From: Steven Rostedt <rostedt@goodmis.org>
To: linux-kernel@vger.kernel.org
Cc: Ingo Molnar <mingo@kernel.org>,
Andrew Morton <akpm@linux-foundation.org>,
Thomas Gleixner <tglx@linutronix.de>
Subject: [for-next][PATCH 3/9] ftrace/x86: Rename MCOUNT_SAVE_FRAME and add more detailed comments
Date: Mon, 01 Dec 2014 14:22:28 -0500 [thread overview]
Message-ID: <20141201192307.057123561@goodmis.org> (raw)
In-Reply-To: 20141201192225.548178029@goodmis.org
[-- Attachment #1: 0003-ftrace-x86-Rename-MCOUNT_SAVE_FRAME-and-add-more-det.patch --]
[-- Type: text/plain, Size: 3395 bytes --]
From: "Steven Rostedt (Red Hat)" <rostedt@goodmis.org>
The name MCOUNT_SAVE_FRAME is rather confusing as it really isn't a
function frame that is saved, but just the required mcount registers
that are needed to be saved before C code may be called. The word
"frame" confuses it as being a function frame which it is not.
Rename MCOUNT_SAVE_FRAME and MCOUNT_RESTORE_FRAME to save_mcount_regs
and restore_mcount_regs respectively. Noticed the lower case, which
keeps it from screaming at the reviewers.
Link: http://lkml.kernel.org/r/CA+55aFwF+qCGSKdGaEgW4p6N65GZ5_XTV=1NbtWDvxnd5yYLiw@mail.gmail.com
Link: http://lkml.kernel.org/r/alpine.DEB.2.11.1411262304010.3961@nanos
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
arch/x86/kernel/mcount_64.S | 32 ++++++++++++++++++++++++--------
1 file changed, 24 insertions(+), 8 deletions(-)
diff --git a/arch/x86/kernel/mcount_64.S b/arch/x86/kernel/mcount_64.S
index 94fe46725fe0..0a693d011980 100644
--- a/arch/x86/kernel/mcount_64.S
+++ b/arch/x86/kernel/mcount_64.S
@@ -21,8 +21,24 @@
# define function_hook mcount
#endif
+/*
+ * gcc -pg option adds a call to 'mcount' in most functions.
+ * When -mfentry is used, the call is to 'fentry' and not 'mcount'
+ * and is done before the function's stack frame is set up.
+ * They both require a set of regs to be saved before calling
+ * any C code and restored before returning back to the function.
+ *
+ * On boot up, all these calls are converted into nops. When tracing
+ * is enabled, the call can jump to either ftrace_caller or
+ * ftrace_regs_caller. Callbacks (tracing functions) that require
+ * ftrace_regs_caller (like kprobes) need to have pt_regs passed to
+ * it. For this reason, the size of the pt_regs structure will be
+ * allocated on the stack and the required mcount registers will
+ * be saved in the locations that pt_regs has them in.
+ */
+
/* skip is set if the stack was already partially adjusted */
-.macro MCOUNT_SAVE_FRAME skip=0
+.macro save_mcount_regs skip=0
/*
* We add enough stack to save all regs.
*/
@@ -39,7 +55,7 @@
movq %rdx, RIP(%rsp)
.endm
-.macro MCOUNT_RESTORE_FRAME skip=0
+.macro restore_mcount_regs skip=0
movq R9(%rsp), %r9
movq R8(%rsp), %r8
movq RDI(%rsp), %rdi
@@ -52,7 +68,7 @@
/* skip is set if stack has been adjusted */
.macro ftrace_caller_setup trace_label skip=0
- MCOUNT_SAVE_FRAME \skip
+ save_mcount_regs \skip
/* Save this location */
GLOBAL(\trace_label)
@@ -121,7 +137,7 @@ GLOBAL(ftrace_call)
restore_frame
- MCOUNT_RESTORE_FRAME
+ restore_mcount_regs
/*
* The copied trampoline must call ftrace_return as it
@@ -196,7 +212,7 @@ GLOBAL(ftrace_regs_call)
movq RBX(%rsp), %rbx
/* skip=8 to skip flags saved in SS */
- MCOUNT_RESTORE_FRAME 8
+ restore_mcount_regs 8
/* Restore flags */
popfq
@@ -240,7 +256,7 @@ trace:
call *ftrace_trace_function
- MCOUNT_RESTORE_FRAME
+ restore_mcount_regs
jmp fgraph_trace
END(function_hook)
@@ -249,7 +265,7 @@ END(function_hook)
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
ENTRY(ftrace_graph_caller)
- MCOUNT_SAVE_FRAME
+ save_mcount_regs
#ifdef CC_USING_FENTRY
leaq SS+16(%rsp), %rdi
@@ -263,7 +279,7 @@ ENTRY(ftrace_graph_caller)
call prepare_ftrace_return
- MCOUNT_RESTORE_FRAME
+ restore_mcount_regs
retq
END(ftrace_graph_caller)
--
2.1.3
next prev parent reply other threads:[~2014-12-01 19:25 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-12-01 19:22 [for-next][PATCH 0/9] ftrace/x86: Updated mcount cleanup for 3.19 Steven Rostedt
2014-12-01 19:22 ` [for-next][PATCH 1/9] ftrace/x86: Have static tracing also use ftrace_caller_setup Steven Rostedt
2014-12-01 19:22 ` [for-next][PATCH 2/9] ftrace/x86: Move MCOUNT_SAVE_FRAME out of header file Steven Rostedt
2014-12-01 19:22 ` Steven Rostedt [this message]
2014-12-01 19:22 ` [for-next][PATCH 4/9] ftrace/x86: Have save_mcount_regs store RIP in %rdi for first parameter Steven Rostedt
2014-12-01 19:22 ` [for-next][PATCH 5/9] ftrace/x86: Simplify save_mcount_regs on getting RIP Steven Rostedt
2014-12-01 19:22 ` [for-next][PATCH 6/9] ftrace/x86: Add macro MCOUNT_REG_SIZE for amount of stack used to save mcount regs Steven Rostedt
2014-12-01 19:22 ` [for-next][PATCH 7/9] ftrace/x86: Have save_mcount_regs macro also save stack frames if needed Steven Rostedt
2014-12-01 19:22 ` [for-next][PATCH 8/9] ftrace/x86: Get rid of ftrace_caller_setup Steven Rostedt
2014-12-01 19:22 ` [for-next][PATCH 9/9] ftrace/fgraph/x86: Have prepare_ftrace_return() take ip as first parameter Steven Rostedt
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20141201192307.057123561@goodmis.org \
--to=rostedt@goodmis.org \
--cc=akpm@linux-foundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=tglx@linutronix.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox