linux-kernel - register_timer_hook fails on realtime kernel with CONFIG_PROFILE

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:	Sun, 06 Aug 2006 01:59:37 -0400
From:	John Kacur <jkacur@...ers.com>
To:	Thomas Gleixner <tglx@...utronix.de>, Ingo Molnar <mingo@...e.hu>
Cc:	john stultz <johnstul@...ibm.com>,
	Steven Rostedt <rostedt@...dmis.org>,
	Darren Hart <dvhltc@...ibm.com>,
	lkml <linux-kernel@...r.kernel.org>
Subject: register_timer_hook fails on realtime kernel with
	CONFIG_PROFILE_NMI

Hi,
I wrote a "Very Simple Timer Notification Module", which calls
register_timer_hook to register a function to be called using the timer
interrupt. This is essentially what OProfile would do in timer interrupt
mode. The standard kernel allows you to do this even if
CONFIG_PROFILE_NMI is defined but the realtime kernel does not. With the
realtime kernel register_timer_hook succeeds but the timer hooks is
never called. From patch-2.6.17-rt8 the following modifies
linux/include/linux/clockchips.h
+#ifndef CONFIG_PROFILE_NMI
+#define CLOCK_CAP_PROFILE	0x000008
+#else
+#define CLOCK_CAP_PROFILE	0x000000
+#endif

So, it's simple enough to work around, by just making CLOCK_CAP_PROFILE
0x000008 regardless of whether CONFIG_PROFILE_NMI is defined or not, but
I would like to understand the intention of the authors here. Is there a
good reason to disallow the timer interrupt mode of profiling if
CONFIG_PROFILE_NMI is defined? (other than arguing that NMI profiling is
superior)

I have included vstnm.tar.bz2 in case it is of any help examinig the
problem. Just do:
tar xjf vstnm.tar.bz2
cd vstnm
make
make install
then use the script load_vstnm to load the module and unload_vstnm to
unload the module. If the timer_tick function is called it will print a
rate limited message to the console that it was called.

Thank you

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/