no, when i do N(N >= 2) blit from N tmp surface to one main surface, the display maybe be confused and the cpu will reach 100%. after i add lock to every DFB interface(Blit, StretchBlit, ....), all works well.
I don't known which one should add lock in multithreaded env. > Hi, > While you are doing some blitting functions, there should be a > hardware_lock in the routine. > This will lock the source and destination surface. > > Is it what you want? > > 2007/12/26, Deng XueFeng <[EMAIL PROTECTED]>: > > Hi > > > > I have some question about the lock use in the multihreaded apps. > > I use directfb 1.0.1 on VIA nano-itx board. > > I got two main surface(graphics surface and video surface) from two > > layer. > > and i use some tmp surface. > > > > then the question are: > > > > 1. when i do two blit at the same time ( one tmp surface to main > > graphics surface, and another to main video graphics), should i use mutex > > lock.? one lock or two ? > > > > 2. when i do two blit from two tmp surface to one main surface, each > > blit overide some part of the main surface. should i use mutex lock? > > > > 3. when i do blit from tmp surface to tmp surface ,should i use mutex > > lock ? > > > > thanks very much. > > > > > > > > > > _______________________________________________ > > directfb-dev mailing list > > directfb-dev@directfb.org > > http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev > > -- Deng XueFeng <[EMAIL PROTECTED]> _______________________________________________ directfb-dev mailing list directfb-dev@directfb.org http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
