Patches 4 and 5 seem agnostic to patches 1-3. I'll pull 4,5 into my
tree and 1-3 can go though tip. I'd like to run theses through my tests.
OK?
-- Steve
On Wed, 16 Mar 2016 15:34:33 +0100
Jiri Olsa wrote:
> Currently dynamic ftrace calls are updated any time
> the
Patches 4 and 5 seem agnostic to patches 1-3. I'll pull 4,5 into my
tree and 1-3 can go though tip. I'd like to run theses through my tests.
OK?
-- Steve
On Wed, 16 Mar 2016 15:34:33 +0100
Jiri Olsa wrote:
> Currently dynamic ftrace calls are updated any time
> the ftrace_ops is
On Wed, Mar 16, 2016 at 03:34:33PM +0100, Jiri Olsa wrote:
> Currently dynamic ftrace calls are updated any time
> the ftrace_ops is un/registered. If we do this update
> only when it's needed, we save lot of time for perf
> system wide ftrace function sampling/counting.
>
> The reason is that
On Wed, Mar 16, 2016 at 03:34:33PM +0100, Jiri Olsa wrote:
> Currently dynamic ftrace calls are updated any time
> the ftrace_ops is un/registered. If we do this update
> only when it's needed, we save lot of time for perf
> system wide ftrace function sampling/counting.
>
> The reason is that
Currently dynamic ftrace calls are updated any time
the ftrace_ops is un/registered. If we do this update
only when it's needed, we save lot of time for perf
system wide ftrace function sampling/counting.
The reason is that for system wide sampling/counting,
perf creates event for each cpu in
Currently dynamic ftrace calls are updated any time
the ftrace_ops is un/registered. If we do this update
only when it's needed, we save lot of time for perf
system wide ftrace function sampling/counting.
The reason is that for system wide sampling/counting,
perf creates event for each cpu in
Currently dynamic ftrace calls are updated any time
the ftrace_ops is un/registered. If we do this update
only when it's needed, we save lot of time for perf
system wide ftrace function sampling/counting.
The reason is that for system wide sampling/counting,
perf creates event for each cpu in
Currently dynamic ftrace calls are updated any time
the ftrace_ops is un/registered. If we do this update
only when it's needed, we save lot of time for perf
system wide ftrace function sampling/counting.
The reason is that for system wide sampling/counting,
perf creates event for each cpu in
8 matches
Mail list logo