LinuxLists.cc - [PATCH net-next v6 0/8] sysctl: Remove sentinel elements from networking

2024-05-01 09:30:23

Subject: [PATCH net-next v6 0/8] sysctl: Remove sentinel elements from networking

From: Joel Granados <[email protected]>

What?
These commits remove the sentinel element (last empty element) from the
sysctl arrays of all the files under the "net/" directory that register
a sysctl array. The merging of the preparation patches [4] to mainline
allows us to just remove sentinel elements without changing behavior.
This is safe because the sysctl registration code (register_sysctl() and
friends) use the array size in addition to checking for a sentinel [1].

Why?
By removing the sysctl sentinel elements we avoid kernel bloat as
ctl_table arrays get moved out of kernel/sysctl.c into their own
respective subsystems. This move was started long ago to avoid merge
conflicts; the sentinel removal bit came after Mathew Wilcox suggested
it to avoid bloating the kernel by one element as arrays moved out. This
patchset will reduce the overall build time size of the kernel and run
time memory bloat by about ~64 bytes per declared ctl_table array (more
info here [5]).

When are we done?
There are 4 patchest (25 commits [2]) that are still outstanding to
completely remove the sentinels: files under "net/" (this patchset),
files under "kernel/" dir, misc dirs (files under mm/ security/ and
others) and the final set that removes the unneeded check for ->procname
== NULL.

Testing:
* Ran sysctl selftests (./tools/testing/selftests/sysctl/sysctl.sh)
* Ran this through 0-day with no errors or warnings

Savings in vmlinux:
A total of 64 bytes per sentinel is saved after removal; I measured in
x86_64 to give an idea of the aggregated savings. The actual savings
will depend on individual kernel configuration.
* bloat-o-meter
- The "yesall" config saves 3976 bytes (bloat-o-meter output [6])
- A reduced config [3] saves 1263 bytes (bloat-o-meter output [7])

Savings in allocated memory:
None in this set but will occur when the superfluous allocations are
removed from proc_sysctl.c. I include it here for context. The
estimated savings during boot for config [3] are 6272 bytes. See [8]
for how to measure it.

Comments/feedback greatly appreciated

Changes in v6:
- Rebased onto net-next/main.
- Besides re-running my cocci scripts, I ran a new find script [9].
Found 0 hits in net/
- Moved "i" variable declaraction out of for() in sysctl_core_net_init
- Removed forgotten sentinel in mpls_table
- Removed CONFIG_AX25_DAMA_SLAVE guard from net/ax25/ax25_ds_timer.c. It
is not needed because that file is compiled only when
CONFIG_AX25_DAMA_SLAVE is set.
- When traversing smc_table, stop on ARRAY_SIZE instead of ARRAY_SIZE-1.
- Link to v5: https://lore.kernel.org/r/[email protected]

Changes in v5:
- Added net files with additional variable to my test .config so the
typo can be caught next time.
- Fixed typo tabel_size -> table_size
- Link to v4: https://lore.kernel.org/r/[email protected]

Changes in v4:
- Keep reverse xmas tree order when introducing new variables
- Use a table_size variable to keep the value of ARRAY_SIZE
- Separated the original "networking: Remove the now superfluous
sentinel elements from ctl_table arra" into smaller commits to ease
review
- Merged x.25 and ax.25 commits together.
- Removed any SOB from the commits that were changed
- Link to v3: https://lore.kernel.org/r/[email protected]

Changes in v3:
- Reworkded ax.25
- Added a BUILD_BUG_ON for the ax.25 commit
- Added a CONFIG_AX25_DAMA_SLAVE guard where needed
- Link to v2: https://lore.kernel.org/r/[email protected]

Changes in v2:
- Rebased to v6.9-rc1
- Removed unneeded comment from sysctl_net_ax25.c
- Link to v1: https://lore.kernel.org/r/[email protected]

Best
Joel

[1] https://lore.kernel.org/all/[email protected]/
[2] https://git.kernel.org/pub/scm/linux/kernel/git/joel.granados/linux.git/tag/?h=sysctl_remove_empty_elem_v5
[3] https://gist.github.com/Joelgranados/feaca7af5537156ca9b73aeaec093171
[4] https://lore.kernel.org/all/ZO5Yx5JFogGi%[email protected]/

[5]
Links Related to the ctl_table sentinel removal:
* Good summaries from Luis:
https://lore.kernel.org/all/ZO5Yx5JFogGi%[email protected]/
https://lore.kernel.org/all/[email protected]/
* Patches adjusting sysctl register calls:
https://lore.kernel.org/all/[email protected]/
https://lore.kernel.org/all/[email protected]/
* Discussions about expectations and approach
https://lore.kernel.org/all/[email protected]
https://lore.kernel.org/all/[email protected]

[6]
add/remove: 0/1 grow/shrink: 2/67 up/down: 76/-4052 (-3976)
Function old new delta
llc_sysctl_init 306 377 +71
nf_log_net_init 866 871 +5
sysctl_core_net_init 375 366 -9
lowpan_frags_init_net 618 598 -20
ip_vs_control_net_init_sysctl 2446 2422 -24
sysctl_route_net_init 521 493 -28
__addrconf_sysctl_register 678 650 -28
xfrm_sysctl_init 405 374 -31
mpls_net_init 367 334 -33
sctp_sysctl_net_register 386 346 -40
__ip_vs_lblcr_init 546 501 -45
__ip_vs_lblc_init 546 501 -45
neigh_sysctl_register 1011 958 -53
mpls_dev_sysctl_register 475 419 -56
ipv6_route_sysctl_init 450 394 -56
xs_tunables_table 448 384 -64
xr_tunables_table 448 384 -64
xfrm_table 320 256 -64
xfrm6_policy_table 128 64 -64
xfrm4_policy_table 128 64 -64
x25_table 448 384 -64
vs_vars 1984 1920 -64
unix_table 128 64 -64
tipc_table 448 384 -64
svcrdma_parm_table 832 768 -64
smc_table 512 448 -64
sctp_table 256 192 -64
sctp_net_table 2304 2240 -64
rxrpc_sysctl_table 704 640 -64
rose_table 704 640 -64
rds_tcp_sysctl_table 192 128 -64
rds_sysctl_rds_table 384 320 -64
rds_ib_sysctl_table 384 320 -64
phonet_table 128 64 -64
nr_table 832 768 -64
nf_log_sysctl_table 768 704 -64
nf_log_sysctl_ftable 128 64 -64
nf_ct_sysctl_table 3200 3136 -64
nf_ct_netfilter_table 128 64 -64
nf_ct_frag6_sysctl_table 256 192 -64
netns_core_table 320 256 -64
net_core_table 2176 2112 -64
neigh_sysctl_template 1416 1352 -64
mptcp_sysctl_table 576 512 -64
mpls_dev_table 128 64 -64
lowpan_frags_ns_ctl_table 256 192 -64
lowpan_frags_ctl_table 128 64 -64
llc_station_table 64 - -64
llc2_timeout_table 320 256 -64
ipv6_table_template 1344 1280 -64
ipv6_route_table_template 768 704 -64
ipv6_rotable 320 256 -64
ipv6_icmp_table_template 448 384 -64
ipv4_table 1024 960 -64
ipv4_route_table 832 768 -64
ipv4_route_netns_table 320 256 -64
ipv4_net_table 7552 7488 -64
ip6_frags_ns_ctl_table 256 192 -64
ip6_frags_ctl_table 128 64 -64
ip4_frags_ns_ctl_table 320 256 -64
ip4_frags_ctl_table 128 64 -64
devinet_sysctl 2184 2120 -64
debug_table 384 320 -64
dccp_default_table 576 512 -64
ctl_forward_entry 128 64 -64
brnf_table 448 384 -64
ax25_param_table 960 896 -64
atalk_table 320 256 -64
addrconf_sysctl 3904 3840 -64
vs_vars_table 256 128 -128
Total: Before=440631035, After=440627059, chg -0.00%

[7]
add/remove: 0/0 grow/shrink: 1/22 up/down: 8/-1263 (-1255)
Function old new delta
sysctl_route_net_init 189 197 +8
__addrconf_sysctl_register 306 294 -12
ipv6_route_sysctl_init 201 185 -16
neigh_sysctl_register 385 366 -19
unix_table 128 64 -64
netns_core_table 256 192 -64
net_core_table 1664 1600 -64
neigh_sysctl_template 1416 1352 -64
ipv6_table_template 1344 1280 -64
ipv6_route_table_template 768 704 -64
ipv6_rotable 192 128 -64
ipv6_icmp_table_template 448 384 -64
ipv4_table 768 704 -64
ipv4_route_table 832 768 -64
ipv4_route_netns_table 320 256 -64
ipv4_net_table 7040 6976 -64
ip6_frags_ns_ctl_table 256 192 -64
ip6_frags_ctl_table 128 64 -64
ip4_frags_ns_ctl_table 320 256 -64
ip4_frags_ctl_table 128 64 -64
devinet_sysctl 2184 2120 -64
ctl_forward_entry 128 64 -64
addrconf_sysctl 3392 3328 -64
Total: Before=8523801, After=8522546, chg -0.01%

[8]
To measure the in memory savings apply this on top of this patchset.

"
diff --git i/fs/proc/proc_sysctl.c w/fs/proc/proc_sysctl.c
index 37cde0efee57..896c498600e8 100644
--- i/fs/proc/proc_sysctl.c
+++ w/fs/proc/proc_sysctl.c
@@ -966,6 +966,7 @@ static struct ctl_dir *new_dir(struct ctl_table_set *set,
table[0].procname = new_name;
table[0].mode = S_IFDIR|S_IRUGO|S_IXUGO;
init_header(&new->header, set->dir.header.root, set, node, table, 1);
+ printk("%ld sysctl saved mem kzalloc\n", sizeof(struct ctl_table));

return new;
}
@@ -1189,6 +1190,7 @@ static struct ctl_table_header *new_links(struct ctl_dir *dir, s>
link_name += len;
link++;
}
+ printk("%ld sysctl saved mem kzalloc\n", sizeof(struct ctl_table));
init_header(links, dir->header.root, dir->header.set, node, link_table,
head->ctl_table_size);
links->nreg = nr_entries;
"
and then run the following bash script in the kernel:

```
accum=0
for n in $(dmesg | grep kzalloc | awk '{print $3}') ; do
accum=$(calc "$accum + $n")
done
echo $accum
```

[9]
```
#!/usr/bin/gawk -f

BEGINFILE {
RS=","
has_struct = 0
}

/(static )?(const )?struct ctl_table/ {
has_struct = 1
}

has_struct && /^(\n)?[\t ]*{(\n)*[\t ]*}/ {
print "Filename : " FILENAME ", Record Number : " FNR
}
```

Signed-off-by: Joel Granados <[email protected]>

--

---
---
Joel Granados (8):
net: Remove the now superfluous sentinel elements from ctl_table array
net: ipv{6,4}: Remove the now superfluous sentinel elements from ctl_table array
net: rds: Remove the now superfluous sentinel elements from ctl_table array
net: sunrpc: Remove the now superfluous sentinel elements from ctl_table array
net: Remove ctl_table sentinel elements from several networking subsystems
netfilter: Remove the now superfluous sentinel elements from ctl_table array
appletalk: Remove the now superfluous sentinel elements from ctl_table array
ax.25: x.25: Remove the now superfluous sentinel elements from ctl_table array

include/net/ax25.h | 2 ++
net/appletalk/sysctl_net_atalk.c | 1 -
net/ax25/ax25_dev.c | 3 +++
net/ax25/ax25_ds_timer.c | 1 +
net/ax25/sysctl_net_ax25.c | 3 +--
net/bridge/br_netfilter_hooks.c | 1 -
net/core/neighbour.c | 5 +----
net/core/sysctl_net_core.c | 13 ++++++-------
net/dccp/sysctl.c | 2 --
net/ieee802154/6lowpan/reassembly.c | 6 +-----
net/ipv4/devinet.c | 5 ++---
net/ipv4/ip_fragment.c | 2 --
net/ipv4/route.c | 8 ++------
net/ipv4/sysctl_net_ipv4.c | 7 +++----
net/ipv4/xfrm4_policy.c | 1 -
net/ipv6/addrconf.c | 8 +++-----
net/ipv6/icmp.c | 1 -
net/ipv6/netfilter/nf_conntrack_reasm.c | 1 -
net/ipv6/reassembly.c | 2 --
net/ipv6/route.c | 5 -----
net/ipv6/sysctl_net_ipv6.c | 8 +++-----
net/ipv6/xfrm6_policy.c | 1 -
net/llc/sysctl_net_llc.c | 8 ++------
net/mpls/af_mpls.c | 13 ++++++-------
net/mptcp/ctrl.c | 1 -
net/netfilter/ipvs/ip_vs_ctl.c | 5 +----
net/netfilter/ipvs/ip_vs_lblc.c | 5 +----
net/netfilter/ipvs/ip_vs_lblcr.c | 5 +----
net/netfilter/nf_conntrack_standalone.c | 6 +-----
net/netfilter/nf_log.c | 3 +--
net/netrom/sysctl_net_netrom.c | 1 -
net/phonet/sysctl.c | 1 -
net/rds/ib_sysctl.c | 1 -
net/rds/sysctl.c | 1 -
net/rds/tcp.c | 1 -
net/rose/sysctl_net_rose.c | 1 -
net/rxrpc/sysctl.c | 1 -
net/sctp/sysctl.c | 10 +++-------
net/smc/smc_sysctl.c | 6 +++---
net/sunrpc/sysctl.c | 1 -
net/sunrpc/xprtrdma/svc_rdma.c | 1 -
net/sunrpc/xprtrdma/transport.c | 1 -
net/sunrpc/xprtsock.c | 1 -
net/tipc/sysctl.c | 1 -
net/unix/sysctl_net_unix.c | 1 -
net/x25/sysctl_net_x25.c | 1 -
net/xfrm/xfrm_sysctl.c | 5 +----
47 files changed, 48 insertions(+), 119 deletions(-)
---
base-commit: c2e6a872bde9912f1a7579639c5ca3adf1003916
change-id: 20240311-jag-sysctl_remset_net-d403a1a93d6b

Best regards,
--
Joel Granados <[email protected]>

2024-05-01 09:30:40

by Joel Granados via B4 Relay

[permalink] [raw]

Subject: [PATCH net-next v6 6/8] netfilter: Remove the now superfluous sentinel elements from ctl_table array

2024-05-01 09:30:49

by Joel Granados via B4 Relay

[permalink] [raw]

Subject: [PATCH net-next v6 2/8] net: ipv{6,4}: Remove the now superfluous sentinel elements from ctl_table array

From: Joel Granados <[email protected]>

This commit comes at the tail end of a greater effort to remove the
empty elements at the end of the ctl_table arrays (sentinels) which
will reduce the overall build time size of the kernel and run time
memory bloat by ~64 bytes per sentinel (further information Link :
https://lore.kernel.org/all/ZO5Yx5JFogGi%[email protected]/)

* Remove sentinel element from ctl_table structs.
* Remove the zeroing out of an array element (to make it look like a
sentinel) in sysctl_route_net_init And ipv6_route_sysctl_init.
This is not longer needed and is safe after commit c899710fe7f9
("networking: Update to register_net_sysctl_sz") added the array size
to the ctl_table registration.
* Remove extra sentinel element in the declaration of devinet_vars.
* Removed the "-1" in __devinet_sysctl_register, sysctl_route_net_init,
ipv6_sysctl_net_init and ipv4_sysctl_init_net that adjusted for having
an extra empty element when looping over ctl_table arrays
* Replace the for loop stop condition in __addrconf_sysctl_register that
tests for procname == NULL with one that depends on array size
* Removing the unprivileged user check in ipv6_route_sysctl_init is
safe as it is replaced by calling ipv6_route_sysctl_table_size;
introduced in commit c899710fe7f9 ("networking: Update to
register_net_sysctl_sz")
* Use a table_size variable to keep the value of ARRAY_SIZE

Signed-off-by: Joel Granados <[email protected]>
---
net/ipv4/devinet.c | 5 ++---
net/ipv4/ip_fragment.c | 2 --
net/ipv4/route.c | 8 ++------
net/ipv4/sysctl_net_ipv4.c | 7 +++----
net/ipv4/xfrm4_policy.c | 1 -
net/ipv6/addrconf.c | 8 +++-----
net/ipv6/icmp.c | 1 -
net/ipv6/reassembly.c | 2 --
net/ipv6/route.c | 5 -----
net/ipv6/sysctl_net_ipv6.c | 8 +++-----
net/ipv6/xfrm6_policy.c | 1 -
11 files changed, 13 insertions(+), 35 deletions(-)

diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c
index 364dbf0cd9bf..a612c57b61c5 100644
--- a/net/ipv4/devinet.c
+++ b/net/ipv4/devinet.c
@@ -2520,7 +2520,7 @@ static int ipv4_doint_and_flush(struct ctl_table *ctl, int write,

static struct devinet_sysctl_table {
struct ctl_table_header *sysctl_header;
- struct ctl_table devinet_vars[__IPV4_DEVCONF_MAX];
+ struct ctl_table devinet_vars[IPV4_DEVCONF_MAX];
} devinet_sysctl = {
.devinet_vars = {
DEVINET_SYSCTL_COMPLEX_ENTRY(FORWARDING, "forwarding",
@@ -2583,7 +2583,7 @@ static int __devinet_sysctl_register(struct net *net, char *dev_name,
if (!t)
goto out;

- for (i = 0; i < ARRAY_SIZE(t->devinet_vars) - 1; i++) {
+ for (i = 0; i < ARRAY_SIZE(t->devinet_vars); i++) {
t->devinet_vars[i].data += (char *)p - (char *)&ipv4_devconf;
t->devinet_vars[i].extra1 = p;
t->devinet_vars[i].extra2 = net;
@@ -2657,7 +2657,6 @@ static struct ctl_table ctl_forward_entry[] = {
.extra1 = &ipv4_devconf,
.extra2 = &init_net,
},
- { },
};
#endif

diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c
index 534b98a0744a..08e2c92e25ab 100644
--- a/net/ipv4/ip_fragment.c
+++ b/net/ipv4/ip_fragment.c
@@ -580,7 +580,6 @@ static struct ctl_table ip4_frags_ns_ctl_table[] = {
.proc_handler = proc_dointvec_minmax,
.extra1 = &dist_min,
},
- { }
};

/* secret interval has been deprecated */
@@ -593,7 +592,6 @@ static struct ctl_table ip4_frags_ctl_table[] = {
.mode = 0644,
.proc_handler = proc_dointvec_jiffies,
},
- { }
};

static int __net_init ip4_frags_ns_ctl_register(struct net *net)
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 0fd9a3d7ac4a..5fd54103174f 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -3496,7 +3496,6 @@ static struct ctl_table ipv4_route_table[] = {
.mode = 0644,
.proc_handler = proc_dointvec,
},
- { }
};

static const char ipv4_route_flush_procname[] = "flush";
@@ -3530,7 +3529,6 @@ static struct ctl_table ipv4_route_netns_table[] = {
.mode = 0644,
.proc_handler = proc_dointvec,
},
- { },
};

static __net_init int sysctl_route_net_init(struct net *net)
@@ -3548,16 +3546,14 @@ static __net_init int sysctl_route_net_init(struct net *net)

/* Don't export non-whitelisted sysctls to unprivileged users */
if (net->user_ns != &init_user_ns) {
- if (tbl[0].procname != ipv4_route_flush_procname) {
- tbl[0].procname = NULL;
+ if (tbl[0].procname != ipv4_route_flush_procname)
table_size = 0;
- }
}

/* Update the variables to point into the current struct net
* except for the first element flush
*/
- for (i = 1; i < ARRAY_SIZE(ipv4_route_netns_table) - 1; i++)
+ for (i = 1; i < table_size; i++)
tbl[i].data += (void *)net - (void *)&init_net;
}
tbl[0].extra1 = net;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index ce5d19978a26..162a0a3b6ba5 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -575,7 +575,6 @@ static struct ctl_table ipv4_table[] = {
.extra1 = &sysctl_fib_sync_mem_min,
.extra2 = &sysctl_fib_sync_mem_max,
},
- { }
};

static struct ctl_table ipv4_net_table[] = {
@@ -1502,11 +1501,11 @@ static struct ctl_table ipv4_net_table[] = {
.proc_handler = proc_dou8vec_minmax,
.extra1 = SYSCTL_ONE,
},
- { }
};

static __net_init int ipv4_sysctl_init_net(struct net *net)
{
+ size_t table_size = ARRAY_SIZE(ipv4_net_table);
struct ctl_table *table;

table = ipv4_net_table;
@@ -1517,7 +1516,7 @@ static __net_init int ipv4_sysctl_init_net(struct net *net)
if (!table)
goto err_alloc;

- for (i = 0; i < ARRAY_SIZE(ipv4_net_table) - 1; i++) {
+ for (i = 0; i < table_size; i++) {
if (table[i].data) {
/* Update the variables to point into
* the current struct net
@@ -1533,7 +1532,7 @@ static __net_init int ipv4_sysctl_init_net(struct net *net)
}

net->ipv4.ipv4_hdr = register_net_sysctl_sz(net, "net/ipv4", table,
- ARRAY_SIZE(ipv4_net_table));
+ table_size);
if (!net->ipv4.ipv4_hdr)
goto err_reg;

diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index fccbbd3e1a4b..0294fef577fa 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -152,7 +152,6 @@ static struct ctl_table xfrm4_policy_table[] = {
.mode = 0644,
.proc_handler = proc_dointvec,
},
- { }
};

static __net_init int xfrm4_net_sysctl_init(struct net *net)
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index 9aa0900abfa1..5c424a0e7232 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -7184,14 +7184,12 @@ static const struct ctl_table addrconf_sysctl[] = {
.extra1 = SYSCTL_ZERO,
.extra2 = SYSCTL_TWO,
},
- {
- /* sentinel */
- }
};

static int __addrconf_sysctl_register(struct net *net, char *dev_name,
struct inet6_dev *idev, struct ipv6_devconf *p)
{
+ size_t table_size = ARRAY_SIZE(addrconf_sysctl);
int i, ifindex;
struct ctl_table *table;
char path[sizeof("net/ipv6/conf/") + IFNAMSIZ];
@@ -7200,7 +7198,7 @@ static int __addrconf_sysctl_register(struct net *net, char *dev_name,
if (!table)
goto out;

- for (i = 0; table[i].data; i++) {
+ for (i = 0; i < table_size; i++) {
table[i].data += (char *)p - (char *)&ipv6_devconf;
/* If one of these is already set, then it is not safe to
* overwrite either of them: this makes proc_dointvec_minmax
@@ -7215,7 +7213,7 @@ static int __addrconf_sysctl_register(struct net *net, char *dev_name,
snprintf(path, sizeof(path), "net/ipv6/conf/%s", dev_name);

p->sysctl_header = register_net_sysctl_sz(net, path, table,
- ARRAY_SIZE(addrconf_sysctl));
+ table_size);
if (!p->sysctl_header)
goto free;

diff --git a/net/ipv6/icmp.c b/net/ipv6/icmp.c
index d285c1f6f1a6..7b31674644ef 100644
--- a/net/ipv6/icmp.c
+++ b/net/ipv6/icmp.c
@@ -1206,7 +1206,6 @@ static struct ctl_table ipv6_icmp_table_template[] = {
.extra1 = SYSCTL_ZERO,
.extra2 = SYSCTL_ONE,
},
- { },
};

struct ctl_table * __net_init ipv6_icmp_sysctl_init(struct net *net)
diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c
index ee95cdcc8747..439f93512b0a 100644
--- a/net/ipv6/reassembly.c
+++ b/net/ipv6/reassembly.c
@@ -436,7 +436,6 @@ static struct ctl_table ip6_frags_ns_ctl_table[] = {
.mode = 0644,
.proc_handler = proc_dointvec_jiffies,
},
- { }
};

/* secret interval has been deprecated */
@@ -449,7 +448,6 @@ static struct ctl_table ip6_frags_ctl_table[] = {
.mode = 0644,
.proc_handler = proc_dointvec_jiffies,
},
- { }
};

static int __net_init ip6_frags_ns_sysctl_register(struct net *net)
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 3e0b2cb20fd2..c43b0616742e 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -6428,7 +6428,6 @@ static struct ctl_table ipv6_route_table_template[] = {
.extra1 = SYSCTL_ZERO,
.extra2 = SYSCTL_ONE,
},
- { }
};

struct ctl_table * __net_init ipv6_route_sysctl_init(struct net *net)
@@ -6452,10 +6451,6 @@ struct ctl_table * __net_init ipv6_route_sysctl_init(struct net *net)
table[8].data = &net->ipv6.sysctl.ip6_rt_min_advmss;
table[9].data = &net->ipv6.sysctl.ip6_rt_gc_min_interval;
table[10].data = &net->ipv6.sysctl.skip_notify_on_dev_down;
-
- /* Don't export sysctls to unprivileged users */
- if (net->user_ns != &init_user_ns)
- table[1].procname = NULL;
}

return table;
diff --git a/net/ipv6/sysctl_net_ipv6.c b/net/ipv6/sysctl_net_ipv6.c
index 75de55f907b0..c060285ff47f 100644
--- a/net/ipv6/sysctl_net_ipv6.c
+++ b/net/ipv6/sysctl_net_ipv6.c
@@ -213,7 +213,6 @@ static struct ctl_table ipv6_table_template[] = {
.proc_handler = proc_doulongvec_minmax,
.extra2 = &ioam6_id_wide_max,
},
- { }
};

static struct ctl_table ipv6_rotable[] = {
@@ -248,11 +247,11 @@ static struct ctl_table ipv6_rotable[] = {
.proc_handler = proc_dointvec,
},
#endif /* CONFIG_NETLABEL */
- { }
};

static int __net_init ipv6_sysctl_net_init(struct net *net)
{
+ size_t table_size = ARRAY_SIZE(ipv6_table_template);
struct ctl_table *ipv6_table;
struct ctl_table *ipv6_route_table;
struct ctl_table *ipv6_icmp_table;
@@ -264,7 +263,7 @@ static int __net_init ipv6_sysctl_net_init(struct net *net)
if (!ipv6_table)
goto out;
/* Update the variables to point into the current struct net */
- for (i = 0; i < ARRAY_SIZE(ipv6_table_template) - 1; i++)
+ for (i = 0; i < table_size; i++)
ipv6_table[i].data += (void *)net - (void *)&init_net;

ipv6_route_table = ipv6_route_sysctl_init(net);
@@ -276,8 +275,7 @@ static int __net_init ipv6_sysctl_net_init(struct net *net)
goto out_ipv6_route_table;

net->ipv6.sysctl.hdr = register_net_sysctl_sz(net, "net/ipv6",
- ipv6_table,
- ARRAY_SIZE(ipv6_table_template));
+ ipv6_table, table_size);
if (!net->ipv6.sysctl.hdr)
goto out_ipv6_icmp_table;

diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index 7924e08ee142..cc885d3aa9e5 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -184,7 +184,6 @@ static struct ctl_table xfrm6_policy_table[] = {
.mode = 0644,
.proc_handler = proc_dointvec,
},
- { }
};

static int __net_init xfrm6_net_sysctl_init(struct net *net)

--
2.43.0

2024-05-01 09:31:08

by Joel Granados via B4 Relay

[permalink] [raw]

Subject: [PATCH net-next v6 5/8] net: Remove ctl_table sentinel elements from several networking subsystems

2024-05-01 09:31:12

by Joel Granados via B4 Relay

[permalink] [raw]

Subject: [PATCH net-next v6 7/8] appletalk: Remove the now superfluous sentinel elements from ctl_table array

2024-05-01 09:31:32

by Joel Granados via B4 Relay

[permalink] [raw]

Subject: [PATCH net-next v6 3/8] net: rds: Remove the now superfluous sentinel elements from ctl_table array

2024-05-01 09:31:42

by Joel Granados via B4 Relay

[permalink] [raw]

Subject: [PATCH net-next v6 8/8] ax.25: x.25: Remove the now superfluous sentinel elements from ctl_table array

2024-05-01 09:32:00

by Joel Granados via B4 Relay

[permalink] [raw]

Subject: [PATCH net-next v6 4/8] net: sunrpc: Remove the now superfluous sentinel elements from ctl_table array

2024-05-01 13:16:21

by Sabrina Dubroca

[permalink] [raw]

Subject: Re: [PATCH net-next v6 8/8] ax.25: x.25: Remove the now superfluous sentinel elements from ctl_table array

2024-05-01, 11:29:32 +0200, Joel Granados via B4 Relay wrote:
> From: Joel Granados <[email protected]>
>
> This commit comes at the tail end of a greater effort to remove the
> empty elements at the end of the ctl_table arrays (sentinels) which will
> reduce the overall build time size of the kernel and run time memory
> bloat by ~64 bytes per sentinel (further information Link :
> https://lore.kernel.org/all/ZO5Yx5JFogGi%[email protected]/)
>
> Avoid a buffer overflow when traversing the ctl_table by ensuring that
> AX25_MAX_VALUES is the same as the size of ax25_param_table. This is
> done with a BUILD_BUG_ON where ax25_param_table is defined and a
> CONFIG_AX25_DAMA_SLAVE guard in the unnamed enum definition as well as
> in the ax25_dev_device_up and ax25_ds_set_timer functions.
^^
nit: not anymore ;)
(but not worth a repost IMO)

> diff --git a/net/ax25/ax25_ds_timer.c b/net/ax25/ax25_ds_timer.c
> index c4f8adbf8144..c50a58d9e368 100644
> --- a/net/ax25/ax25_ds_timer.c
> +++ b/net/ax25/ax25_ds_timer.c
> @@ -55,6 +55,7 @@ void ax25_ds_set_timer(ax25_dev *ax25_dev)
> ax25_dev->dama.slave_timeout =
> msecs_to_jiffies(ax25_dev->values[AX25_VALUES_DS_TIMEOUT]) / 10;
> mod_timer(&ax25_dev->dama.slave_timer, jiffies + HZ);
> + return;

nit: return not needed here since we're already at the bottom of the
function, but probably not worth a repost of the series.

> }

--
Sabrina

2024-05-01 13:22:30

by Sabrina Dubroca

[permalink] [raw]

Subject: Re: [PATCH net-next v6 0/8] sysctl: Remove sentinel elements from networking

2024-05-01, 11:29:24 +0200, Joel Granados via B4 Relay wrote:
> From: Joel Granados <[email protected]>
>
> What?
> These commits remove the sentinel element (last empty element) from the
> sysctl arrays of all the files under the "net/" directory that register
> a sysctl array. The merging of the preparation patches [4] to mainline
> allows us to just remove sentinel elements without changing behavior.
> This is safe because the sysctl registration code (register_sysctl() and
> friends) use the array size in addition to checking for a sentinel [1].
>
> Why?
> By removing the sysctl sentinel elements we avoid kernel bloat as
> ctl_table arrays get moved out of kernel/sysctl.c into their own
> respective subsystems. This move was started long ago to avoid merge
> conflicts; the sentinel removal bit came after Mathew Wilcox suggested
> it to avoid bloating the kernel by one element as arrays moved out. This
> patchset will reduce the overall build time size of the kernel and run
> time memory bloat by about ~64 bytes per declared ctl_table array (more
> info here [5]).
>
> When are we done?
> There are 4 patchest (25 commits [2]) that are still outstanding to
> completely remove the sentinels: files under "net/" (this patchset),
> files under "kernel/" dir, misc dirs (files under mm/ security/ and
> others) and the final set that removes the unneeded check for ->procname
> == NULL.
>
> Testing:
> * Ran sysctl selftests (./tools/testing/selftests/sysctl/sysctl.sh)
> * Ran this through 0-day with no errors or warnings
>
> Savings in vmlinux:
> A total of 64 bytes per sentinel is saved after removal; I measured in
> x86_64 to give an idea of the aggregated savings. The actual savings
> will depend on individual kernel configuration.
> * bloat-o-meter
> - The "yesall" config saves 3976 bytes (bloat-o-meter output [6])
> - A reduced config [3] saves 1263 bytes (bloat-o-meter output [7])
>
> Savings in allocated memory:
> None in this set but will occur when the superfluous allocations are
> removed from proc_sysctl.c. I include it here for context. The
> estimated savings during boot for config [3] are 6272 bytes. See [8]
> for how to measure it.
>
> Comments/feedback greatly appreciated
>
> Changes in v6:
> - Rebased onto net-next/main.
> - Besides re-running my cocci scripts, I ran a new find script [9].
> Found 0 hits in net/
> - Moved "i" variable declaraction out of for() in sysctl_core_net_init
> - Removed forgotten sentinel in mpls_table
> - Removed CONFIG_AX25_DAMA_SLAVE guard from net/ax25/ax25_ds_timer.c. It
> is not needed because that file is compiled only when
> CONFIG_AX25_DAMA_SLAVE is set.
> - When traversing smc_table, stop on ARRAY_SIZE instead of ARRAY_SIZE-1.
> - Link to v5: https://lore.kernel.org/r/[email protected]

I pointed out a few tiny details in the ax25 patch but either way, the
series looks good to me. Thanks!

Series:
Reviewed-by: Sabrina Dubroca <[email protected]>

Note that you could have kept the ack/reviewed-by on patch 4 since it
was not modified. Jeff and Chuck, your reviews got lost in the repost.

--
Sabrina

2024-05-03 02:27:42

by Allison Henderson

[permalink] [raw]

Subject: Re: [PATCH net-next v6 3/8] net: rds: Remove the now superfluous sentinel elements from ctl_table array

On Wed, 2024-05-01 at 11:29 +0200, Joel Granados via B4 Relay wrote:
> From: Joel Granados <[email protected]>
>
> This commit comes at the tail end of a greater effort to remove the
> empty elements at the end of the ctl_table arrays (sentinels) which
> will reduce the overall build time size of the kernel and run time
> memory bloat by ~64 bytes per sentinel (further information Link :
> https://lore.kernel.org/all/ZO5Yx5JFogGi%[email protected]/
> )
>
> * Remove sentinel element from ctl_table structs.
>
> Signed-off-by: Joel Granados <[email protected]>
These changes look fine to me. Thank you!
Acked-by: Allison Henderson <[email protected]>

> ---
> net/rds/ib_sysctl.c | 1 -
> net/rds/sysctl.c    | 1 -
> net/rds/tcp.c       | 1 -
> 3 files changed, 3 deletions(-)
>
> diff --git a/net/rds/ib_sysctl.c b/net/rds/ib_sysctl.c
> index e4e41b3afce7..2af678e71e3c 100644
> --- a/net/rds/ib_sysctl.c
> +++ b/net/rds/ib_sysctl.c
> @@ -103,7 +103,6 @@ static struct ctl_table rds_ib_sysctl_table[] = {
>                 .mode           = 0644,
>                 .proc_handler   = proc_dointvec,
>         },
> -       { }
> };
>
> void rds_ib_sysctl_exit(void)
> diff --git a/net/rds/sysctl.c b/net/rds/sysctl.c
> index e381bbcd9cc1..025f518a4349 100644
> --- a/net/rds/sysctl.c
> +++ b/net/rds/sysctl.c
> @@ -89,7 +89,6 @@ static struct ctl_table rds_sysctl_rds_table[] = {
>                 .mode           = 0644,
>                 .proc_handler   = proc_dointvec,
>         },
> -       { }
> };
>
> void rds_sysctl_exit(void)
> diff --git a/net/rds/tcp.c b/net/rds/tcp.c
> index 2dba7505b414..d8111ac83bb6 100644
> --- a/net/rds/tcp.c
> +++ b/net/rds/tcp.c
> @@ -86,7 +86,6 @@ static struct ctl_table rds_tcp_sysctl_table[] = {
>                 .proc_handler   = rds_tcp_skbuf_handler,
>                 .extra1         = &rds_tcp_min_rcvbuf,
>         },
> -       { }
> };
>
> u32 rds_tcp_write_seq(struct rds_tcp_connection *tc)
>

2024-05-03 12:18:33

by Joel Granados

[permalink] [raw]

Subject: Re: [PATCH net-next v6 8/8] ax.25: x.25: Remove the now superfluous sentinel elements from ctl_table array

On Wed, May 01, 2024 at 03:15:54PM +0200, Sabrina Dubroca wrote:
> 2024-05-01, 11:29:32 +0200, Joel Granados via B4 Relay wrote:
> > From: Joel Granados <[email protected]>
> >
> > This commit comes at the tail end of a greater effort to remove the
> > empty elements at the end of the ctl_table arrays (sentinels) which will
> > reduce the overall build time size of the kernel and run time memory
> > bloat by ~64 bytes per sentinel (further information Link :
> > https://lore.kernel.org/all/ZO5Yx5JFogGi%[email protected]/)
> >
> > Avoid a buffer overflow when traversing the ctl_table by ensuring that
> > AX25_MAX_VALUES is the same as the size of ax25_param_table. This is
> > done with a BUILD_BUG_ON where ax25_param_table is defined and a
> > CONFIG_AX25_DAMA_SLAVE guard in the unnamed enum definition as well as
> > in the ax25_dev_device_up and ax25_ds_set_timer functions.
> ^^
> nit: not anymore ;)
> (but not worth a repost IMO)
>
>
> > diff --git a/net/ax25/ax25_ds_timer.c b/net/ax25/ax25_ds_timer.c
> > index c4f8adbf8144..c50a58d9e368 100644
> > --- a/net/ax25/ax25_ds_timer.c
> > +++ b/net/ax25/ax25_ds_timer.c
> > @@ -55,6 +55,7 @@ void ax25_ds_set_timer(ax25_dev *ax25_dev)
> > ax25_dev->dama.slave_timeout =
> > msecs_to_jiffies(ax25_dev->values[AX25_VALUES_DS_TIMEOUT]) / 10;
> > mod_timer(&ax25_dev->dama.slave_timer, jiffies + HZ);
> > + return;
>
> nit: return not needed here since we're already at the bottom of the
> function, but probably not worth a repost of the series.
>
Thx. I will not repost, but I have changed them locally so they are
there in case a V7 is required.

Best
--

Joel Granados

Attachments:

(No filename) (1.72 kB)
signature.asc (673.00 B)
Download all attachments

2024-05-07 08:02:38

by Joel Granados

[permalink] [raw]

Subject: Re: [PATCH net-next v6 8/8] ax.25: x.25: Remove the now superfluous sentinel elements from ctl_table array

On Fri, May 03, 2024 at 06:23:14PM +0300, Dan Carpenter wrote:
> On Fri, May 03, 2024 at 02:18:11PM +0200, Joel Granados wrote:
> > On Wed, May 01, 2024 at 03:15:54PM +0200, Sabrina Dubroca wrote:
> > > 2024-05-01, 11:29:32 +0200, Joel Granados via B4 Relay wrote:
> > > > From: Joel Granados <[email protected]>
> > > > diff --git a/net/ax25/ax25_ds_timer.c b/net/ax25/ax25_ds_timer.c
> > > > index c4f8adbf8144..c50a58d9e368 100644
> > > > --- a/net/ax25/ax25_ds_timer.c
> > > > +++ b/net/ax25/ax25_ds_timer.c
> > > > @@ -55,6 +55,7 @@ void ax25_ds_set_timer(ax25_dev *ax25_dev)
> > > > ax25_dev->dama.slave_timeout =
> > > > msecs_to_jiffies(ax25_dev->values[AX25_VALUES_DS_TIMEOUT]) / 10;
> > > > mod_timer(&ax25_dev->dama.slave_timer, jiffies + HZ);
> > > > + return;
> > >
> > > nit: return not needed here since we're already at the bottom of the
> > > function, but probably not worth a repost of the series.
> > >
> > Thx. I will not repost, but I have changed them locally so they are
> > there in case a V7 is required.
> >
>
> It's a checkpatch.pl -f warning so we probably will want to fix it
> eventually.

According to [1] the patchset has already been applied. So I'll just
send another patch for it to be applied on top.

Thx for pointing this out.

[1] https://patchwork.kernel.org/project/netdevbpf/patch/[email protected]/
--

Joel Granados

Attachments:

(No filename) (1.42 kB)
signature.asc (673.00 B)
Download all attachments