From: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
To: Anderw Morton <akpm@osdl.org>
Cc: LKML <linux-kernel@vger.kernel.org>,
Keith Owens <kaos@americas.sgi.com>,
Dean Nelson <dnc@americas.sgi.com>,
Tony Luck <tony.luck@intel.com>,
Anath Mavinakayanahalli <ananth@in.ibm.com>,
Prasanna Panchamukhi <prasanna@in.ibm.com>,
Dave M <davem@davemloft.net>, Andi Kleen <ak@suse.de>,
Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Subject: [(take 2)patch 4/7] Notify page fault call chain for powerpc
Date: Thu, 20 Apr 2006 16:25:00 -0700 [thread overview]
Message-ID: <20060420233912.045800746@csdlinux-2.jf.intel.com> (raw)
In-Reply-To: 20060420232456.712271992@csdlinux-2.jf.intel.com
[-- Attachment #1: notify_page_fault_powerpc.patch --]
[-- Type: text/plain, Size: 3101 bytes --]
Overloading of page fault notification with the
notify_die() has performance issues(since the
only interested components for page fault is
kprobes and/or kdb) and hence this patch introduces
the new notifier call chain exclusively for page
fault notifications their by avoiding notifying
unnecessary components in the do_page_fault() code
path.
Changes since previous version:
1) Moved the implementaion of register/unregister_page_fault_notifier()
call to arch/xxx/mm/fault.c
2) Added #ifdef CONFIG_KPROBES around this register/unregister calls
as currently kprobes is the only potential user.
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
---
arch/powerpc/mm/fault.c | 36 +++++++++++++++++++++++++++++++++++-
include/asm-powerpc/kdebug.h | 2 ++
2 files changed, 37 insertions(+), 1 deletion(-)
Index: linux-2.6.17-rc1-mm3/arch/powerpc/mm/fault.c
===================================================================
--- linux-2.6.17-rc1-mm3.orig/arch/powerpc/mm/fault.c
+++ linux-2.6.17-rc1-mm3/arch/powerpc/mm/fault.c
@@ -40,6 +40,40 @@
#include <asm/kdebug.h>
#include <asm/siginfo.h>
+#ifdef CONFIG_KPROBES
+ATOMIC_NOTIFIER_HEAD(notify_page_fault_chain);
+
+/* Hook to register for page fault notifications */
+int register_page_fault_notifier(struct notifier_block *nb)
+{
+ return atomic_notifier_chain_register(¬ify_page_fault_chain, nb);
+}
+
+int unregister_page_fault_notifier(struct notifier_block *nb)
+{
+ return atomic_notifier_chain_unregister(¬ify_page_fault_chain, nb);
+}
+
+static inline int notify_page_fault(enum die_val val, const char *str,
+ struct pt_regs *regs, long err, int trap, int sig)
+{
+ struct die_args args = {
+ .regs = regs,
+ .str = str,
+ .err = err,
+ .trapnr = trap,
+ .signr = sig
+ };
+ return atomic_notifier_call_chain(¬ify_page_fault_chain, val, &args);
+}
+#else
+static inline int notify_page_fault(enum die_val val, const char *str,
+ struct pt_regs *regs, long err, int trap, int sig)
+{
+ return NOTIFY_DONE;
+}
+#endif
+
/*
* Check whether the instruction at regs->nip is a store using
* an update addressing form which will update r1.
@@ -142,7 +176,7 @@ int __kprobes do_page_fault(struct pt_re
is_write = error_code & ESR_DST;
#endif /* CONFIG_4xx || CONFIG_BOOKE */
- if (notify_die(DIE_PAGE_FAULT, "page_fault", regs, error_code,
+ if (notify_page_fault(DIE_PAGE_FAULT, "page_fault", regs, error_code,
11, SIGSEGV) == NOTIFY_STOP)
return 0;
Index: linux-2.6.17-rc1-mm3/include/asm-powerpc/kdebug.h
===================================================================
--- linux-2.6.17-rc1-mm3.orig/include/asm-powerpc/kdebug.h
+++ linux-2.6.17-rc1-mm3/include/asm-powerpc/kdebug.h
@@ -18,6 +18,8 @@ struct die_args {
extern int register_die_notifier(struct notifier_block *);
extern int unregister_die_notifier(struct notifier_block *);
+extern int register_page_fault_notifier(struct notifier_block *);
+extern int unregister_page_fault_notifier(struct notifier_block *);
extern struct atomic_notifier_head powerpc_die_chain;
/* Grossly misnamed. */
--
next prev parent reply other threads:[~2006-04-20 23:51 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-04-20 23:24 [(take 2)patch 0/7] Notify page fault call chain Anil S Keshavamurthy
2006-04-20 23:24 ` [(take 2)patch 1/7] Notify page fault call chain for x86_64 Anil S Keshavamurthy
2006-04-20 23:24 ` [(take 2)patch 2/7] Notify page fault call chain for i386 Anil S Keshavamurthy
2006-04-20 23:24 ` [(take 2)patch 3/7] Notify page fault call chain for ia64 Anil S Keshavamurthy
2006-04-20 23:25 ` Anil S Keshavamurthy [this message]
2006-04-20 23:25 ` [(take 2)patch 5/7] Notify page fault call chain for sparc64 Anil S Keshavamurthy
2006-04-20 23:25 ` [(take 2)patch 6/7] Kprobes registers for notify page fault Anil S Keshavamurthy
2006-04-21 0:07 ` Keshavamurthy Anil S
2006-04-21 0:14 ` Keith Owens
2006-04-21 0:38 ` Keshavamurthy Anil S
2006-04-21 0:53 ` Keith Owens
2006-04-21 1:03 ` Keshavamurthy Anil S
2006-04-20 23:25 ` [(take 2)patch 7/7] Notify page fault call chain Anil S Keshavamurthy
2006-04-24 9:19 ` bibo mao
2006-04-25 1:10 ` Keith Owens
2006-04-25 5:43 ` Keshavamurthy Anil S
2006-04-24 19:28 ` [(take 2)patch 0/7] " Robin Holt
2006-04-25 6:01 ` Keshavamurthy Anil S
2006-04-25 10:15 ` Robin Holt
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=20060420233912.045800746@csdlinux-2.jf.intel.com \
--to=anil.s.keshavamurthy@intel.com \
--cc=ak@suse.de \
--cc=akpm@osdl.org \
--cc=ananth@in.ibm.com \
--cc=davem@davemloft.net \
--cc=dnc@americas.sgi.com \
--cc=kaos@americas.sgi.com \
--cc=linux-kernel@vger.kernel.org \
--cc=prasanna@in.ibm.com \
--cc=tony.luck@intel.com \
/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