On Sat, 2009-02-28 at 08:45 +0000, Fridtjof Busse wrote:
> I agree that this is not a critical error, but nonethelesse a kernel
> oops is not exactly a "nice to have".
That was my point in mentioning TEST_SUSPEND_SECONDS. We are discussing
increasing the current 5-second limit to avoid these false alarms.
Before we can do that however, we have to look at a wide range of
systems and scenarios so as to choose a value that will avoid false
alarms but not miss real problems.
On Sat, 2009-02-28 at 08:45 +0000, Fridtjof Busse wrote:
> I agree that this is not a critical error, but nonethelesse a kernel
> oops is not exactly a "nice to have".
That was my point in mentioning TEST_SUSPEND_ SECONDS. We are discussing
increasing the current 5-second limit to avoid these false alarms.
Before we can do that however, we have to look at a wide range of
systems and scenarios so as to choose a value that will avoid false
alarms but not miss real problems.