tech-userlevel archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: fixing libpthread noload, another approach

On Sun, Feb 3, 2013 at 1:28 AM, Emmanuel Dreyfus <> 
> A first proposal to fix the problem has been to implement minimal mutex
> and condvar operations in libc stubs: if they initialize enough fields,
> then operations can resume safely after libpthread is loaded. The
> drawback of this approach is that there is a performance hit for
> programs that are not linked with -lpthread.
> Here is another proposal:

As Alan suggested in,
could you please write a document that answers all those questions for
the various alternatives you are considering, before writing the code?
 (Such document will be helpful to decide what to do now, and later in
the future to showcase that various options were considered and to
justify why a particular option was chosen/discarded.)

Julio Merino / @jmmv

Home | Main Index | Thread Index | Old Index