On 06/06/2013 06:49 AM, Dave Chinner wrote: >> > How [patch 09/35]'s inode_lru_isolate() avoids this bug I don't know. >> > Perhaps it doesn't. > The LRU_RETRY cse is supposed to handle this. However, the LRU_RETRY > return code is now buggy and you've caught that. It'll need fixing. > My original code only had inode_lru_isolate() drop the lru lock, and > it would return LRU_RETRY which would restart the scan of the list > from the start, thereby avoiding those problems. > Yes, I have changed that, but I wasn't aware that your original intention for restarting from the beginning was to avoid such problems. And having only half the brain Andrew has, I didn't notice it myself. I will fix this somehow while trying to keep the behavior Mel insisted on; iow; not retrying forever. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>