->ops->get_timeline_name(fence)) - __field(unsigned int, context) - __field(unsigned int, seqno) - - __string(waiting_driver, f1->ops->get_driver_name(f1)) - __string(waiting_timeline, f1->ops->get_timeline_name(f1)) - __field(unsigned int, waiting_context) - __field(unsigned int, waiting_seqno) - ), - - TP_fast_assign( - __assign_str(driver, fence->ops->get_driver_name(fence)) - __assign_str(timeline, fence->ops->get_timeline_name(fence)) - __entry->context = fence->context; - __entry->seqno = fence->seqno; - - __assign_str(waiting_driver, f1->ops->get_driver_name(f1)) - __assign_str(waiting_timeline, f1->ops->get_timeline_name(f1)) - __entry->waiting_context = f1->context; - __entry->waiting_seqno = f1->seqno; - - ), - - TP_printk("driver=%s timeline=%s context=%u seqno=%u " \ - "waits on driver=%s timeline=%s context=%u seqno=%u", - __get_str(driver), __get_str(timeline), __entry->context, - __entry->seqno, - __get_str(waiting_driver), __get_str(waiting_timeline), - __entry->waiting_context, __entry->waiting_seqno) -); - DECLARE_EVENT_CLASS(dma_fence,
TP_PROTO(struct dma_fence *fence),
Am 13.10.2017 um 16:06 schrieb Steven Rostedt:
From: Steven Rostedt (VMware) rostedt@goodmis.org
Commit e941759c74 ("fence: dma-buf cross-device synchronization") added trace event fence_annotate_wait_on, but never used it. It was renamed to dma_fence_annotate_wait_on by commit f54d186700 ("dma-buf: Rename struct fence to dma_fence") but still not used. As defined trace events have data structures and functions created for them, it is a waste of memory if they are not used. Remove the unused trace event.
Signed-off-by: Steven Rostedt (VMware) rostedt@goodmis.org
Reviewed-by: Christian König christian.koenig@amd.com
Index: linux-trace.git/drivers/dma-buf/dma-fence.c
--- linux-trace.git.orig/drivers/dma-buf/dma-fence.c +++ linux-trace.git/drivers/dma-buf/dma-fence.c @@ -27,7 +27,6 @@ #define CREATE_TRACE_POINTS #include <trace/events/dma_fence.h>
-EXPORT_TRACEPOINT_SYMBOL(dma_fence_annotate_wait_on); EXPORT_TRACEPOINT_SYMBOL(dma_fence_emit); EXPORT_TRACEPOINT_SYMBOL(dma_fence_enable_signal);
Index: linux-trace.git/include/trace/events/dma_fence.h
--- linux-trace.git.orig/include/trace/events/dma_fence.h +++ linux-trace.git/include/trace/events/dma_fence.h @@ -8,46 +8,6 @@
struct dma_fence;
-TRACE_EVENT(dma_fence_annotate_wait_on,
- /* fence: the fence waiting on f1, f1: the fence to be waited on. */
- TP_PROTO(struct dma_fence *fence, struct dma_fence *f1),
- TP_ARGS(fence, f1),
- TP_STRUCT__entry(
__string(driver, fence->ops->get_driver_name(fence))__string(timeline, fence->ops->get_timeline_name(fence))__field(unsigned int, context)__field(unsigned int, seqno)__string(waiting_driver, f1->ops->get_driver_name(f1))__string(waiting_timeline, f1->ops->get_timeline_name(f1))__field(unsigned int, waiting_context)__field(unsigned int, waiting_seqno)- ),
- TP_fast_assign(
__assign_str(driver, fence->ops->get_driver_name(fence))__assign_str(timeline, fence->ops->get_timeline_name(fence))__entry->context = fence->context;__entry->seqno = fence->seqno;__assign_str(waiting_driver, f1->ops->get_driver_name(f1))__assign_str(waiting_timeline, f1->ops->get_timeline_name(f1))__entry->waiting_context = f1->context;__entry->waiting_seqno = f1->seqno;- ),
- TP_printk("driver=%s timeline=%s context=%u seqno=%u " \
"waits on driver=%s timeline=%s context=%u seqno=%u",__get_str(driver), __get_str(timeline), __entry->context,__entry->seqno,__get_str(waiting_driver), __get_str(waiting_timeline),__entry->waiting_context, __entry->waiting_seqno)-);
DECLARE_EVENT_CLASS(dma_fence,
TP_PROTO(struct dma_fence *fence),
dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel
On Fri, Oct 13, 2017 at 05:27:59PM +0200, Christian König wrote:
Am 13.10.2017 um 16:06 schrieb Steven Rostedt:
From: Steven Rostedt (VMware) rostedt@goodmis.org
Commit e941759c74 ("fence: dma-buf cross-device synchronization") added trace event fence_annotate_wait_on, but never used it. It was renamed to dma_fence_annotate_wait_on by commit f54d186700 ("dma-buf: Rename struct fence to dma_fence") but still not used. As defined trace events have data structures and functions created for them, it is a waste of memory if they are not used. Remove the unused trace event.
Signed-off-by: Steven Rostedt (VMware) rostedt@goodmis.org
Reviewed-by: Christian König christian.koenig@amd.com
Since you have commit rights to drm-misc I assume you'll push this too.
Thanks, Daniel
Index: linux-trace.git/drivers/dma-buf/dma-fence.c
--- linux-trace.git.orig/drivers/dma-buf/dma-fence.c +++ linux-trace.git/drivers/dma-buf/dma-fence.c @@ -27,7 +27,6 @@ #define CREATE_TRACE_POINTS #include <trace/events/dma_fence.h> -EXPORT_TRACEPOINT_SYMBOL(dma_fence_annotate_wait_on); EXPORT_TRACEPOINT_SYMBOL(dma_fence_emit); EXPORT_TRACEPOINT_SYMBOL(dma_fence_enable_signal); Index: linux-trace.git/include/trace/events/dma_fence.h =================================================================== --- linux-trace.git.orig/include/trace/events/dma_fence.h +++ linux-trace.git/include/trace/events/dma_fence.h @@ -8,46 +8,6 @@ struct dma_fence; -TRACE_EVENT(dma_fence_annotate_wait_on,
- /* fence: the fence waiting on f1, f1: the fence to be waited on. */
- TP_PROTO(struct dma_fence *fence, struct dma_fence *f1),
- TP_ARGS(fence, f1),
- TP_STRUCT__entry(
__string(driver, fence->ops->get_driver_name(fence))__string(timeline, fence->ops->get_timeline_name(fence))__field(unsigned int, context)__field(unsigned int, seqno)__string(waiting_driver, f1->ops->get_driver_name(f1))__string(waiting_timeline, f1->ops->get_timeline_name(f1))__field(unsigned int, waiting_context)__field(unsigned int, waiting_seqno)- ),
- TP_fast_assign(
__assign_str(driver, fence->ops->get_driver_name(fence))__assign_str(timeline, fence->ops->get_timeline_name(fence))__entry->context = fence->context;__entry->seqno = fence->seqno;__assign_str(waiting_driver, f1->ops->get_driver_name(f1))__assign_str(waiting_timeline, f1->ops->get_timeline_name(f1))__entry->waiting_context = f1->context;__entry->waiting_seqno = f1->seqno;- ),
- TP_printk("driver=%s timeline=%s context=%u seqno=%u " \
"waits on driver=%s timeline=%s context=%u seqno=%u",__get_str(driver), __get_str(timeline), __entry->context,__entry->seqno,__get_str(waiting_driver), __get_str(waiting_timeline),__entry->waiting_context, __entry->waiting_seqno)-);
- DECLARE_EVENT_CLASS(dma_fence, TP_PROTO(struct dma_fence *fence),
dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel
On Mon, 16 Oct 2017 11:15:45 +0200 Daniel Vetter daniel@ffwll.ch wrote:
On Fri, Oct 13, 2017 at 05:27:59PM +0200, Christian König wrote:
Am 13.10.2017 um 16:06 schrieb Steven Rostedt:
From: Steven Rostedt (VMware) rostedt@goodmis.org
Commit e941759c74 ("fence: dma-buf cross-device synchronization") added trace event fence_annotate_wait_on, but never used it. It was renamed to dma_fence_annotate_wait_on by commit f54d186700 ("dma-buf: Rename struct fence to dma_fence") but still not used. As defined trace events have data structures and functions created for them, it is a waste of memory if they are not used. Remove the unused trace event.
Signed-off-by: Steven Rostedt (VMware) rostedt@goodmis.org
Reviewed-by: Christian König christian.koenig@amd.com
Since you have commit rights to drm-misc I assume you'll push this too.
Will it cause conflict if I pull it through my tree? It's just removing unused traecpoints. I added it to my tree already (but haven't pushed it to linux-next yet), with the Reviewed by from Christian. Is that OK? It shouldn't cause any merge conflicts that Linus can't handle.
-- Steve
Hi Steven,
On 16 October 2017 at 20:07, Steven Rostedt rostedt@goodmis.org wrote:
On Mon, 16 Oct 2017 11:15:45 +0200 Daniel Vetter daniel@ffwll.ch wrote:
On Fri, Oct 13, 2017 at 05:27:59PM +0200, Christian König wrote:
Am 13.10.2017 um 16:06 schrieb Steven Rostedt:
From: Steven Rostedt (VMware) rostedt@goodmis.org
Commit e941759c74 ("fence: dma-buf cross-device synchronization") added trace event fence_annotate_wait_on, but never used it. It was renamed to dma_fence_annotate_wait_on by commit f54d186700 ("dma-buf: Rename struct fence to dma_fence") but still not used. As defined trace events have data structures and functions created for them, it is a waste of memory if they are not used. Remove the unused trace event.
Signed-off-by: Steven Rostedt (VMware) rostedt@goodmis.org
Reviewed-by: Christian König christian.koenig@amd.com
Since you have commit rights to drm-misc I assume you'll push this too.
Will it cause conflict if I pull it through my tree? It's just removing unused traecpoints. I added it to my tree already (but haven't pushed it to linux-next yet), with the Reviewed by from Christian. Is that OK? It shouldn't cause any merge conflicts that Linus can't handle.
I suspect it should be ok; please do feel free to add my Acked-by: Sumit Semwal sumit.semwal@linaro.org if you wish :)
-- Steve _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel
Best, Sumit.
On Mon, 16 Oct 2017 20:49:11 +0530 Sumit Semwal sumit.semwal@linaro.org wrote:
I suspect it should be ok; please do feel free to add my Acked-by: Sumit Semwal sumit.semwal@linaro.org if you wish :)
Done. Thanks!
-- Steve
dri-devel@lists.freedesktop.org
-
Christian König -
Daniel Vetter -
Steven Rostedt -
Sumit Semwal