Skip to contentSkip to conversation

sh: remove stray markers

PATCH In review Christoph Hellwig 3 participants 6 messages May 6, 2009 – May 7, 2009 SuperH

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 sh: remove stray markers

arch/sh has a couple of stray markers without any users introduced in commit 3d58695edbfac785161bf282dc11fd42a483d6c9. Remove them in preparation of removing the markers in favour of the TRACE_EVENT macro (and also because we don't keep dead code around).

Paul, are you okay putting this in via the tracing tree? I'd like to have all marker removals queued up there so we can kill it in one go once 2.6.31 opens.

arch/sh/kernel/process_32.c +0 -2
Index: linux-2.6-tip/arch/sh/kernel/process_32.c
===================================================================
@@ -119,8 +119,6 @@
 	pid = do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0,
 		      &regs, 0, NULL, NULL);
 
-	trace_mark(kernel_arch_kthread_create, "pid %d fn %p", pid, fn);
-
 	return pid;
 }
 
arch/sh/kernel/process_64.c +1 -6
Index: linux-2.6-tip/arch/sh/kernel/process_64.c
===================================================================
@@ -323,7 +323,6 @@
 int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
 {
 	struct pt_regs regs;
-	int pid;
 
 	memset(&regs, 0, sizeof(regs));
 	regs.regs[2] = (unsigned long)arg;
@@ -333,12 +332,8 @@
 	regs.sr = (1 << 30);
 
 	/* Ok, create the new process.. */
-	pid = do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0,
+	return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0,
 		      &regs, 0, NULL, NULL);
-
-	trace_mark(kernel_arch_kthread_create, "pid %d fn %p", pid, fn);
-
-	return pid;
 }
 
 /*
arch/sh/kernel/sys_sh.c +0 -2
Index: linux-2.6-tip/arch/sh/kernel/sys_sh.c
===================================================================
@@ -79,8 +79,6 @@
 	version = call >> 16; /* hack for backward compatibility */
 	call &= 0xffff;
 
-	trace_mark(kernel_arch_ipc_call, "call %u first %d", call, first);
-
 	if (call <= SEMTIMEDOP)
 		switch (call) {
 		case SEMOP:
arch/sh/mm/fault_32.c +0 -4
Index: linux-2.6-tip/arch/sh/mm/fault_32.c
===================================================================
@@ -249,9 +249,6 @@
 {
 	int ret = 0;
 
-	trace_mark(kernel_arch_trap_entry, "trap_id %d ip #p%ld",
-		   trap >> 5, instruction_pointer(regs));
-
 #ifdef CONFIG_KPROBES
 	if (!user_mode(regs)) {
 		preempt_disable();
@@ -327,6 +324,5 @@
 
 	ret = 0;
 out:
-	trace_mark(kernel_arch_trap_exit, MARK_NOARGS);
 	return ret;
 }
#1 lore raw
Quoted 9 lines · Christoph Hellwig

arch/sh has a couple of stray markers without any users introduced in commit 3d58695edbfac785161bf282dc11fd42a483d6c9. Remove them in preparation of removing the markers in favour of the TRACE_EVENT macro (and also because we don't keep dead code around).

Paul, are you okay putting this in via the tracing tree? I'd like to have all marker removals queued up there so we can kill it in one go once 2.6.31 opens.

It's fine with me to have this included in the SH tree as well. The process.c files tend to be frequently modified in architectures so having overlap between tip/tracing and the SH tree could become awkward to Paul.

Also, if there's some current use of these markers then it's OK to not do this change right now but replace them with tracepoints in the v2.6.31 merge window, after both the SH and the tracing infrastructure changes are upstream.

Ingo

#2 lore raw
Quoted 4 lines · Ingo Molnar

Also, if there's some current use of these markers then it's OK to not do this change right now but replace them with tracepoints in the v2.6.31 merge window, after both the SH and the tracing infrastructure changes are upstream.

There's not user of these. The two only uses sets of markers are sputrace, and kvmtrace. I've sent a patch for the first, and the latter is work in progress.

#3 lore raw
Quoted 7 lines · Christoph Hellwig
Quoted 4 lines · Ingo Molnar

Also, if there's some current use of these markers then it's OK to not do this change right now but replace them with tracepoints in the v2.6.31 merge window, after both the SH and the tracing infrastructure changes are upstream.

There's not user of these. [...]

I'll defer to Paul to decide whether SH needs those markers. If he added them to ease his own debugging (via whatever external approach), then that's an OK goal and we can handle this all politely without destroying functionality, in the next merge window.

Ingo

#4 lore raw
Quoted 9 lines · Ingo Molnar
Quoted 6 lines · Christoph Hellwig
Quoted 4 lines · Ingo Molnar

Also, if there's some current use of these markers then it's OK to not do this change right now but replace them with tracepoints in the v2.6.31 merge window, after both the SH and the tracing infrastructure changes are upstream.

Quoted 7 lines
Quoted 1 line

There's not user of these. [...]

I'll defer to Paul to decide whether SH needs those markers. If he added them to ease his own debugging (via whatever external approach), then that's an OK goal and we can handle this all politely without destroying functionality, in the next merge window.

They were primarily added to aid some downstream folks that were just adding these on their own and generally not paying attention to where the markers were being placed on subsequent merges. I don't feel too strongly one way or the other, but as they have proven to be useful for some people, I would rather transition over to tracepoints during the merge window than kill them off prematurely.

#5 lore raw
Quoted 17 lines · Paul Mundt
Quoted 9 lines · Ingo Molnar
Quoted 6 lines · Christoph Hellwig
Quoted 4 lines · Ingo Molnar

Also, if there's some current use of these markers then it's OK to not do this change right now but replace them with tracepoints in the v2.6.31 merge window, after both the SH and the tracing infrastructure changes are upstream.

Quoted 6 lines
Quoted 1 line

There's not user of these. [...]

I'll defer to Paul to decide whether SH needs those markers. If he added them to ease his own debugging (via whatever external approach), then that's an OK goal and we can handle this all politely without destroying functionality, in the next merge window.

Quoted 7 lines

They were primarily added to aid some downstream folks that were just adding these on their own and generally not paying attention to where the markers were being placed on subsequent merges. I don't feel too strongly one way or the other, but as they have proven to be useful for some people, I would rather transition over to tracepoints during the merge window than kill them off prematurely.

fair enough!

Ingo