Re: RFC: should a va_bytes option be added to vn_getsize_locked()?
- In reply to: Rick Macklem : "RFC: should a va_bytes option be added to vn_getsize_locked()?"
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Sun, 10 Mar 2024 18:52:17 UTC
On Sat, Mar 09, 2024 at 04:59:49PM -0800, Rick Macklem wrote: > Hi, > > I would like to compare va_size to va_bytes in vn_generic_copy_file_range(), > as a heuristic to check for a sparse file (only works for non-compressed > file systems). > > The call to VOP_GETATTR(invp, ..) was replaced by vn_getsize_locked() > in vn_generic_copy_file_range(). > > To get va_bytes I can either modify the code to again use VOP_GETATTR() > or I could add an additional return argument to vn_getsize_locked(). > Since vn_getsize_locked() is descibed as a first step towards not using > VOP_GETATTR() it sounds like adding an agument to vn_getsize_locked() > is not the preferred alternative, but I thought I'd ask. For me it sounds as very specific usage. You might be better served by directly using VOP_GETATTR() then IMO.