mm: Remove slab destructors from kmem_cache_create().
Slab destructors were no longer supported after Christoph's c59def9f222d44bb7e2f0a559f2906191a0862d7 change. They've been BUGs for both slab and slub, and slob never supported them either.
This rips out support for the dtor pointer from kmem_cache_create() completely and fixes up every single callsite in the kernel (there were about 224, not including the slab allocator definitions themselves, or the documentation references).
Perhaps this is something that should go in later, as it does cause a bit of churn across the tree. However, there shouldn't be anything here that causes trouble for git merges, as it's all pretty straightforward removal. I've placed it in a git tree to pull as it's quite large, and not terribly interesting.
If you want this single change, please pull from:
master.kernel.org:/pub/scm/linux/kernel/git/lethal/sh-2.6.git kmem_death
Which contains:
Paul Mundt (1):
mm: Remove slab destructors from kmem_cache_create().
arch/arm/plat-s3c24xx/dma.c | 2 +- arch/arm26/mm/memc.c | 4 ++-- arch/i386/mm/init.c | 3 +-- arch/ia64/ia32/ia32_support.c | 2 +- arch/powerpc/kernel/rtas_flash.c | 2 +- arch/powerpc/mm/hugetlbpage.c | 2 +- arch/powerpc/mm/init_64.c | 3 +-- arch/powerpc/platforms/cell/spufs/inode.c | 2 +- arch/sh/kernel/cpu/sh4/sq.c | 3 +-- arch/sh/mm/pmb.c | 2 +- arch/sparc64/mm/tsb.c | 3 +-- block/bsg.c | 2 +- block/ll_rw_blk.c | 6 +++--- drivers/acpi/osl.c | 2 +- drivers/block/aoe/aoeblk.c | 4 ++-- drivers/ieee1394/eth1394.c | 2 +- drivers/infiniband/core/mad.c | 1 - drivers/infiniband/hw/amso1100/c2_vq.c | 2 +- drivers/infiniband/hw/ehca/ehca_av.c | 2 +- drivers/infiniband/hw/ehca/ehca_cq.c | 2 +- drivers/infiniband/hw/ehca/ehca_main.c | 2 +- drivers/infiniband/hw/ehca/ehca_mrmw.c | 4 ++-- drivers/infiniband/hw/ehca/ehca_pd.c | 2 +- drivers/infiniband/hw/ehca/ehca_qp.c | 2 +- drivers/infiniband/ulp/iser/iscsi_iser.c | 2 +- drivers/kvm/mmu.c | 8 ++++---- drivers/md/raid5.c | 4 ++-- drivers/message/i2o/i2o_block.c | 3 +-- drivers/mtd/ubi/eba.c | 2 +- drivers/mtd/ubi/wl.c | 2 +- drivers/s390/block/dasd_devmap.c | 2 +- drivers/s390/scsi/zfcp_aux.c | 6 +++--- drivers/scsi/aic94xx/aic94xx_init.c | 4 ++-- drivers/scsi/libsas/sas_init.c | 2 +- drivers/scsi/qla2xxx/qla_os.c | 2 +- drivers/scsi/qla4xxx/ql4_os.c | 2 +- drivers/scsi/scsi.c | 2 +- drivers/scsi/scsi_lib.c | 4 ++-- drivers/scsi/scsi_tgt_lib.c | 2 +- drivers/usb/host/uhci-hcd.c | 2 +- drivers/usb/mon/mon_text.c | 2 +- fs/adfs/super.c | 4 ++-- fs/affs/super.c | 2 +- fs/afs/super.c | 3 +-- fs/befs/linuxvfs.c | 4 ++-- fs/bfs/inode.c | 4 ++-- fs/bio.c | 2 +- fs/block_dev.c | 2 +- fs/cifs/cifsfs.c | 10 +++++----- fs/coda/inode.c | 4 ++-- fs/configfs/mount.c | 2 +- fs/dcache.c | 4 ++-- fs/dcookies.c | 2 +- fs/dlm/lowcomms.c | 2 +- fs/dlm/memory.c | 2 +- fs/dnotify.c | 2 +- fs/dquot.c | 4 ++-- fs/ecryptfs/main.c | 2 +- fs/efs/super.c | 4 ++-- fs/eventpoll.c | 4 ++-- fs/ext2/super.c | 4 ++-- fs/ext3/super.c | 2 +- fs/ext4/super.c | 2 +- fs/fat/cache.c | 2 +- fs/fat/inode.c | 2 +- fs/fcntl.c | 2 +- fs/freevxfs/vxfs_super.c | 4 ++-- fs/fuse/dev.c | 2 +- fs/fuse/inode.c | 2 +- fs/gfs2/main.c | 6 +++--- fs/hfs/super.c | 2 +- fs/hfsplus/super.c | 2 +- fs/hpfs/super.c | 4 ++-- fs/hugetlbfs/inode.c | 2 +- fs/inode.c | 3 +-- fs/inotify_user.c | 4 ++-- fs/isofs/inode.c | 2 +- fs/jbd/journal.c | 8 +++----- fs/jbd/revoke.c | 4 ++-- fs/jbd2/journal.c | 8 +++----- fs/jbd2/revoke.c | 4 ++-- fs/jffs2/malloc.c | 18 +++++++++--------- fs/jffs2/super.c | 2 +- fs/jfs/jfs_metapage.c | 2 +- fs/jfs/super.c | 2 +- fs/locks.c | 2 +- fs/mbcache.c | 2 +- fs/minix/inode.c | 4 ++-- fs/namespace.c | 2 +- fs/ncpfs/inode.c | 4 ++-- fs/nfs/direct.c | 2 +- fs/nfs/inode.c | 4 ++-- fs/nfs/pagelist.c | 2 +- fs/nfs/read.c | 2 +- fs/nfs/write.c | 2 +- fs/nfsd/nfs4state.c | 8 ++++---- fs/ntfs/super.c | 10 +++++----- fs/ocfs2/dlm/dlmfs.c | 2 +- fs/ocfs2/dlm/dlmmaster.c | 2 +- fs/ocfs2/super.c | 2 +- fs/ocfs2/uptodate.c | 2 +- fs/openpromfs/inode.c | 2 +- fs/proc/inode.c | 4 ++-- fs/qnx4/inode.c | 2 +- fs/reiserfs/super.c | 2 +- fs/romfs/inode.c | 4 ++-- fs/smbfs/inode.c | 4 ++-- fs/smbfs/request.c | 2 +- fs/sysfs/mount.c | 2 +- fs/sysv/inode.c | 2 +- fs/udf/super.c | 2 +- fs/ufs/super.c | 4 ++-- fs/xfs/linux-2.6/kmem.h | 4 ++-- include/linux/i2o.h | 3 +-- include/linux/slab.h | 3 +-- ipc/mqueue.c | 2 +- kernel/fork.c | 18 +++++++++--------- kernel/nsproxy.c | 2 +- kernel/posix-timers.c | 2 +- kernel/user.c | 2 +- lib/idr.c | 2 +- lib/radix-tree.c | 2 +- mm/mempolicy.c | 4 ++-- mm/rmap.c | 2 +- mm/shmem.c | 2 +- mm/slab.c | 17 +++++++---------- mm/slob.c | 3 +-- mm/slub.c | 4 +--- net/bridge/br_fdb.c | 2 +- net/core/flow.c | 2 +- net/core/neighbour.c | 2 +- net/core/skbuff.c | 4 ++-- net/core/sock.c | 6 +++--- net/dccp/ackvec.c | 4 ++-- net/dccp/ccid.c | 2 +- net/dccp/ccids/lib/loss_interval.c | 2 +- net/dccp/ccids/lib/packet_history.c | 4 ++-- net/dccp/proto.c | 2 +- net/decnet/dn_route.c | 2 +- net/decnet/dn_table.c | 2 +- net/ipv4/fib_hash.c | 4 ++-- net/ipv4/fib_trie.c | 2 +- net/ipv4/inetpeer.c | 2 +- net/ipv4/ipmr.c | 2 +- net/ipv4/ipvs/ip_vs_conn.c | 2 +- net/ipv4/route.c | 2 +- net/ipv4/tcp.c | 2 +- net/ipv6/ip6_fib.c | 2 +- net/ipv6/route.c | 2 +- net/ipv6/xfrm6_tunnel.c | 2 +- net/netfilter/nf_conntrack_core.c | 2 +- net/netfilter/nf_conntrack_expect.c | 2 +- net/netfilter/xt_hashlimit.c | 2 +- net/rxrpc/af_rxrpc.c | 2 +- net/sctp/protocol.c | 4 ++-- net/socket.c | 3 +-- net/sunrpc/rpc_pipe.c | 2 +- net/sunrpc/sched.c | 4 ++-- net/tipc/handler.c | 2 +- net/xfrm/xfrm_input.c | 2 +- net/xfrm/xfrm_policy.c | 2 +- security/keys/key.c | 2 +- security/selinux/avc.c | 2 +- security/selinux/hooks.c | 2 +- security/selinux/ss/avtab.c | 2 +- 165 files changed, 247 insertions(+), 268 deletions(-)
Quoted 6 lines · Paul Mundt
Perhaps this is something that should go in later, as it does cause a bit of churn across the tree. However, there shouldn't be anything here that causes trouble for git merges, as it's all pretty straightforward removal. I've placed it in a git tree to pull as it's quite large, and not terribly interesting.
Ok, I thought twice about this, but
(a) I think most of the merging for 2.6.23 has been done
(b) there's really no point in holding on to that thing
(c) the patch looked beatiful in colorized version in gitk
so I decided I might as well just merge it. Looks fine to me, and there really is no point holding on to an interface that we've never really used, and that we don't even support any more.
Linus
Quoted 4 lines · Paul Mundt
Slab destructors were no longer supported after Christoph's
c59def9f222d44bb7e2f0a559f2906191a0862d7change. They've been BUGs for both slab and slub, and slob never supported them either.
SLOB of course did support destructors, otherwise it wouldn't have worked.
But good riddance to destructors nonetheless.
Signature · 1 line
Mathematics is the supreme nostalgia of our time.
Quoted 9 lines · Matt Mackall
Quoted 4 lines · Paul Mundt
Slab destructors were no longer supported after Christoph's
c59def9f222d44bb7e2f0a559f2906191a0862d7change. They've been BUGs for both slab and slub, and slob never supported them either.SLOB of course did support destructors, otherwise it wouldn't have worked.
Ah, I just noticed that there was no reference to the dtor in SLOB at all apart from the kmem_cache_create() op. It looks like they were called back in to and removed by an earlier patch, so the 'never' statement is disinformation on my part. Sorry for the confusion!
Please do not merge this patch. I will reuse the destructor field for the kmem_cache_ops in the slab defrag patchset. If you apply this patch then I need to readd the parameter later.
Quoted 3 lines · Christoph Lameter
Please do not merge this patch. I will reuse the destructor field for the
kmem_cache_opsin the slab defrag patchset. If you apply this patch then I need to readd the parameter later.
The vast majority of slabs won't have any cache ops, right? Wouldn't it be cleaner to have a kmem_set_cache_ops() for the few that do?
While we're on the subject, has anyone reviewed your defrag stuff?
I read through it a while back and my initial impression is a) it's awfully complex and invasive and b) it probably won't protect all workloads.
As this and several other vexing problems stem from the dcache design, I'd rather abandon the dcache out in the woods so we can replace it with something less troublesome.
Signature · 1 line
Mathematics is the supreme nostalgia of our time.