On 7/22/05, Mulyadi Santosa <mulyadi.santosa@xxxxxxxxx> wrote: > > Note that "deferred" allocation doesn't always happen, especially when > you use GFP_ATOMIC. Deferred allocation only happens if memory > allocator can not satisfy your request due to lack of free RAM, so it > sleeps and wait until some amount of memory is released.... > You are right Mulyadi, but I want to make it clear that waiting and sleeping for the memory allocation (when no free memory available) when you use the flag GFP_KERNEL, when GFP_ATOMIC flag is used; it mentions that allocate memory from the pool where kernel keeps memory for fullfilment of fast memory allocation requirement, so in case of GFP_ATOMIC if no memory available then your allocation request will be rejected immediately rather waiting for memory to be freed .... *_ATOMIC* operations are __always __ non-sleeping operations where waiting is not allowed -- Fawad Lateef -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
