3b8
Subj : Re: deadlock avoidance
To   : comp.programming.threads
From : Pierre Barbier de Reuille
Date : Tue Jul 05 2005 01:16 pm

Satendra a écrit :
> The article is sure interesting .... but at a few places ....
> The author points out "C functions using global or "static" variables
> are surely not reentrant .... " but in my humble opinion there can be
> reentrant functions which may modify static or global variables for
> many purposes.  "Every access to shared data must be guarded with a
> mutex acquisition .. " I disagree to this too .. operation like
> unconditional assignment etc are atomic enough to be done without a
> mutex lock ... I feel use of "Every" " Always" etc need a relook
> 

Well, there are for sure some imprecision ... but I feel they should be
lifted by the other parts of the articles. IMO, an introduction is
allowed to oversee some details, and these kind of details were ok for me :)

Pierre

.

0