diff mbox series

tracing: always define trace_{irq,preempt}_{enable_disable}

Message ID 20171019083230.2450779-1-arnd@arndb.de
State Superseded
Headers show
Series tracing: always define trace_{irq,preempt}_{enable_disable} | expand

Commit Message

Arnd Bergmann Oct. 19, 2017, 8:32 a.m. UTC
We get a build error in the irqsoff tracer in some configurations:

kernel/trace/trace_irqsoff.c: In function 'trace_preempt_on':
kernel/trace/trace_irqsoff.c:855:2: error: implicit declaration of function 'trace_preempt_enable_rcuidle'; did you mean 'trace_irq_enable_rcuidle'? [-Werror=implicit-function-declaration]
  trace_preempt_enable_rcuidle(a0, a1);

The problem is that trace_preempt_enable_rcuidle() has different
definition based on multiple Kconfig symbols, but not all combinations
have a valid definition.

This changes the conditions so that we always get exactly one
definition of each of the four tracing macros. I have not tried
to verify that these definitions are sensible, but now we
can build all randconfig combinations again.

Fixes: d59158162e03 ("tracing: Add support for preempt and irq enable/disable events")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

---
 include/trace/events/preemptirq.h | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

-- 
2.9.0

Comments

Joel Fernandes Oct. 19, 2017, 3:42 p.m. UTC | #1
Hi Arnd,

On Thu, Oct 19, 2017 at 1:32 AM, Arnd Bergmann <arnd@arndb.de> wrote:
> We get a build error in the irqsoff tracer in some configurations:

>

> kernel/trace/trace_irqsoff.c: In function 'trace_preempt_on':

> kernel/trace/trace_irqsoff.c:855:2: error: implicit declaration of function 'trace_preempt_enable_rcuidle'; did you mean 'trace_irq_enable_rcuidle'? [-Werror=implicit-function-declaration]

>   trace_preempt_enable_rcuidle(a0, a1);

>

> The problem is that trace_preempt_enable_rcuidle() has different

> definition based on multiple Kconfig symbols, but not all combinations

> have a valid definition.

>

> This changes the conditions so that we always get exactly one

> definition of each of the four tracing macros. I have not tried

> to verify that these definitions are sensible, but now we

> can build all randconfig combinations again.

>


Thanks for catching this. I didn't follow why it breaks for you,
especially I'm troubled by your proposal of defining the empty macro
for !defined(CONFIG_DEBUG_PREEMPT) in your patch. Could you provide
your config sample and architecture you're building for? I'm guessing
its ARM but let me know. I will try to build it and reproduce it.

thanks,
Joel
Arnd Bergmann Oct. 19, 2017, 4:04 p.m. UTC | #2
On Thu, Oct 19, 2017 at 5:42 PM, Joel Fernandes <joelaf@google.com> wrote:
> Hi Arnd,

>

> On Thu, Oct 19, 2017 at 1:32 AM, Arnd Bergmann <arnd@arndb.de> wrote:

>> We get a build error in the irqsoff tracer in some configurations:

>>

>> kernel/trace/trace_irqsoff.c: In function 'trace_preempt_on':

>> kernel/trace/trace_irqsoff.c:855:2: error: implicit declaration of function 'trace_preempt_enable_rcuidle'; did you mean 'trace_irq_enable_rcuidle'? [-Werror=implicit-function-declaration]

>>   trace_preempt_enable_rcuidle(a0, a1);

>>

>> The problem is that trace_preempt_enable_rcuidle() has different

>> definition based on multiple Kconfig symbols, but not all combinations

>> have a valid definition.

>>

>> This changes the conditions so that we always get exactly one

>> definition of each of the four tracing macros. I have not tried

>> to verify that these definitions are sensible, but now we

>> can build all randconfig combinations again.

>>

>

> Thanks for catching this. I didn't follow why it breaks for you,

> especially I'm troubled by your proposal of defining the empty macro

> for !defined(CONFIG_DEBUG_PREEMPT) in your patch. Could you provide

> your config sample and architecture you're building for? I'm guessing

> its ARM but let me know. I will try to build it and reproduce it.


I ran into it during an arm64 randconfig build, uploaded the config file to
https://pastebin.com/s5AiqXTM

        Arnd

> thanks,

> Joel
Joel Fernandes Oct. 19, 2017, 6:42 p.m. UTC | #3
Hi,

On Thu, Oct 19, 2017 at 1:32 AM, Arnd Bergmann <arnd@arndb.de> wrote:
> We get a build error in the irqsoff tracer in some configurations:

>

> kernel/trace/trace_irqsoff.c: In function 'trace_preempt_on':

> kernel/trace/trace_irqsoff.c:855:2: error: implicit declaration of function 'trace_preempt_enable_rcuidle'; did you mean 'trace_irq_enable_rcuidle'? [-Werror=implicit-function-declaration]

>   trace_preempt_enable_rcuidle(a0, a1);

>

> The problem is that trace_preempt_enable_rcuidle() has different

> definition based on multiple Kconfig symbols, but not all combinations

> have a valid definition.

>

> This changes the conditions so that we always get exactly one

> definition of each of the four tracing macros. I have not tried

> to verify that these definitions are sensible, but now we

> can build all randconfig combinations again.

>

> Fixes: d59158162e03 ("tracing: Add support for preempt and irq enable/disable events")

> Signed-off-by: Arnd Bergmann <arnd@arndb.de>



Yes this build condition is possible. Thanks for finding it. I agree
with this patch, I will be simplifying this much more once I integrate
PROVE_LOCKING to use tracepoint probes but this is Ok with me for now.

Acked-by: Joel Fernandes <joelaf@google.com>


thanks!

- Joel

<snip>
Steven Rostedt Oct. 31, 2017, 2:13 p.m. UTC | #4
On Thu, 19 Oct 2017 10:32:13 +0200
Arnd Bergmann <arnd@arndb.de> wrote:

> We get a build error in the irqsoff tracer in some configurations:

> 

> kernel/trace/trace_irqsoff.c: In function 'trace_preempt_on':

> kernel/trace/trace_irqsoff.c:855:2: error: implicit declaration of function 'trace_preempt_enable_rcuidle'; did you mean 'trace_irq_enable_rcuidle'? [-Werror=implicit-function-declaration]

>   trace_preempt_enable_rcuidle(a0, a1);

> 

> The problem is that trace_preempt_enable_rcuidle() has different

> definition based on multiple Kconfig symbols, but not all combinations

> have a valid definition.

> 

> This changes the conditions so that we always get exactly one

> definition of each of the four tracing macros. I have not tried

> to verify that these definitions are sensible, but now we

> can build all randconfig combinations again.

> 

> Fixes: d59158162e03 ("tracing: Add support for preempt and irq enable/disable events")

> Signed-off-by: Arnd Bergmann <arnd@arndb.de>


Applied. Thanks Arnd.

-- Steve
diff mbox series

Patch

diff --git a/include/trace/events/preemptirq.h b/include/trace/events/preemptirq.h
index f5024c560d8f..9c4eb33c5a1d 100644
--- a/include/trace/events/preemptirq.h
+++ b/include/trace/events/preemptirq.h
@@ -56,15 +56,18 @@  DEFINE_EVENT(preemptirq_template, preempt_enable,
 
 #include <trace/define_trace.h>
 
-#else /* !CONFIG_PREEMPTIRQ_EVENTS */
+#endif /* !CONFIG_PREEMPTIRQ_EVENTS */
 
+#if !defined(CONFIG_PREEMPTIRQ_EVENTS) || defined(CONFIG_PROVE_LOCKING)
 #define trace_irq_enable(...)
 #define trace_irq_disable(...)
-#define trace_preempt_enable(...)
-#define trace_preempt_disable(...)
 #define trace_irq_enable_rcuidle(...)
 #define trace_irq_disable_rcuidle(...)
+#endif
+
+#if !defined(CONFIG_PREEMPTIRQ_EVENTS) || !defined(CONFIG_DEBUG_PREEMPT)
+#define trace_preempt_enable(...)
+#define trace_preempt_disable(...)
 #define trace_preempt_enable_rcuidle(...)
 #define trace_preempt_disable_rcuidle(...)
-
 #endif