[PR18457] Don't require rtld lock to compute DTV addr for static TLS

[Siddhesh Poyarekar]

On Fri, Jun 05, 2015 at 01:23:45AM -0300, Alexandre Oliva wrote:
> It is undisputably enough to fix the performance regression that in
> certain cases introduced new deadlocks to the pile of other possible
> deadlocks that variants of the testcase could exercise.  It fixes a
> small problem, and it's a trivial patch.
> 
> Sure, it doesn't fix a much larger preexisting problem, that's being
> misrepresented as a TLS issue, but that is actually a far more pervasive
> dynamic loader problem.  I have no interest or time in tackling that.
> 
> Sure, if you guys want to leave the regression unfixed until someone
> figures out some way to fix a larger but mostly unrelated problem, it's
> your choice.

No, IMO we should fix this regression in the interim and your approach
in general seems fine to me.  I intended to point out that it doesn't
fix the larger problem and I think we're on the same page on that.

I haven't reviewed the patch in detail nor your conversation with
Torvald yet, so I can't comment on whether Torvald and I agree,
although from a quick skimming of the thread ISTM that we're talking
about different things.

Siddhesh
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 473 bytes
Desc: not available
URL: <http://sourceware.org/pipermail/libc-alpha/attachments/20150605/8db5ae4e/attachment.sig>