Dear Raj, everyone I try to help....here it is.....CMIIW > I have a doubt regarding the working of the VMA. How is a program > executable's text region shared among processes that are running the > same program ? I would like to understand how finally the same pages > are shared between the 2 processes . Well, first of all, VMA structures are unique for each process. The exception is thread, which is sharing text region, global data segment (not the thread specific one). You must also notice, VMAs of different processes might map same page(s). The only scenario I can think, when two or more processes share VMA of text region (pay attention, you are asking about VMA, not the page or page frame) are in multi threading. Since they execute same code (in most cases, new thread are created with clone() using CLONE_VM), it makes sense to share the text region for efficiency. Code region is something that is read only too, so sharing them won't hurt :) Same pages (we are now talking about pages) can be mapped into different VMA of entirely different process, for example mmap()-ing shared libraries. Why it can happen? during file read, several pages landed into page cache. The pages on pages cache is always the first place to be looked up when reading a file, before directly fetching data block (if none is exist on page cache). So, successive read of the file (including mmap()) might just map the page into different VMA Rik, Greg, everyone....feel free to CMIIW regards Mulyadi -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
