On 5/22/22 12:32 AM, Christoph Hellwig wrote: > So why don't we just define a __file_remove_privs that gets the iocb > flags passed an an extra argument, and then make file_remove_privs pass > 0 flags, and maybe also add a kiocb_remove_privs wrapper for callers > that have the kiocb. Same for the timestamp update, btw. That seems > less churn and less code overall. I introduced a __file_remove_privs() function with an additional parameter that takes the iocb_flags.
