Re: [PATCH v3 1/4] networking: Remove the now superfluous sentinel elements from ctl_table array

From: Joel Granados
Date: Tue Apr 16 2024 - 05:46:28 EST

Next message: Chen Wang: "[PATCH 0/3] mmc: sdhci-of-dwcmshc: support Sophgo SG2042"
Previous message: Chen Wang: "[PATCH 1/1] mmc: sdhci-of-dwcmshc: add callback framework for expansion"
In reply to: Joel Granados: "Re: [PATCH v3 1/4] networking: Remove the now superfluous sentinel elements from ctl_table array"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Mon, Apr 15, 2024 at 04:12:10PM -0700, Kuniyuki Iwashima wrote:
> From: Joel Granados via B4 Relay <devnull+j.granados.samsung.com@xxxxxxxxxx>
> Date: Fri, 12 Apr 2024 16:48:29 +0200
> > From: Joel Granados <j.granados@xxxxxxxxxxx>
...
> > Signed-off-by: Joel Granados <j.granados@xxxxxxxxxxx>
> > ---
> > net/core/neighbour.c | 5 +----
> > net/core/sysctl_net_core.c | 9 ++++-----
> > 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 | 5 +----
> > net/ipv6/icmp.c | 1 -
> > net/ipv6/reassembly.c | 2 --
> > net/ipv6/route.c | 5 -----
> > net/ipv6/sysctl_net_ipv6.c | 4 +---
> > net/ipv6/xfrm6_policy.c | 1 -
> > net/llc/sysctl_net_llc.c | 8 ++------
> > net/mpls/af_mpls.c | 3 +--
> > net/mptcp/ctrl.c | 1 -
> > 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 | 6 +-----
> > net/smc/smc_sysctl.c | 1 -
> > 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 +----
> > 35 files changed, 20 insertions(+), 81 deletions(-)
>
> You may want to split patch based on subsystem or the type of changes
> to make review easier.

That is fair. It is a big chunk:). I'll put the trivial patches together
to avoid having an 18 commits to instead have 8. This is my proposal
based on MAINTAINERS file:

### Not in MAINTAINERS / Orphaned
net/core/neighbour.c
net/core/sysctl_net_core.c
net/ieee802154/6lowpan/reassembly.c
net/mpls/af_mpls.c
net/unix/sysctl_net_unix.c
net/dccp/sysctl.c

### NETWORKING
net/ipv4/devinet.c
net/ipv4/ip_fragment.c
net/ipv4/route.c
net/ipv4/sysctl_net_ipv4.c
net/ipv4/xfrm4_policy.c
net/ipv6/addrconf.c
net/ipv6/icmp.c
net/ipv6/reassembly.c
net/ipv6/route.c
net/ipv6/sysctl_net_ipv6.c
net/ipv6/xfrm6_policy.c

### RDS
net/rds/ib_sysctl.c
net/rds/sysctl.c
net/rds/tcp.c

### SUNRPC
net/sunrpc/sysctl.c
net/sunrpc/xprtrdma/svc_rdma.c
net/sunrpc/xprtrdma/transport.c
net/sunrpc/xprtsock.c

### LLC/MTCP/NETROM/PHONET/ROSE/RXRPC/SCTP/SMC/TIPC/x.25/IPSEC
net/llc/sysctl_net_llc.c
net/mptcp/ctrl.c
net/netrom/sysctl_net_netrom.c
net/phonet/sysctl.c
net/rose/sysctl_net_rose.c
net/rxrpc/sysctl.c
net/sctp/sysctl.c
net/smc/smc_sysctl.c
net/tipc/sysctl.c
net/x25/sysctl_net_x25.c
net/xfrm/xfrm_sysctl.c

>
>
> >
> > diff --git a/net/core/neighbour.c b/net/core/neighbour.c
> > index 552719c3bbc3..b0327402b3e6 100644
> > --- a/net/core/neighbour.c
> > +++ b/net/core/neighbour.c
> > @@ -3728,7 +3728,7 @@ static int neigh_proc_base_reachable_time(struct ctl_table *ctl, int write,
> >
> > static struct neigh_sysctl_table {
> > struct ctl_table_header *sysctl_header;
> > - struct ctl_table neigh_vars[NEIGH_VAR_MAX + 1];
> > + struct ctl_table neigh_vars[NEIGH_VAR_MAX];
> > } neigh_sysctl_template __read_mostly = {
> > .neigh_vars = {
> > NEIGH_SYSCTL_ZERO_INTMAX_ENTRY(MCAST_PROBES, "mcast_solicit"),
> > @@ -3779,7 +3779,6 @@ static struct neigh_sysctl_table {
> > .extra2 = SYSCTL_INT_MAX,
> > .proc_handler = proc_dointvec_minmax,
> > },
> > - {},
> > },
> > };
> >
> > @@ -3807,8 +3806,6 @@ int neigh_sysctl_register(struct net_device *dev, struct neigh_parms *p,
> > if (dev) {
> > dev_name_source = dev->name;
> > /* Terminate the table early */
>
> You can remove this comment.
Why? I do not think we should remove it because it is what the change to
neigh_vars_size is doing.

>
>
> > - memset(&t->neigh_vars[NEIGH_VAR_GC_INTERVAL], 0,
> > - sizeof(t->neigh_vars[NEIGH_VAR_GC_INTERVAL]));
> > neigh_vars_size = NEIGH_VAR_BASE_REACHABLE_TIME_MS + 1;
> > } else {
> > struct neigh_table *tbl = p->tbl;
> > diff --git a/net/core/sysctl_net_core.c b/net/core/sysctl_net_core.c
> > index 6973dda3abda..46f5143e86be 100644
> > --- a/net/core/sysctl_net_core.c
> > +++ b/net/core/sysctl_net_core.c
> > @@ -660,7 +660,6 @@ static struct ctl_table net_core_table[] = {
> > .proc_handler = proc_dointvec_minmax,
> > .extra1 = SYSCTL_ZERO,
> > },
> > - { }
> > };
> >
> > static struct ctl_table netns_core_table[] = {
> > @@ -697,7 +696,6 @@ static struct ctl_table netns_core_table[] = {
> > .extra2 = SYSCTL_ONE,
> > .proc_handler = proc_dou8vec_minmax,
> > },
> > - { }
> > };
> >
> > static int __init fb_tunnels_only_for_init_net_sysctl_setup(char *str)
> > @@ -715,7 +713,8 @@ __setup("fb_tunnels=", fb_tunnels_only_for_init_net_sysctl_setup);
> >
> > static __net_init int sysctl_core_net_init(struct net *net)
> > {
> > - struct ctl_table *tbl, *tmp;
> > + struct ctl_table *tbl;
> > + size_t table_size = ARRAY_SIZE(netns_core_table);
>
> When you add a new variable, please keep reverse xmas tree.
Thx for pointing this out. Was not aware of this quirk in net code. Will
include it for my next version.

>
> Also, you can reuse this variable for the following
> register_net_sysctl_sz(), but it's inconsistent in the
> this patch..
>
> table_size
> * sysctl_route_net_init
> * ipv4_sysctl_init_net
>
> ARRAY_SIZE
> * __addrconf_sysctl_register
> * ipv6_sysctl_net_init
> * mpls_dev_sysctl_register
> * sctp_sysctl_net_register
>
>
> >
> > tbl = netns_core_table;
> > if (!net_eq(net, &init_net)) {
> > @@ -723,8 +722,8 @@ static __net_init int sysctl_core_net_init(struct net *net)
> > if (tbl == NULL)
> > goto err_dup;
> >
> > - for (tmp = tbl; tmp->procname; tmp++)
> > - tmp->data += (char *)net - (char *)&init_net;
> > + for (int i = 0; i < table_size; ++i)
> > .devinet_vars = {
...
> > .extra1 = SYSCTL_ONE,
> > },
> > - { }
> > };
> >
> > static __net_init int ipv4_sysctl_init_net(struct net *net)
> > {
> > struct ctl_table *table;
> > + size_t table_size = ARRAY_SIZE(ipv4_net_table);
>
> nit: keep reverse xmax tree order.
Ok.

>
>
> >
> > table = ipv4_net_table;
> > if (!net_eq(net, &init_net)) {
> > @@ -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++) {
...
> >
> > static int __addrconf_sysctl_register(struct net *net, char *dev_name,
> > @@ -7197,7 +7194,7 @@ 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 < ARRAY_SIZE(addrconf_sysctl); i++) {
>
> ^^^
Did you mean reuse variable here? What does this mean?

>
>
> > 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
> > diff --git a/net/ipv6/icmp.c b/net/ipv6/icmp.c
> > index 1635da07285f..91cbf8e8009f 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,
...
> > - { }
> > };
> >
> > static int __net_init ipv6_sysctl_net_init(struct net *net)
> > @@ -264,7 +262,7 @@ 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 < ARRAY_SIZE(ipv6_table_template); i++)
>
> ^^^
Did you mean reuse variable here? What does this mean?
>
>
> > ipv6_table[i].data += (void *)net - (void *)&init_net;
> >
> > ipv6_route_table = ipv6_route_sysctl_init(net);
> > diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
> > index 42fb6996b077..499b5f5c19fc 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)
> > diff --git a/net/llc/sysctl_net_llc.c b/net/llc/sysctl_net_llc.c
> > index 8443a6d841b0..72e101135f8c 100644
> > --- a/net/llc/sysctl_net_llc.c
> > +++ b/net/llc/sysctl_net_llc.c
> > @@ -44,11 +44,6 @@ static struct ctl_table llc2_timeout_table[] = {
> > .mode = 0644,
> > .proc_handler = proc_dointvec_jiffies,
> > },
> > - { },
> > -};
> > -
> > -static struct ctl_table llc_station_table[] = {
> > - { },
> > };
> >
> > static struct ctl_table_header *llc2_timeout_header;
> > @@ -56,8 +51,9 @@ static struct ctl_table_header *llc_station_header;
> >
> > int __init llc_sysctl_init(void)
> > {
> > + struct ctl_table empty[1] = {};
> > llc2_timeout_header = register_net_sysctl(&init_net, "net/llc/llc2/timeout", llc2_timeout_table);
> > - llc_station_header = register_net_sysctl(&init_net, "net/llc/station", llc_station_table);
> > + llc_station_header = register_net_sysctl_sz(&init_net, "net/llc/station", empty, 0);
>
> Do we really need this ... ??
That is a good question, but its something that needs to be address
outside this patchset. I'm just keeping current behaviour.

>
>
> >
> > if (!llc2_timeout_header || !llc_station_header) {
> > llc_sysctl_exit();
> > diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c
> > index 6dab883a08dd..e163fac55ffa 100644
> > --- a/net/mpls/af_mpls.c
> > +++ b/net/mpls/af_mpls.c
> > @@ -1393,7 +1393,6 @@ static const struct ctl_table mpls_dev_table[] = {
> > .proc_handler = mpls_conf_proc,
> > .data = MPLS_PERDEV_SYSCTL_OFFSET(input_enabled),
> > },
> > - { }
> > };
> >
> > static int mpls_dev_sysctl_register(struct net_device *dev,
> > @@ -2689,7 +2688,7 @@ static int mpls_net_init(struct net *net)
> > /* Table data contains only offsets relative to the base of
> > * the mdev at this point, so make them absolute.
> > */
> > - for (i = 0; i < ARRAY_SIZE(mpls_table) - 1; i++)
> > + for (i = 0; i < ARRAY_SIZE(mpls_table); i++)
>
> ^^^
>
>
> > table[i].data = (char *)net + (uintptr_t)table[i].data;
> >
> > net->mpls.ctl = register_net_sysctl_sz(net, "net/mpls", table,
> > diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
> > index 13fe0748dde8..8bf7c26a0878 100644
> > --- a/net/mptcp/ctrl.c
> > +++ b/net/mptcp/ctrl.c
...
> >
> > net->xfrm.sysctl_hdr = register_net_sysctl_sz(net, "net/core", table,
> > table_size);
> >
> > --
> > 2.43.0
> B

Thx for the review

Best
--

Joel Granados

Attachment: signature.asc
Description: PGP signature

Next message: Chen Wang: "[PATCH 0/3] mmc: sdhci-of-dwcmshc: support Sophgo SG2042"
Previous message: Chen Wang: "[PATCH 1/1] mmc: sdhci-of-dwcmshc: add callback framework for expansion"
In reply to: Joel Granados: "Re: [PATCH v3 1/4] networking: Remove the now superfluous sentinel elements from ctl_table array"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]