From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
stable@vger.kernel.org,
Dominique Martinet <dominique.martinet@cea.fr>,
Jun Piao <piaojun@huawei.com>,
Matthew Wilcox <willy@infradead.org>, Greg Kurz <groug@kaod.org>,
Sasha Levin <sashal@kernel.org>
Subject: [PATCH 4.19 10/72] 9p: add a per-client fcall kmem_cache
Date: Tue, 2 Jul 2019 10:01:11 +0200 [thread overview]
Message-ID: <20190702080125.107936681@linuxfoundation.org> (raw)
In-Reply-To: <20190702080124.564652899@linuxfoundation.org>
[ Upstream commit 91a76be37ff89795526c452a6799576b03bec501 ]
Having a specific cache for the fcall allocations helps speed up
end-to-end latency.
The caches will automatically be merged if there are multiple caches
of items with the same size so we do not need to try to share a cache
between different clients of the same size.
Since the msize is negotiated with the server, only allocate the cache
after that negotiation has happened - previous allocations or
allocations of different sizes (e.g. zero-copy fcall) are made with
kmalloc directly.
Some figures on two beefy VMs with Connect-IB (sriov) / trans=rdma,
with ior running 32 processes in parallel doing small 32 bytes IOs:
- no alloc (4.18-rc7 request cache): 65.4k req/s
- non-power of two alloc, no patch: 61.6k req/s
- power of two alloc, no patch: 62.2k req/s
- non-power of two alloc, with patch: 64.7k req/s
- power of two alloc, with patch: 65.1k req/s
Link: http://lkml.kernel.org/r/1532943263-24378-2-git-send-email-asmadeus@codewreck.org
Signed-off-by: Dominique Martinet <dominique.martinet@cea.fr>
Acked-by: Jun Piao <piaojun@huawei.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Greg Kurz <groug@kaod.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
include/net/9p/9p.h | 4 ++++
include/net/9p/client.h | 1 +
net/9p/client.c | 37 ++++++++++++++++++++++++++++++++-----
3 files changed, 37 insertions(+), 5 deletions(-)
diff --git a/include/net/9p/9p.h b/include/net/9p/9p.h
index b8eb51a661e5..4ab293f574e0 100644
--- a/include/net/9p/9p.h
+++ b/include/net/9p/9p.h
@@ -336,6 +336,9 @@ enum p9_qid_t {
#define P9_NOFID (u32)(~0)
#define P9_MAXWELEM 16
+/* Minimal header size: size[4] type[1] tag[2] */
+#define P9_HDRSZ 7
+
/* ample room for Twrite/Rread header */
#define P9_IOHDRSZ 24
@@ -558,6 +561,7 @@ struct p9_fcall {
size_t offset;
size_t capacity;
+ struct kmem_cache *cache;
u8 *sdata;
};
diff --git a/include/net/9p/client.h b/include/net/9p/client.h
index c2671d40bb6b..735f3979d559 100644
--- a/include/net/9p/client.h
+++ b/include/net/9p/client.h
@@ -123,6 +123,7 @@ struct p9_client {
struct p9_trans_module *trans_mod;
enum p9_trans_status status;
void *trans;
+ struct kmem_cache *fcall_cache;
union {
struct {
diff --git a/net/9p/client.c b/net/9p/client.c
index 83e39fef58e1..7ef54719c6f7 100644
--- a/net/9p/client.c
+++ b/net/9p/client.c
@@ -237,9 +237,16 @@ static int parse_opts(char *opts, struct p9_client *clnt)
return ret;
}
-static int p9_fcall_init(struct p9_fcall *fc, int alloc_msize)
+static int p9_fcall_init(struct p9_client *c, struct p9_fcall *fc,
+ int alloc_msize)
{
- fc->sdata = kmalloc(alloc_msize, GFP_NOFS);
+ if (likely(c->fcall_cache) && alloc_msize == c->msize) {
+ fc->sdata = kmem_cache_alloc(c->fcall_cache, GFP_NOFS);
+ fc->cache = c->fcall_cache;
+ } else {
+ fc->sdata = kmalloc(alloc_msize, GFP_NOFS);
+ fc->cache = NULL;
+ }
if (!fc->sdata)
return -ENOMEM;
fc->capacity = alloc_msize;
@@ -248,7 +255,16 @@ static int p9_fcall_init(struct p9_fcall *fc, int alloc_msize)
void p9_fcall_fini(struct p9_fcall *fc)
{
- kfree(fc->sdata);
+ /* sdata can be NULL for interrupted requests in trans_rdma,
+ * and kmem_cache_free does not do NULL-check for us
+ */
+ if (unlikely(!fc->sdata))
+ return;
+
+ if (fc->cache)
+ kmem_cache_free(fc->cache, fc->sdata);
+ else
+ kfree(fc->sdata);
}
EXPORT_SYMBOL(p9_fcall_fini);
@@ -273,9 +289,9 @@ p9_tag_alloc(struct p9_client *c, int8_t type, unsigned int max_size)
if (!req)
return NULL;
- if (p9_fcall_init(&req->tc, alloc_msize))
+ if (p9_fcall_init(c, &req->tc, alloc_msize))
goto free_req;
- if (p9_fcall_init(&req->rc, alloc_msize))
+ if (p9_fcall_init(c, &req->rc, alloc_msize))
goto free;
p9pdu_reset(&req->tc);
@@ -965,6 +981,7 @@ struct p9_client *p9_client_create(const char *dev_name, char *options)
clnt->trans_mod = NULL;
clnt->trans = NULL;
+ clnt->fcall_cache = NULL;
client_id = utsname()->nodename;
memcpy(clnt->name, client_id, strlen(client_id) + 1);
@@ -1008,6 +1025,15 @@ struct p9_client *p9_client_create(const char *dev_name, char *options)
if (err)
goto close_trans;
+ /* P9_HDRSZ + 4 is the smallest packet header we can have that is
+ * followed by data accessed from userspace by read
+ */
+ clnt->fcall_cache =
+ kmem_cache_create_usercopy("9p-fcall-cache", clnt->msize,
+ 0, 0, P9_HDRSZ + 4,
+ clnt->msize - (P9_HDRSZ + 4),
+ NULL);
+
return clnt;
close_trans:
@@ -1039,6 +1065,7 @@ void p9_client_destroy(struct p9_client *clnt)
p9_tag_cleanup(clnt);
+ kmem_cache_destroy(clnt->fcall_cache);
kfree(clnt);
}
EXPORT_SYMBOL(p9_client_destroy);
--
2.20.1
next prev parent reply other threads:[~2019-07-02 8:15 UTC|newest]
Thread overview: 84+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-07-02 8:01 [PATCH 4.19 00/72] 4.19.57-stable review Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 01/72] perf ui helpline: Use strlcpy() as a shorter form of strncpy() + explicit set nul Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 02/72] perf help: Remove needless use of strncpy() Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 03/72] perf header: Fix unchecked usage " Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 04/72] arm64: Dont unconditionally add -Wno-psabi to KBUILD_CFLAGS Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 05/72] Revert "x86/uaccess, ftrace: Fix ftrace_likely_update() vs. SMAP" Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 06/72] IB/hfi1: Close PSM sdma_progress sleep window Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 07/72] 9p/xen: fix check for xenbus_read error in front_probe Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 08/72] 9p: Use a slab for allocating requests Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 09/72] 9p: embed fcall in req to round down buffer allocs Greg Kroah-Hartman
2019-07-02 8:01 ` Greg Kroah-Hartman [this message]
2019-07-02 8:01 ` [PATCH 4.19 11/72] 9p: rename p9_free_req() function Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 12/72] 9p: Add refcount to p9_req_t Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 13/72] 9p/rdma: do not disconnect on down_interruptible EAGAIN Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 14/72] 9p: Rename req to rreq in trans_fd Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 15/72] 9p: acl: fix uninitialized iattr access Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 16/72] 9p/rdma: remove useless check in cm_event_handler Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 17/72] 9p: p9dirent_read: check network-provided name length Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 18/72] 9p: potential NULL dereference Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 19/72] 9p/trans_fd: abort p9_read_work if req status changed Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 20/72] 9p/trans_fd: put worker reqs on destroy Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 21/72] net/9p: include trans_common.h to fix missing prototype warning Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 22/72] qmi_wwan: Fix out-of-bounds read Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 23/72] Revert "usb: dwc3: gadget: Clear req->needs_extra_trb flag on cleanup" Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 24/72] usb: dwc3: gadget: combine unaligned and zero flags Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 25/72] usb: dwc3: gadget: track number of TRBs per request Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 26/72] usb: dwc3: gadget: use num_trbs when skipping TRBs on ->dequeue() Greg Kroah-Hartman
2019-07-03 2:03 ` Sasha Levin
2019-07-03 7:20 ` Greg Kroah-Hartman
2019-07-03 19:59 ` Sasha Levin
2019-07-02 8:01 ` [PATCH 4.19 27/72] usb: dwc3: gadget: extract dwc3_gadget_ep_skip_trbs() Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 28/72] usb: dwc3: gadget: introduce cancelled_list Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 29/72] usb: dwc3: gadget: move requests to cancelled_list Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 30/72] usb: dwc3: gadget: remove wait_end_transfer Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 31/72] usb: dwc3: gadget: Clear req->needs_extra_trb flag on cleanup Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 32/72] fs/proc/array.c: allow reporting eip/esp for all coredumping threads Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 33/72] mm/mempolicy.c: fix an incorrect rebind node in mpol_rebind_nodemask Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 34/72] fs/binfmt_flat.c: make load_flat_shared_library() work Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 35/72] clk: socfpga: stratix10: fix divider entry for the emac clocks Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 36/72] mm: soft-offline: return -EBUSY if set_hwpoison_free_buddy_page() fails Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 37/72] mm: hugetlb: soft-offline: dissolve_free_huge_page() return zero on !PageHuge Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 38/72] mm/page_idle.c: fix oops because end_pfn is larger than max_pfn Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 39/72] dm log writes: make sure super sector log updates are written in order Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 40/72] scsi: vmw_pscsi: Fix use-after-free in pvscsi_queue_lck() Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 41/72] x86/speculation: Allow guests to use SSBD even if host does not Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 42/72] x86/microcode: Fix the microcode load on CPU hotplug for real Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 43/72] x86/resctrl: Prevent possible overrun during bitmap operations Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 44/72] KVM: x86/mmu: Allocate PAE root array when using SVMs 32-bit NPT Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 45/72] NFS/flexfiles: Use the correct TCP timeout for flexfiles I/O Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 46/72] cpu/speculation: Warn on unsupported mitigations= parameter Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 47/72] SUNRPC: Clean up initialisation of the struct rpc_rqst Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 48/72] irqchip/mips-gic: Use the correct local interrupt map registers Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 49/72] eeprom: at24: fix unexpected timeout under high load Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 50/72] af_packet: Block execution of tasks waiting for transmit to complete in AF_PACKET Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 51/72] bonding: Always enable vlan tx offload Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 52/72] ipv4: Use return value of inet_iif() for __raw_v4_lookup in the while loop Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 53/72] net/packet: fix memory leak in packet_set_ring() Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 54/72] net: remove duplicate fetch in sock_getsockopt Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 55/72] net: stmmac: fixed new system time seconds value calculation Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 56/72] net: stmmac: set IC bit when transmitting frames with HW timestamp Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 57/72] sctp: change to hold sk after auth shkey is created successfully Greg Kroah-Hartman
2019-07-02 8:01 ` [PATCH 4.19 58/72] team: Always enable vlan tx offload Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 59/72] tipc: change to use register_pernet_device Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 60/72] tipc: check msg->req data len in tipc_nl_compat_bearer_disable Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 61/72] tun: wake up waitqueues after IFF_UP is set Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 62/72] bpf: simplify definition of BPF_FIB_LOOKUP related flags Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 63/72] bpf: lpm_trie: check left child of last leftmost node for NULL Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 64/72] bpf: fix nested bpf tracepoints with per-cpu data Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 65/72] bpf: fix unconnected udp hooks Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 66/72] bpf: udp: Avoid calling reuseports bpf_prog from udp_gro Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 67/72] bpf: udp: ipv6: Avoid running reuseports bpf_prog from __udp6_lib_err Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 68/72] arm64: futex: Avoid copying out uninitialised stack in failed cmpxchg() Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 69/72] bpf, arm64: use more scalable stadd over ldxr / stxr loop in xadd Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 70/72] futex: Update comments and docs about return values of arch futex code Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 71/72] RDMA: Directly cast the sockaddr union to sockaddr Greg Kroah-Hartman
2019-07-02 8:02 ` [PATCH 4.19 72/72] tipc: pass tunnel dev as NULL to udp_tunnel(6)_xmit_skb Greg Kroah-Hartman
2019-07-02 12:32 ` [PATCH 4.19 00/72] 4.19.57-stable review kernelci.org bot
2019-07-02 16:54 ` Naresh Kamboju
2019-07-02 20:23 ` Guenter Roeck
2019-07-03 14:46 ` Greg Kroah-Hartman
2019-07-02 21:08 ` Kelsey Skunberg
2019-07-02 22:52 ` shuah
2019-07-03 10:21 ` Jon Hunter
2019-07-04 5:29 ` Bharath Vedartham
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=20190702080125.107936681@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=dominique.martinet@cea.fr \
--cc=groug@kaod.org \
--cc=linux-kernel@vger.kernel.org \
--cc=piaojun@huawei.com \
--cc=sashal@kernel.org \
--cc=stable@vger.kernel.org \
--cc=willy@infradead.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).