All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Steve Wise" <swise-7bPotxP6k4+P2YhJcF5u+vpXobYPEAuW@public.gmane.org>
To: 'Moni Shoua' <monis-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>,
	dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	matanb-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org,
	leon-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org,
	majd-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org,
	liranl-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org,
	'Kamal Heib' <kamalh-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>,
	'Amir Vadai' <amirv-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Subject: RE: [PATCH rdma-next 11/29] IB/rxe: Allocation pool for RDMA objects
Date: Wed, 15 Jun 2016 11:17:06 -0500	[thread overview]
Message-ID: <011101d1c721$5c22d590$146880b0$@opengridcomputing.com> (raw)
In-Reply-To: <1464886657-14258-12-git-send-email-monis-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>


 
> Manage and allocate pool of objects with given limit on number of
> elements.  Gets parameters from rxe_type_info. Pool elements are
> allocated out of a slab cache.  Objects that are using this facility
> are: PD, QP, SRQ, CQ, MR, FMR, MW, etc.
> 
> Signed-off-by: Kamal Heib <kamalh-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
> Signed-off-by: Amir Vadai <amirv-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
> Signed-off-by: Moni Shoua <monis-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
> Reviewed-by: Haggai Eran <haggaie-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
> ---
>  drivers/infiniband/hw/rxe/rxe_pool.c | 510
> +++++++++++++++++++++++++++++++++++
>  drivers/infiniband/hw/rxe/rxe_pool.h | 164 +++++++++++
>  2 files changed, 674 insertions(+)
>  create mode 100644 drivers/infiniband/hw/rxe/rxe_pool.c
>  create mode 100644 drivers/infiniband/hw/rxe/rxe_pool.h
> 
> diff --git a/drivers/infiniband/hw/rxe/rxe_pool.c
> b/drivers/infiniband/hw/rxe/rxe_pool.c
> new file mode 100644
> index 0000000..5a7da6b
> --- /dev/null
> +++ b/drivers/infiniband/hw/rxe/rxe_pool.c
> @@ -0,0 +1,510 @@
> +/*
> + * Copyright (c) 2016 Mellanox Technologies Ltd. All rights reserved.
> + * Copyright (c) 2015 System Fabric Works, Inc. All rights reserved.
> + *
> + * This software is available to you under a choice of one of two
> + * licenses.  You may choose to be licensed under the terms of the GNU
> + * General Public License (GPL) Version 2, available from the file
> + * COPYING in the main directory of this source tree, or the
> + * OpenIB.org BSD license below:
> + *
> + *	   Redistribution and use in source and binary forms, with or
> + *	   without modification, are permitted provided that the following
> + *	   conditions are met:
> + *
> + *		- Redistributions of source code must retain the above
> + *		  copyright notice, this list of conditions and the following
> + *		  disclaimer.
> + *
> + *		- Redistributions in binary form must reproduce the above
> + *		  copyright notice, this list of conditions and the following
> + *		  disclaimer in the documentation and/or other materials
> + *		  provided with the distribution.
> + *
> + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
> + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
> + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
> + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
> HOLDERS
> + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
> + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
> + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
> + * SOFTWARE.
> + */
> +
> +#include "rxe.h"
> +#include "rxe_loc.h"
> +
> +/* info about object pools
> + * note that mr, fmr and mw share a single index space
> + * so that one can map an lkey to the correct type of object
> + */
> +struct rxe_type_info rxe_type_info[RXE_NUM_TYPES] = {
> +	[RXE_TYPE_UC] = {
> +		.name		= "uc",
> +		.size		= sizeof(struct rxe_ucontext),
> +	},
> +	[RXE_TYPE_PD] = {
> +		.name		= "pd",
> +		.size		= sizeof(struct rxe_pd),
> +	},
> +	[RXE_TYPE_AH] = {
> +		.name		= "ah",
> +		.size		= sizeof(struct rxe_ah),
> +		.flags		= RXE_POOL_ATOMIC,
> +	},
> +	[RXE_TYPE_SRQ] = {
> +		.name		= "srq",
> +		.size		= sizeof(struct rxe_srq),
> +		.flags		= RXE_POOL_INDEX,
> +		.min_index	= RXE_MIN_SRQ_INDEX,
> +		.max_index	= RXE_MAX_SRQ_INDEX,
> +	},
> +	[RXE_TYPE_QP] = {
> +		.name		= "qp",
> +		.size		= sizeof(struct rxe_qp),
> +		.cleanup	= rxe_qp_cleanup,
> +		.flags		= RXE_POOL_INDEX,
> +		.min_index	= RXE_MIN_QP_INDEX,
> +		.max_index	= RXE_MAX_QP_INDEX,
> +	},
> +	[RXE_TYPE_CQ] = {
> +		.name		= "cq",
> +		.size		= sizeof(struct rxe_cq),
> +		.cleanup	= rxe_cq_cleanup,
> +	},
> +	[RXE_TYPE_MR] = {
> +		.name		= "mr",
> +		.size		= sizeof(struct rxe_mem),
> +		.cleanup	= rxe_mem_cleanup,
> +		.flags		= RXE_POOL_INDEX,
> +		.max_index	= RXE_MAX_MR_INDEX,
> +		.min_index	= RXE_MIN_MR_INDEX,
> +	},
> +	[RXE_TYPE_FMR] = {
> +		.name		= "fmr",
> +		.size		= sizeof(struct rxe_mem),
> +		.cleanup	= rxe_mem_cleanup,
> +		.flags		= RXE_POOL_INDEX,
> +		.max_index	= RXE_MAX_FMR_INDEX,
> +		.min_index	= RXE_MIN_FMR_INDEX,
> +	},
> +	[RXE_TYPE_MW] = {
> +		.name		= "mw",
> +		.size		= sizeof(struct rxe_mem),
> +		.flags		= RXE_POOL_INDEX,
> +		.max_index	= RXE_MAX_MW_INDEX,
> +		.min_index	= RXE_MIN_MW_INDEX,
> +	},
> +	[RXE_TYPE_MC_GRP] = {
> +		.name		= "mc_grp",
> +		.size		= sizeof(struct rxe_mc_grp),
> +		.cleanup	= rxe_mc_cleanup,
> +		.flags		= RXE_POOL_KEY,
> +		.key_offset	= offsetof(struct rxe_mc_grp, mgid),
> +		.key_size	= sizeof(union ib_gid),
> +	},
> +	[RXE_TYPE_MC_ELEM] = {
> +		.name		= "mc_elem",
> +		.size		= sizeof(struct rxe_mc_elem),
> +		.flags		= RXE_POOL_ATOMIC,
> +	},
> +};
> +

Perhaps the above slab names should be prefixed with "rxe-"?  EG: "rxe-cq",
"rxe-qp", etc.  They look very non-descriptive when looking at them in slabinfo:

[root@stevo2 linux-2.6]# head /proc/slabinfo
slabinfo - version: 2.1
# name            <active_objs> <num_objs> <objsize> <objperslab> <pagesperslab>
: tunables <limit> <batchcount> <sharedfactor> : slabdata <active_slabs>
<num_slabs> <sharedavail>
uc                     0      0    352   11    1 : tunables   54   27    8 :
slabdata      0      0      0
cq                     6     34    240   17    1 : tunables  120   60    8 :
slabdata      2      2      0
qp                     4      6   1920    2    1 : tunables   24   12    8 :
slabdata      3      3      0
pd                     4     36    112   36    1 : tunables  120   60    8 :
slabdata      1      1      0


Also when using rxe with nvmf (among other bugs I'm chasing down), when you
unload ib_rxe, there are allocated cq objects left in the the slab cache:

kmem_cache_destroy cq: Slab cache still has objects
CPU: 5 PID: 4147 Comm: rmmod Tainted: G            E   4.7.0-rc2-nvmf-all.3+rxe+
#51
Hardware name: Supermicro X9DR3-F/X9DR3-F, BIOS 3.2a 07/09/2015
 0000000000000000 ffff881007b33d98 ffffffff812d1359 ffff88103bd65098
 ffffea00378983d0 ffff881007b33da8 ffff881078b32240 ffff881007b33df8
 ffffffff8117619c ffff881007b33da8 ffff881007b33da8 ffff88103bec6960
Call Trace:
 [<ffffffff812d1359>] dump_stack+0x51/0x78
 [<ffffffff8117619c>] kmem_cache_destroy+0x12c/0x150
 [<ffffffffa04c62cc>] rxe_cache_exit+0x1c/0x40 [ib_rxe]
 [<ffffffffa04ce4cf>] rxe_module_exit+0x13/0x23 [ib_rxe]
 [<ffffffff810e1d15>] SyS_delete_module+0x185/0x1d0
 [<ffffffff8100278e>] ? syscall_trace_enter_phase2+0x6e/0x190
 [<ffffffff81002915>] ? syscall_trace_enter+0x65/0x70
 [<ffffffff81002d4d>] do_syscall_64+0x6d/0x160
 [<ffffffff8161857c>] entry_SYSCALL64_slow_path+0x25/0x25
rxe: unloaded

Steve.

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

  parent reply	other threads:[~2016-06-15 16:17 UTC|newest]

Thread overview: 67+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-06-02 16:57 [PATCH rdma-next 00/29] Soft RoCE driver Moni Shoua
     [not found] ` <1464886657-14258-1-git-send-email-monis-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-06-02 16:57   ` [PATCH rdma-next 01/29] IB/rxe: IBA header types and methods Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 02/29] IB/rxe: Bit mask and lengths declaration for different opcodes Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 03/29] IB/rxe: Default rxe device and port parameters Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 04/29] IB/rxe: External interface to lower level modules Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 05/29] IB/rxe: Misc local interfaces between files Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 06/29] IB/rxe: Add maintainer for rxe driver Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 07/29] IB/rxe: Work request's opcode information table Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 08/29] IB/rxe: User/kernel shared queues infrastructure Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 09/29] IB/rxe: Common user/kernel queue implementation Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 10/29] IB/rxe: Interface to ib_core Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 11/29] IB/rxe: Allocation pool for RDMA objects Moni Shoua
     [not found]     ` <1464886657-14258-12-git-send-email-monis-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-06-15 16:17       ` Steve Wise [this message]
2016-06-15 16:45         ` Steve Wise
2016-06-15 16:54           ` Moni Shoua
2016-06-19 15:01           ` Yonatan Cohen
     [not found]             ` <8806657f-7aaa-03a3-8cb4-d04ad03a4822-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-06-20 21:18               ` Steve Wise
2016-06-15 16:52         ` Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 12/29] IB/rxe: RXE tasks handling Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 13/29] IB/rxe: Address vector manipulation functions Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 14/29] IB/rxe: Shared Receive Queue (SRQ) " Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 15/29] IB/rxe: Completion Queue (CQ) " Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 16/29] IB/rxe: Queue Pair (QP) handling Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 17/29] IB/rxe: Memory Region (MR) manioulation Moni Shoua
     [not found]     ` <1464886657-14258-18-git-send-email-monis-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-06-03  5:24       ` Christoph Hellwig
     [not found]         ` <20160603052440.GA31196-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2016-06-05 13:25           ` Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 18/29] IB/rxe: Add multicast infrastructure Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 19/29] IB/rxe: Received packets handling Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 20/29] IB/rxe: Completion handling Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 21/29] IB/rxe: QP request handling Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 22/29] IB/rxe: QP response handling Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 23/29] IB/rxe: Dummy DMA callbacks for RXE device Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 24/29] IB/rxe: Invariant CRC implementation Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 25/29] IB/rxe: Module init hooks Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 26/29] IB/rxe: Interface to netdev stack Moni Shoua
     [not found]     ` <1464886657-14258-27-git-send-email-monis-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-06-02 17:31       ` Jason Gunthorpe
     [not found]         ` <20160602173158.GA17320-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2016-06-05  9:42           ` Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 27/29] IB/rxe: sysfs interface to RXE Moni Shoua
     [not found]     ` <1464886657-14258-28-git-send-email-monis-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-06-03  6:00       ` Bart Van Assche
     [not found]         ` <47f2db91-1d9a-148b-f57c-1b0a3e62d95b-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-06-05  9:34           ` Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 28/29] IB/rxe: Shared objects between user and kernel Moni Shoua
     [not found]     ` <1464886657-14258-29-git-send-email-monis-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-06-09  6:09       ` Leon Romanovsky
     [not found]         ` <20160609060931.GM3663-2ukJVAZIZ/Y@public.gmane.org>
2016-06-13  7:22           ` Moni Shoua
2016-06-02 16:57   ` [PATCH rdma-next 29/29] IB/rxe: Add Soft-RoCE to kbuild and makefiles Moni Shoua
2016-06-02 18:05   ` [PATCH rdma-next 00/29] Soft RoCE driver Steve Wise
2016-06-08 17:25     ` Hefty, Sean
2016-06-09 13:17     ` Moni Shoua
     [not found]       ` <CAG9sBKPa0qQ4uuhoWLfZaXP9E+DnLrkOcU-2yNXbdZFQKppN8w-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2016-06-09 14:45         ` Dennis Dalessandro
     [not found]           ` <20160609144554.GA14212-W4f6Xiosr+yv7QzWx2u06xL4W9x8LtSr@public.gmane.org>
2016-06-09 15:12             ` Moni Shoua
2016-06-09 19:16             ` Leon Romanovsky
     [not found]               ` <20160609191608.GA5408-2ukJVAZIZ/Y@public.gmane.org>
2016-06-09 19:47                 ` Dennis Dalessandro
     [not found]                   ` <20160609194758.GA9017-W4f6Xiosr+yv7QzWx2u06xL4W9x8LtSr@public.gmane.org>
2016-06-12  9:51                     ` Moni Shoua
     [not found]                       ` <CAG9sBKN3+Yf7dUG0ORNFNG=j+AgR1eq5mvPQxx6jZEXEGRpk3w-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2016-06-13 15:08                         ` Hefty, Sean
     [not found]                           ` <1828884A29C6694DAF28B7E6B8A82373AB0617C2-P5GAC/sN6hkd3b2yrw5b5LfspsVTdybXVpNB7YpNyf8@public.gmane.org>
2016-06-14 15:31                             ` Moni Shoua
2016-06-03  5:25   ` Christoph Hellwig
     [not found]     ` <20160603052539.GB31196-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2016-06-08  8:49       ` Moni Shoua
     [not found]         ` <CAG9sBKN9HQKhRgXWZ8JJwwUdEBaP9VnGwr790C=q4w0XDxkbtw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2016-06-08  9:52           ` Nicholas A. Bellinger
     [not found]             ` <1465379530.5365.91.camel-XoQW25Eq2zviZyQQd+hFbcojREIfoBdhmpATvIKMPHk@public.gmane.org>
2016-06-08 10:41               ` Leon Romanovsky
     [not found]                 ` <20160608104123.GJ3663-2ukJVAZIZ/Y@public.gmane.org>
2016-06-08 12:20                   ` Doug Ledford
     [not found]                     ` <5e1bd9ce-f802-9237-14a0-7cdc2be5089b-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2016-06-08 13:10                       ` Leon Romanovsky
2016-06-08 12:27           ` Sagi Grimberg
     [not found]             ` <57580F4D.10802-ImC7XgPzLAfvYQKSrp0J2Q@public.gmane.org>
2016-06-08 13:53               ` Steve Wise
2016-06-03 17:04   ` Christoph Lameter
     [not found]     ` <alpine.DEB.2.20.1606031203510.30612-wcBtFHqTun5QOdAKl3ChDw@public.gmane.org>
2016-06-05 13:30       ` Moni Shoua
2016-06-08 16:51   ` Steve Wise
2016-06-08 17:40     ` Steve Wise
2016-06-09  7:05       ` Moni Shoua

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='011101d1c721$5c22d590$146880b0$@opengridcomputing.com' \
    --to=swise-7bpotxp6k4+p2yhjcf5u+vpxobypeauw@public.gmane.org \
    --cc=amirv-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    --cc=dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
    --cc=kamalh-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    --cc=leon-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    --cc=linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
    --cc=liranl-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    --cc=majd-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    --cc=matanb-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    --cc=monis-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.