From ebcfc8a53450fd71443668191507770e17c409bd Mon Sep 17 00:00:00 2001 From: Brian Behlendorf Date: Tue, 4 Sep 2012 19:00:59 -0700 Subject: [PATCH] Disable page allocation warnings for ARC buffers Buffers for the ARC are normally backed by the SPL virtual slab. However, if memory is low, AND no slab objects are available, AND a new slab cannot be quickly constructed a new emergency object will be directly allocated. These objects can be as large as order 5 on a system with 4k pages. And because they are allocated with KM_PUSHPAGE, to avoid a potential deadlock, they are not allowed to initiate I/O to satisfy the allocation. This can result in the occasional allocation failure. However, since these allocations are allowed to block and perform operations such as memory compaction they will eventually succeed. Since this is not unexpected (just unlikely) behavior this patch disables the warning for the allocation failure. Signed-off-by: Brian Behlendorf Issue #465 --- module/zfs/zio.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/module/zfs/zio.c b/module/zfs/zio.c index fe2bdc867fea..6333c787d246 100644 --- a/module/zfs/zio.c +++ b/module/zfs/zio.c @@ -259,7 +259,7 @@ zio_buf_alloc(size_t size) ASSERT(c < SPA_MAXBLOCKSIZE >> SPA_MINBLOCKSHIFT); - return (kmem_cache_alloc(zio_buf_cache[c], KM_PUSHPAGE)); + return (kmem_cache_alloc(zio_buf_cache[c], KM_PUSHPAGE | KM_NODEBUG)); } /* @@ -275,7 +275,8 @@ zio_data_buf_alloc(size_t size) ASSERT(c < SPA_MAXBLOCKSIZE >> SPA_MINBLOCKSHIFT); - return (kmem_cache_alloc(zio_data_buf_cache[c], KM_PUSHPAGE)); + return (kmem_cache_alloc(zio_data_buf_cache[c], + KM_PUSHPAGE | KM_NODEBUG)); } void