From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: stable@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
patches@lists.linux.dev, Eric Dumazet <edumazet@google.com>,
Jakub Kicinski <kuba@kernel.org>,
Suraj Jitindar Singh <surajjs@amazon.com>
Subject: [PATCH 5.4 36/38] net/dst: use a smaller percpu_counter batch for dst entries accounting
Date: Sat, 13 Jan 2024 10:50:12 +0100 [thread overview]
Message-ID: <20240113094207.556743428@linuxfoundation.org> (raw)
In-Reply-To: <20240113094206.455533180@linuxfoundation.org>
5.4-stable review patch. If anyone has any objections, please let me know.
------------------
From: Eric Dumazet <edumazet@google.com>
commit cf86a086a18095e33e0637cb78cda1fcf5280852 upstream.
percpu_counter_add() uses a default batch size which is quite big
on platforms with 256 cpus. (2*256 -> 512)
This means dst_entries_get_fast() can be off by +/- 2*(nr_cpus^2)
(131072 on servers with 256 cpus)
Reduce the batch size to something more reasonable, and
add logic to ip6_dst_gc() to call dst_entries_get_slow()
before calling the _very_ expensive fib6_run_gc() function.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Suraj Jitindar Singh <surajjs@amazon.com>
Cc: <stable@vger.kernel.org> # 5.4.x
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
include/net/dst_ops.h | 4 +++-
net/core/dst.c | 8 ++++----
net/ipv6/route.c | 3 +++
3 files changed, 10 insertions(+), 5 deletions(-)
--- a/include/net/dst_ops.h
+++ b/include/net/dst_ops.h
@@ -53,9 +53,11 @@ static inline int dst_entries_get_slow(s
return percpu_counter_sum_positive(&dst->pcpuc_entries);
}
+#define DST_PERCPU_COUNTER_BATCH 32
static inline void dst_entries_add(struct dst_ops *dst, int val)
{
- percpu_counter_add(&dst->pcpuc_entries, val);
+ percpu_counter_add_batch(&dst->pcpuc_entries, val,
+ DST_PERCPU_COUNTER_BATCH);
}
static inline int dst_entries_init(struct dst_ops *dst)
--- a/net/core/dst.c
+++ b/net/core/dst.c
@@ -81,11 +81,11 @@ void *dst_alloc(struct dst_ops *ops, str
{
struct dst_entry *dst;
- if (ops->gc && dst_entries_get_fast(ops) > ops->gc_thresh) {
+ if (ops->gc &&
+ !(flags & DST_NOCOUNT) &&
+ dst_entries_get_fast(ops) > ops->gc_thresh) {
if (ops->gc(ops)) {
- printk_ratelimited(KERN_NOTICE "Route cache is full: "
- "consider increasing sysctl "
- "net.ipv[4|6].route.max_size.\n");
+ pr_notice_ratelimited("Route cache is full: consider increasing sysctl net.ipv6.route.max_size.\n");
return NULL;
}
}
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -3218,6 +3218,9 @@ static int ip6_dst_gc(struct dst_ops *op
int entries;
entries = dst_entries_get_fast(ops);
+ if (entries > rt_max_size)
+ entries = dst_entries_get_slow(ops);
+
if (time_after(rt_last_gc + rt_min_interval, jiffies) &&
entries <= rt_max_size)
goto out;
next prev parent reply other threads:[~2024-01-13 9:57 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-13 9:49 [PATCH 5.4 00/38] 5.4.267-rc1 review Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 01/38] nfc: llcp_core: Hold a ref to llcp_local->dev when holding a ref to llcp_local Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 02/38] i40e: Fix filter input checks to prevent config with invalid values Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 03/38] net: sched: em_text: fix possible memory leak in em_text_destroy() Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 04/38] net: Implement missing getsockopt(SO_TIMESTAMPING_NEW) Greg Kroah-Hartman
2024-01-13 19:25 ` Jörn-Thorben Hinz
2024-01-13 9:49 ` [PATCH 5.4 05/38] can: raw: add support for SO_TXTIME/SCM_TXTIME Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 06/38] can: raw: add support for SO_MARK Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 07/38] net-timestamp: extend SOF_TIMESTAMPING_OPT_ID to HW timestamps Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 08/38] ARM: sun9i: smp: Fix array-index-out-of-bounds read in sunxi_mc_smp_init Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 09/38] net: bcmgenet: Fix FCS generation for fragmented skbuffs Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 10/38] net: Save and restore msg_namelen in sock_sendmsg Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 11/38] i40e: fix use-after-free in i40e_aqc_add_filters() Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 12/38] ASoC: meson: g12a: extract codec-to-codec utils Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 13/38] ASoC: meson: g12a-tohdmitx: Validate written enum values Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 14/38] ASoC: meson: g12a-tohdmitx: Fix event generation for S/PDIF mux Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 15/38] i40e: Restore VF MSI-X state during PCI reset Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 16/38] net/qla3xxx: switch from pci_ to dma_ API Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 17/38] net/qla3xxx: fix potential memleak in ql_alloc_buffer_queues Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 18/38] asix: Add check for usbnet_get_endpoints Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 19/38] bnxt_en: Remove mis-applied code from bnxt_cfg_ntp_filters() Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 20/38] net: Implement missing SO_TIMESTAMPING_NEW cmsg support Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 21/38] mm/memory-failure: check the mapcount of the precise page Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 22/38] firewire: ohci: suppress unexpected system reboot in AMD Ryzen machines and ASM108x/VT630x PCIe cards Greg Kroah-Hartman
2024-01-13 9:49 ` [PATCH 5.4 23/38] i2c: core: Fix atomic xfer check for non-preempt config Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 24/38] mm: fix unmap_mapping_range high bits shift bug Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 25/38] mmc: rpmb: fixes pause retune on all RPMB partitions Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 26/38] mmc: core: Cancel delayed work before releasing host Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 27/38] mmc: sdhci-sprd: Fix eMMC init failure after hw reset Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 28/38] ath10k: Wait until copy complete is actually done before completing Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 29/38] ath10k: Add interrupt summary based CE processing Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 30/38] ath10k: Keep track of which interrupts fired, dont poll them Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 31/38] ath10k: Get rid of "per_ce_irq" hw param Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 32/38] net: tls, update curr on splice as well Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 33/38] netfilter: nf_tables: Reject tables of unsupported family Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 34/38] PCI: Extract ATS disabling to a helper function Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 35/38] PCI: Disable ATS for specific Intel IPU E2000 devices Greg Kroah-Hartman
2024-01-13 9:50 ` Greg Kroah-Hartman [this message]
2024-01-13 9:50 ` [PATCH 5.4 37/38] ipv6: make ip6_rt_gc_expire an atomic_t Greg Kroah-Hartman
2024-01-13 9:50 ` [PATCH 5.4 38/38] ipv6: remove max_size check inline with ipv4 Greg Kroah-Hartman
2024-01-15 9:24 ` [PATCH 5.4 00/38] 5.4.267-rc1 review Naresh Kamboju
2024-01-15 10:23 ` Jon Hunter
2024-01-15 16:25 ` Harshit Mogalapalli
2024-01-15 19:09 ` Florian Fainelli
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=20240113094207.556743428@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=edumazet@google.com \
--cc=kuba@kernel.org \
--cc=patches@lists.linux.dev \
--cc=stable@vger.kernel.org \
--cc=surajjs@amazon.com \
/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