Skip to contentSkip to conversation

kmsg_dump: Don't run on non-error paths by default

PATCH In review Matthew Garrett 3 participants 6 messages Feb 8, 2012 1 message not in the archive

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 kmsg_dump: Don't run on non-error paths by default

Since 04c6862c055fb687c90d9652f32c11a063df15cf kmsg_dump() gets run on normal paths including poweroff and reboot. This is less than ideal given pstore implementations that can only represent single backtraces, since a reboot may overwrite a stored oops before it's been picked up by userspace. In addition, some pstore backends may have low performance and provide a significant delay in reboot as a result.

This patch adds a printk.always_kmsg_dump kernel parameter (which can also be changed from userspace). Without it, the code will only be run on failure paths rather than on normal paths. The option can be enabled in environments where there's a desire to attempt to audit whether or not a reboot was cleanly requested or not.

---
 include/linux/kmsg_dump.h |    9 +++++++--
 kernel/printk.c           |    6 ++++++
 2 files changed, 13 insertions(+), 2 deletions(-)
include/linux/kmsg_dump.h +7 -2
index fee6631..35f7237 100644
@@ -15,13 +15,18 @@
 #include <linux/errno.h>
 #include <linux/list.h>
 
+/*
+ * Keep this list arranged in rough order of priority. Anything listed after
+ * KMSG_DUMP_OOPS will not be logged by default unless printk.always_kmsg_dump
+ * is passed to the kernel.
+ */
 enum kmsg_dump_reason {
-	KMSG_DUMP_OOPS,
 	KMSG_DUMP_PANIC,
+	KMSG_DUMP_OOPS,
+	KMSG_DUMP_EMERG,
 	KMSG_DUMP_RESTART,
 	KMSG_DUMP_HALT,
 	KMSG_DUMP_POWEROFF,
-	KMSG_DUMP_EMERG,
 };
 
 /**
kernel/printk.c +6 -0
index 13c0a11..e7d07f5 100644
@@ -702,6 +702,9 @@ static bool printk_time = 0;
 #endif
 module_param_named(time, printk_time, bool, S_IRUGO | S_IWUSR);
 
+static bool always_kmsg_dump;
+module_param_named(always_kmsg_dump, always_kmsg_dump, bool, S_IRUGO | S_IWUSR);
+
 /* Check if we have any console registered that can be called early in boot. */
 static int have_callable_console(void)
 {
@@ -1732,6 +1735,9 @@ void kmsg_dump(enum kmsg_dump_reason reason)
 	unsigned long l1, l2;
 	unsigned long flags;
 
+	if (reason > KMSG_DUMP_PANIC && !always_kmsg_dump)
+		return;
+
 	/* Theoretically, the log could move on after we do this, but
 	   there's not a lot we can do about that. The new messages
 	   will overwrite the start of what we dump. */
Signature · 1 line
1.7.7.6
#1 lore raw

On Wed, Feb 08, 2012 at 11:14:06AM -0500, Matthew Garrett wrote:

[..]

Quoted 32 lines
include/linux/kmsg_dump.h +7 -2
+/*
+ * Keep this list arranged in rough order of priority. Anything listed after
+ * KMSG_DUMP_OOPS will not be logged by default unless printk.always_kmsg_dump
+ * is passed to the kernel.
+ */
 enum kmsg_dump_reason {
-	KMSG_DUMP_OOPS,
 	KMSG_DUMP_PANIC,
+	KMSG_DUMP_OOPS,
+	KMSG_DUMP_EMERG,
 	KMSG_DUMP_RESTART,
 	KMSG_DUMP_HALT,
 	KMSG_DUMP_POWEROFF,
-	KMSG_DUMP_EMERG,
 };
 
 /**
kernel/printk.c +3 -0
index 13c0a11..e7d07f5 100644
@@ -702,6 +702,9 @@ static bool printk_time = 0;
 #endif
 module_param_named(time, printk_time, bool, S_IRUGO | S_IWUSR);
 
+static bool always_kmsg_dump;
+module_param_named(always_kmsg_dump, always_kmsg_dump, bool, S_IRUGO | S_IWUSR);
+
 /* Check if we have any console registered that can be called early in boot. */
 static int have_callable_console(void)
 {

kernel/printk.c void kmsg_dump(enum kmsg_dump_reason reason) · line 1739

Quoted 5 lines
kernel/printk.c +2 -0
@@ -1732,6 +1735,9 @@ void kmsg_dump(enum kmsg_dump_reason reason)
 	unsigned long l1, l2;
 	unsigned long flags;
 
+	if (reason > KMSG_DUMP_PANIC && !always_kmsg_dump)
+		return;

Did you mean reason > KMSG_DUMP_OOPS to enable oops logging by default?

Given the fact that not everybody likes kmsg_dump() and it is not known how stable it with various backends, will it make sense to keep it disabled by default and provide a knob to enable it (instead of always_kmsg_dump).

Thanks
Vivek

#2 lore raw
Quoted 4 lines · Vivek Goyal

kernel/printk.c void kmsg_dump(enum kmsg_dump_reason reason) · line 1739

Quoted 2 lines
kernel/printk.c +2 -0
+	if (reason > KMSG_DUMP_PANIC && !always_kmsg_dump)
+		return;

Did you mean reason > KMSG_DUMP_OOPS to enable oops logging by default?

Bah. Yes, I think my test system was set to panic on oops so I didn't catch that.

Quoted 3 lines

Given the fact that not everybody likes kmsg_dump() and it is not known how stable it with various backends, will it make sense to keep it disabled by default and provide a knob to enable it (instead of always_kmsg_dump).

I think logging panics and oopses by default is a useful feature. It may be desirable to have a mechanism to disable it if some other way of handling that (serial console, kdump) is available.

Signature · 1 line
Matthew Garrett | [email protected]
#3 lore raw
Quoted 10 lines · Matthew Garrett
Quoted 4 lines · Vivek Goyal

kernel/printk.c void kmsg_dump(enum kmsg_dump_reason reason) · line 1739

Quoted 2 lines
kernel/printk.c +2 -0
+	if (reason > KMSG_DUMP_PANIC && !always_kmsg_dump)
+		return;

Did you mean reason > KMSG_DUMP_OOPS to enable oops logging by default?

Bah. Yes, I think my test system was set to panic on oops so I didn't catch that.

Quoted 7 lines
Quoted 3 lines

Given the fact that not everybody likes kmsg_dump() and it is not known how stable it with various backends, will it make sense to keep it disabled by default and provide a knob to enable it (instead of always_kmsg_dump).

I think logging panics and oopses by default is a useful feature. It may be desirable to have a mechanism to disable it if some other way of handling that (serial console, kdump) is available.

Ok, keeping it enabled by default for panic and oops might be better.

Thinking more about it. This problem sounds similar to different kernel log levels where we enable desired level of logging. May be instead of having a boolean knob, we can have multivalue knob.

Thanks
Vivek

Acked-by on v1 1/1
#4 lore raw
In reply to Matthew Garrett

This patch, introducing tunable parameter, is reasonable to me.

Acked-by Seiji Aguchi <seiji.aguchi.hds.com>

-----Original Message-----
From: Matthew Garrett [mailto:[email protected]]
Sent: Wednesday, February 08, 2012 11:14 AM
To: [email protected]
Cc: Seiji Aguchi; [email protected]; [email protected]; Matthew Garrett
Subject: [PATCH] kmsg_dump: Don't run on non-error paths by default

Since 04c6862c055fb687c90d9652f32c11a063df15cf kmsg_dump() gets run on normal paths including poweroff and reboot. This is less than ideal given pstore implementations that can only represent single backtraces, since a reboot may overwrite a stored oops before it's been picked up by userspace. In addition, some pstore backends may have low performance and provide a significant delay in reboot as a result.

This patch adds a printk.always_kmsg_dump kernel parameter (which can also be changed from userspace). Without it, the code will only be run on failure paths rather than on normal paths. The option can be enabled in environments where there's a desire to attempt to audit whether or not a reboot was cleanly requested or not.

---
 include/linux/kmsg_dump.h |    9 +++++++--
 kernel/printk.c           |    6 ++++++
 2 files changed, 13 insertions(+), 2 deletions(-)
include/linux/kmsg_dump.h +9 -2
@@ -15,13 +15,18 @@
 #include <linux/errno.h>
 #include <linux/list.h>
 
+/*
+ * Keep this list arranged in rough order of priority. Anything listed 
+after
+ * KMSG_DUMP_OOPS will not be logged by default unless 
+printk.always_kmsg_dump
+ * is passed to the kernel.
+ */
 enum kmsg_dump_reason {
-	KMSG_DUMP_OOPS,
 	KMSG_DUMP_PANIC,
+	KMSG_DUMP_OOPS,
+	KMSG_DUMP_EMERG,
 	KMSG_DUMP_RESTART,
 	KMSG_DUMP_HALT,
 	KMSG_DUMP_POWEROFF,
-	KMSG_DUMP_EMERG,
 };

/**

kernel/printk.c +4 -0
@@ -702,6 +702,9 @@ static bool printk_time = 0;  #endif  module_param_named(time, printk_time, bool, S_IRUGO | S_IWUSR);
 
+static bool always_kmsg_dump;
+module_param_named(always_kmsg_dump, always_kmsg_dump, bool, S_IRUGO | 
+S_IWUSR);
+
 /* Check if we have any console registered that can be called early in boot. */  static int have_callable_console(void)  { @@ -1732,6 +1735,9 @@ void kmsg_dump(enum kmsg_dump_reason reason)
 	unsigned long l1, l2;
 	unsigned long flags;
 

+ if (reason > KMSG_DUMP_PANIC && !always_kmsg_dump)
+ return;
+
/* Theoretically, the log could move on after we do this, but
there's not a lot we can do about that. The new messages will overwrite the start of what we dump. */
--
1.7.7.6

A message that is not in the archive (sent off-list or not yet received): 32727E9A83EE9A42A1F0906295A3A77B2E598C0A2F@USINDEVS01.corp.hds.com

Acked-by Acked-by
#6 lore raw

There was a mistake in writing for my e-mail address.

-----Original Message-----
From: Seiji Aguchi
Sent: Wednesday, February 08, 2012 3:24 PM
To: 'Matthew Garrett'; [email protected]
Cc: [email protected]; [email protected]
Subject: RE: [PATCH] kmsg_dump: Don't run on non-error paths by default

This patch, introducing tunable parameter, is reasonable to me.

Acked-by Seiji Aguchi <seiji.aguchi.hds.com>

-----Original Message-----
From: Matthew Garrett [mailto:[email protected]]
Sent: Wednesday, February 08, 2012 11:14 AM
To: [email protected]
Cc: Seiji Aguchi; [email protected]; [email protected]; Matthew Garrett
Subject: [PATCH] kmsg_dump: Don't run on non-error paths by default

Since 04c6862c055fb687c90d9652f32c11a063df15cf kmsg_dump() gets run on normal paths including poweroff and reboot. This is less than ideal given pstore implementations that can only represent single backtraces, since a reboot may overwrite a stored oops before it's been picked up by userspace. In addition, some pstore backends may have low performance and provide a significant delay in reboot as a result.

This patch adds a printk.always_kmsg_dump kernel parameter (which can also be changed from userspace). Without it, the code will only be run on failure paths rather than on normal paths. The option can be enabled in environments where there's a desire to attempt to audit whether or not a reboot was cleanly requested or not.

---
 include/linux/kmsg_dump.h |    9 +++++++--
 kernel/printk.c           |    6 ++++++
 2 files changed, 13 insertions(+), 2 deletions(-)
include/linux/kmsg_dump.h +9 -2
@@ -15,13 +15,18 @@
 #include <linux/errno.h>
 #include <linux/list.h>
 
+/*
+ * Keep this list arranged in rough order of priority. Anything listed 
+after
+ * KMSG_DUMP_OOPS will not be logged by default unless 
+printk.always_kmsg_dump
+ * is passed to the kernel.
+ */
 enum kmsg_dump_reason {
-	KMSG_DUMP_OOPS,
 	KMSG_DUMP_PANIC,
+	KMSG_DUMP_OOPS,
+	KMSG_DUMP_EMERG,
 	KMSG_DUMP_RESTART,
 	KMSG_DUMP_HALT,
 	KMSG_DUMP_POWEROFF,
-	KMSG_DUMP_EMERG,
 };

/**

kernel/printk.c +4 -0
@@ -702,6 +702,9 @@ static bool printk_time = 0;  #endif  module_param_named(time, printk_time, bool, S_IRUGO | S_IWUSR);
 
+static bool always_kmsg_dump;
+module_param_named(always_kmsg_dump, always_kmsg_dump, bool, S_IRUGO | 
+S_IWUSR);
+
 /* Check if we have any console registered that can be called early in boot. */  static int have_callable_console(void)  { @@ -1732,6 +1735,9 @@ void kmsg_dump(enum kmsg_dump_reason reason)
 	unsigned long l1, l2;
 	unsigned long flags;
 

+ if (reason > KMSG_DUMP_PANIC && !always_kmsg_dump)
+ return;
+
/* Theoretically, the log could move on after we do this, but
there's not a lot we can do about that. The new messages will overwrite the start of what we dump. */
--
1.7.7.6