Date: Mon, 19 Aug 2002 12:56:51 -0700 From: Terry Lambert <tlambert2@mindspring.com> To: Poul-Henning Kamp <phk@freebsd.org> Cc: current@freebsd.org, hackers@freebsd.org Subject: Re: phk's JKH list Message-ID: <3D614D83.3B08815B@mindspring.com> References: <3592.1029771744@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
Poul-Henning Kamp wrote: > I've started to type in my mental sticky notes, have at it: > > http://people.freebsd.org/~phk/TODO/ | Kernel stack depth warning. | | Write a function which checks the amount of kernel stack used and | print a warning if it exceeds a sysctl-able limit. | The inserted calls should compile to nothing unless the right kernel | options is used. | Instrument the kernel strategically and examine the result. If this is just for testing purposes, then you probably want to look at `-finstrument-functions'; by ensuring that stacks are all aligned on an 8K or so boundary, you can write small assembly functions named: void __cyg_profile_func_enter (void *this_fn, void *call_site); void __cyg_profile_func_exit (void *this_fn, void *call_site); And check the stack pointer against the previous alignment address (example: stack address mod 8K := current stack depth). I've used these for non-statistical profiling (e.g. recording specific arc's in gprof-usable buckets for kernel profiling, rather than relying on a statistics clock and the current PC), but the principle should work the same for this purpose. -- Terry To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3D614D83.3B08815B>