> > However, we should look at how FreeBSD has implemented posix_fallocate > and see if we can incorporate that code. Or at least come up with a > compatable implementation so that there is a possibility of moving disks > between these systems.
FreeBSD has a default implementation of vop_fallocate() which zero out the allocated blocks. This code is independent of underlying filesystem and is executed for posix_fallocate() when underlying filesystem doesn't implement vop_fallocate(). Until we implement vop_fallocate() for ffs with one of the approaches discussed above (possibly without filling zeros to the allocated blocks), should I implement genfs_fallocate() with an approach similar to FreeBSD? Regards, Hrishikesh On Tue, Nov 19, 2019 at 11:10 PM Mouse <mo...@rodents-montreal.org> wrote: > >> You don't have to know where to write. Strictly speaking you only > >> need to know that the space is reserved. > > Right. > > > One thing that I am missing here is: if you would keep info about > reservatio$ > > You don't need to know history. You just need to know how many blocks > are available to be newly allocated each of two ways. > > In your example, you have 1200 blocks available and you posix_fallocate > 1000 of them. Then you have 1200 blocks marked free with a free-block > count of only 200; how it got to that state is not relevant. (Whether > you need to have the number 1200 stored somewhere also - or instead - > is a question I would have to dig into FFS internals at some length to > answer, and indeed the answer would likely depend on the details of how > all this got implemented.) > > Obviously, anything that deals with free blocks needs to be updated, > and you need two forms of allocate-a-block, one for replacing > "allocated hole" blocks with real content and one for allocating new > blocks. > > > If you request block-allocator from fallocate to reserve these block, to > mak$ > > No; there's no need to allocate specific blocks. All you need to do is > ensure that, when it comes time to put real data at that spot in the > file, a block will be available. Which block it is is not important. > > /~\ The ASCII Mouse > \ / Ribbon Campaign > X Against HTML mo...@rodents-montreal.org > / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B >
