From mboxrd@z Thu Jan 1 00:00:00 1970 From: Josh Durgin Subject: Re: [PATCH 1/4] rbd: allocate image requests with a slab allocator Date: Thu, 02 May 2013 09:19:39 -0700 Message-ID: <5182921B.7000900@inktank.com> References: <51818A5C.1080903@inktank.com> <51818AA5.6070200@inktank.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from mail-pb0-f42.google.com ([209.85.160.42]:43938 "EHLO mail-pb0-f42.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757990Ab3EBQTN (ORCPT ); Thu, 2 May 2013 12:19:13 -0400 Received: by mail-pb0-f42.google.com with SMTP id up7so427862pbc.1 for ; Thu, 02 May 2013 09:19:13 -0700 (PDT) In-Reply-To: <51818AA5.6070200@inktank.com> Sender: ceph-devel-owner@vger.kernel.org List-ID: To: Alex Elder Cc: ceph-devel@vger.kernel.org It looks like the GFP flags used throughout request creation could be cleanup up (there's a mix of ATOMIC, KERNEL, and NOIO). I just noticed since this changes a GFP_ATOMIC to GFP_KERNEL. Making them consistent can be a later patch though. Reviewed-by: Josh Durgin On 05/01/2013 02:35 PM, Alex Elder wrote: > Create a slab cache to manage rbd_img_request allocation. Nothing > too fancy at this point--we'll still initialize everything at > allocation time (no constructor) > > This is part of: > http://tracker.ceph.com/issues/3926 > > Signed-off-by: Alex Elder > --- > drivers/block/rbd.c | 39 ++++++++++++++++++++++++++++++++++----- > 1 file changed, 34 insertions(+), 5 deletions(-) > > diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c > index a6e5fe3..005c397 100644 > --- a/drivers/block/rbd.c > +++ b/drivers/block/rbd.c > @@ -40,6 +40,7 @@ > #include > #include > #include > +#include > > #include "rbd_types.h" > > @@ -344,6 +345,8 @@ static DEFINE_SPINLOCK(rbd_dev_list_lock); > static LIST_HEAD(rbd_client_list); /* clients */ > static DEFINE_SPINLOCK(rbd_client_list_lock); > > +static struct kmem_cache *rbd_img_request_cache; > + > static int rbd_img_request_submit(struct rbd_img_request *img_request); > > static void rbd_dev_device_release(struct device *dev); > @@ -1821,7 +1824,7 @@ static struct rbd_img_request *rbd_img_request_create( > { > struct rbd_img_request *img_request; > > - img_request = kmalloc(sizeof (*img_request), GFP_ATOMIC); > + img_request = kmem_cache_alloc(rbd_img_request_cache, GFP_KERNEL); > if (!img_request) > return NULL; > > @@ -1884,7 +1887,7 @@ static void rbd_img_request_destroy(struct kref *kref) > if (img_request_child_test(img_request)) > rbd_obj_request_put(img_request->obj_request); > > - kfree(img_request); > + kmem_cache_free(rbd_img_request_cache, img_request); > } > > static bool rbd_img_obj_end_request(struct rbd_obj_request *obj_request) > @@ -4992,6 +4995,26 @@ static void rbd_sysfs_cleanup(void) > device_unregister(&rbd_root_dev); > } > > +static int rbd_slab_init(void) > +{ > + rbd_assert(!rbd_img_request_cache); > + rbd_img_request_cache = kmem_cache_create("rbd_img_request", > + sizeof (struct rbd_img_request), > + __alignof__(struct rbd_img_request), > + 0, NULL); > + if (rbd_img_request_cache) > + return 0; > + > + return -ENOMEM; > +} > + > +static void rbd_slab_exit(void) > +{ > + rbd_assert(rbd_img_request_cache); > + kmem_cache_destroy(rbd_img_request_cache); > + rbd_img_request_cache = NULL; > +} > + > static int __init rbd_init(void) > { > int rc; > @@ -5001,16 +5024,22 @@ static int __init rbd_init(void) > > return -EINVAL; > } > - rc = rbd_sysfs_init(); > + rc = rbd_slab_init(); > if (rc) > return rc; > - pr_info("loaded " RBD_DRV_NAME_LONG "\n"); > - return 0; > + rc = rbd_sysfs_init(); > + if (rc) > + rbd_slab_exit(); > + else > + pr_info("loaded " RBD_DRV_NAME_LONG "\n"); > + > + return rc; > } > > static void __exit rbd_exit(void) > { > rbd_sysfs_cleanup(); > + rbd_slab_exit(); > } > > module_init(rbd_init); >