Chuck Silvers <[email protected]> wrote: > what would you do differently based on the result of that check? > and in what context would this check be performed? > the check may or may not be useful for your purpose, > depend on what you want to do with it.
This is for PUFFS page cache write-once feature I am working on. Right now data enters the page cache while we do not know if storage backend is available. Cache flush may later fail because of EDQUOT or ESPACE, which means the process cannot exit: it remains forever in state DE. I plan to allocate storage backend by reading and writing the affected area in the file, but I want to do it only once: if the file area is already mapped in memory, it means I already did the operation and want data to go to the page cache without the write-once operation. -- Emmanuel Dreyfus http://hcpnet.free.fr/pubz [email protected]
