Messages in this thread | | | Subject | Re: [git patch] fuse fixes | From | Miklos Szeredi <> | Date | Thu, 27 Apr 2006 16:38:50 +0200 |
| |
> >> This function is called from everywhere, and so, it looks like it should > >> use SLAB_NOFS rather than SLAB_KERNEL. I would audit every GFP_KERNEL > >> and SLAB_KERNEL usage, and consider replacing with SLAB_NOFS or GFP_NOFS. > > > > GFP_NOFS doesn't make much sense, since mm never calls back into FUSE > > anyway: FUSE writes through the page-cache, and hence never dirties > > any pages. > > > > I'll add a comment to fuse_request_alloc(). > > If you're using loop, particularly something insane like swapping over > loop, "the path" will certainly want to know that its passing through > the VFS layer, regardless of specific page cache behavior, AFAICS.
IIRC "loop" decouples the base filesystem from any user of the loop device by a kernel thread and mediating buffers.
Miklos - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |