Messages in this thread | | | Date | Thu, 19 Oct 2000 17:28:57 -0700 (PDT) | From | Linus Torvalds <> | Subject | Re: 2.4.0-test10-pre3:Oops in mm/filemap.c:filemap_write_pa |
| |
On 20 Oct 2000, Trond Myklebust wrote: > > > invalidate_inode_pages nfs_zap_caches nfs_lock fcntl_setlk > > do_fcntl sys_fcntl > > > So I guess that NFS locking is really bad if the region is > > mmapped! > > Yep, but that's a symptom, not a cause. We want to be able to run > invalidate_inode_pages() safely at any moment, since the need can be > triggered externally (because the server and client page caches > disagree).
Well, the thing is, that if somebody has a page mapped, there's nothing we can do if the server and client disagrees.
We just cannot invalidate the page - people would actually lose data, and would lose major local consistency guarantees that UNIX filesystems are supposed to get.
So it's a matter of one sort of consistency against another - and I think the local consistency requirements are the stricter ones, considering that we don't even _know_ whether the server has really changed those pages or not..
Linus
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
| |