Messages in this thread | | | Date | Tue, 01 Apr 2008 18:14:50 +0300 | From | Adrian Hunter <> | Subject | Re: [RFC PATCH 20/26] UBIFS: add VFS operations |
| |
Artem Bityutskiy wrote: > Pekka Enberg wrote: >>> Yeah, this is also a guarding thing. When a dirty page is released >>> the budget which was allocated for it has to be freed. If this function >>> is called, then the budget was not freed, which must never happen. >> >> So what guarantees that no one calls invalidate_complete_page() or >> fallback_migrate_page(), for example? > > At the first glance it looks like it might be called, but for clean pages, > which is not the problem, but there is this assert which may give fake > alarm. Need to look closer at this. Thanks for the note.
Dirty pages are not released.
In UBIFS, clean pages do not have PagePrivate(page) set and so releasepage() is not called.
| |