Messages in this thread | | | Date | Wed, 22 Feb 2006 10:39:50 +1100 | From | David Gibson <> | Subject | Re: RFC: Block reservation for hugetlbfs |
| |
On Tue, Feb 21, 2006 at 03:18:59PM +1100, Nick Piggin wrote: > David Gibson wrote: > >These days, hugepages are demand-allocated at first fault time. > >There's a somewhat dubious (and racy) heuristic when making a new > >mmap() to check if there are enough available hugepages to fully > >satisfy that mapping. > > > >A particularly obvious case where the heuristic breaks down is where a > >process maps its hugepages not as a single chunk, but as a bunch of > >individually mmap()ed (or shmat()ed) blocks without touching and > >instantiating the blocks in between allocations. In this case the > >size of each block is compared against the total number of available > >hugepages. It's thus easy for the process to become overcommitted, > >because each block mapping will succeed, although the total number of > >hugepages required by all blocks exceeds the number available. In > >particular, this defeats such a program which will detect a mapping > >failure and adjust its hugepage usage downward accordingly. > > > >The patch below is a draft attempt to address this problem, by > >strictly reserving a number of physical hugepages for hugepages inodes > >which have been mapped, but not instatiated. MAP_SHARED mappings are > >thus "safe" - they will fail on mmap(), not later with a SIGBUS. > >MAP_PRIVATE mappings can still SIGBUS. > > > >This patch appears to address the problem at hand - it allows DB2 to > >start correctly, for instance, which previously suffered the failure > >described above. I'm almost certain I'm missing some locking or other > >synchronization - I am entirely bewildered as to what I need to hold > >to safely update i_blocks as below. Corrections for my ignorance > >solicited... > > > >Signed-off-by: David Gibson <dwg@au1.ibm.com> > > > > This introduces > tree_lock(r) -> hugetlb_lock > > And we already have > hugetlb_lock -> lru_lock > > So we now have tree_lock(r) -> lru_lock, which would deadlock > against lru_lock -> tree_lock(w), right? > > From a quick glance it looks safe, but I'd _really_ rather not > introduce something like this.
Urg.. good point. I hadn't even thought of that consequence - I was more worried about whether I need i_lock or i_mutex to protect my updates to i_blocks.
Would hugetlb_lock -> tree_lock(r) be any preferable (I think that's a possible alternative).
-- David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson - 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/
| |