Messages in this thread | | | Date | Thu, 1 May 2008 16:30:39 -0700 | From | Venki Pallipadi <> | Subject | Re: huge gcc 4.1.{0,1} __weak problem |
| |
On Thu, May 01, 2008 at 04:21:16PM -0700, Tom Rini wrote: > On Thu, May 01, 2008 at 03:42:38PM -0700, Andrew Morton wrote: > > [snip] > > Is there some vaguely maintainable workaround we can do? If the problem > > only affects completely-empty weak functions then we could put something in > > them to make them non-empty? > > My memory is a tiny bit hazy (it was a while ago), but no, it's not just > empty functions (again, I _think_ I hit it with a generic vs arch weak > function). >
Other thing we observed was: this does not just depend on the __weak function definition. It also depends on where the function is called from.
__weak function with single return statement, did not get inlined when called from say
caller() { function(); }
but got inlined when called as in
caller() { for (;;) { function(); } }
Thanks, Venki
| |