Messages in this thread Patch in this message |  | | Date | Mon, 14 Mar 2011 01:08:47 -0700 (PDT) | From | Hugh Dickins <> | Subject | [PATCH] thp+memcg-numa: fix BUG at include/linux/mm.h:370! |
| |
THP's collapse_huge_page() has an understandable but ugly difference in when its huge page is allocated: inside if NUMA but outside if not. It's hardly surprising that the memcg failure path forgot that, freeing the page in the non-NUMA case, then hitting a VM_BUG_ON in get_page() (or even worse, using the freed page).
Signed-off-by: Hugh Dickins <hughd@google.com> ---
mm/huge_memory.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
--- 2.6.38-rc8/mm/huge_memory.c 2011-03-08 09:27:16.000000000 -0800 +++ linux/mm/huge_memory.c 2011-03-13 18:26:21.000000000 -0700 @@ -1762,6 +1762,10 @@ static void collapse_huge_page(struct mm #ifndef CONFIG_NUMA VM_BUG_ON(!*hpage); new_page = *hpage; + if (unlikely(mem_cgroup_newpage_charge(new_page, mm, GFP_KERNEL))) { + up_read(&mm->mmap_sem); + return; + } #else VM_BUG_ON(*hpage); /* @@ -1781,12 +1785,12 @@ static void collapse_huge_page(struct mm *hpage = ERR_PTR(-ENOMEM); return; } -#endif if (unlikely(mem_cgroup_newpage_charge(new_page, mm, GFP_KERNEL))) { up_read(&mm->mmap_sem); put_page(new_page); return; } +#endif /* after allocating the hugepage upgrade to mmap_sem write mode */ up_read(&mm->mmap_sem);
|  |