2013-10-18 15:43:34 +00:00
|
|
|
/*
|
|
|
|
* drivers/net/bond/bond_options.c - bonding options
|
|
|
|
* Copyright (c) 2013 Jiri Pirko <jiri@resnulli.us>
|
2013-12-12 22:09:55 +00:00
|
|
|
* Copyright (c) 2013 Scott Feldman <sfeldma@cumulusnetworks.com>
|
2013-10-18 15:43:34 +00:00
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or
|
|
|
|
* (at your option) any later version.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
|
|
|
|
|
|
|
|
#include <linux/errno.h>
|
|
|
|
#include <linux/if.h>
|
2013-10-18 15:43:35 +00:00
|
|
|
#include <linux/netdevice.h>
|
|
|
|
#include <linux/rwlock.h>
|
|
|
|
#include <linux/rcupdate.h>
|
bonding: add infrastructure for an option API
This patch adds the necessary basic infrastructure to support
centralized and unified option manipulation API for the bonding. The new
structure bond_option will be used to describe each option with its
dependencies on modes which will be checked automatically thus removing a
lot of duplicated code. Also automatic range checking is added for
some options. Currently the option setting function requires RTNL to
be acquired prior to calling it, since many options already rely on RTNL
it seemed like the best choice to protect all against common race
conditions.
In order to add an option the following steps need to be done:
1. Add an entry BOND_OPT_<option> to bond_options.h so it gets a unique id
and a bit corresponding to the id
2. Add a bond_option entry to the bond_opts[] array in bond_options.c which
describes the option, its dependencies and its manipulation function
3. Add code to export the option through sysfs and/or as a module parameter
(the sysfs export will be made automatically in the future)
The options can have different flags set, currently the following are
supported:
BOND_OPTFLAG_NOSLAVES - require that the bond device has no slaves prior
to setting the option
BOND_OPTFLAG_IFDOWN - require that the bond device is down prior to
setting the option
BOND_OPTFLAG_RAWVAL - don't parse the value but return it raw for the
option to parse
There's a new value structure to describe different types of values
which can have the following flags:
BOND_VALFLAG_DEFAULT - marks the default option (permanent string alias
to this option is "default")
BOND_VALFLAG_MIN - the minimum value that this option can have
BOND_VALFLAG_MAX - the maximum value that this option can have
An example would be nice here, so if we have an option which can have
the values "off"(2), "special"(4, default) and supports a range, say
16 - 32, it should be defined as follows:
"off", 2,
"special", 4, BOND_VALFLAG_DEFAULT,
"rangemin", 16, BOND_VALFLAG_MIN,
"rangemax", 32, BOND_VALFLAG_MAX
So we have the valid intervals: [2, 2], [4, 4], [16, 32]
Also the valid strings: "off" = 2, "special" and "default" = 4
"rangemin" = 16, "rangemax" = 32
BOND_VALFLAG_(MIN|MAX) can be used to specify a valid range for an
option, if MIN is omitted then 0 is considered as a minimum. If an
exact match is found in the values[] table it will be returned,
otherwise the range is tried (if available).
The option parameter passing is done by using a special structure called
bond_opt_value which can take either a string or a value to parse. One
of the bond_opt_init(val|str) macros should be used depending on which
one does the user want to parse (string or value). Then a call to
__bond_opt_set should be done under RTNL.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-01-22 13:53:16 +00:00
|
|
|
#include <linux/ctype.h>
|
2013-10-18 15:43:34 +00:00
|
|
|
#include "bonding.h"
|
|
|
|
|
2014-01-22 13:53:17 +00:00
|
|
|
static struct bond_opt_value bond_mode_tbl[] = {
|
|
|
|
{ "balance-rr", BOND_MODE_ROUNDROBIN, BOND_VALFLAG_DEFAULT},
|
|
|
|
{ "active-backup", BOND_MODE_ACTIVEBACKUP, 0},
|
|
|
|
{ "balance-xor", BOND_MODE_XOR, 0},
|
|
|
|
{ "broadcast", BOND_MODE_BROADCAST, 0},
|
|
|
|
{ "802.3ad", BOND_MODE_8023AD, 0},
|
|
|
|
{ "balance-tlb", BOND_MODE_TLB, 0},
|
|
|
|
{ "balance-alb", BOND_MODE_ALB, 0},
|
|
|
|
{ NULL, -1, 0},
|
|
|
|
};
|
|
|
|
|
2014-01-22 13:53:18 +00:00
|
|
|
static struct bond_opt_value bond_pps_tbl[] = {
|
|
|
|
{ "default", 1, BOND_VALFLAG_DEFAULT},
|
|
|
|
{ "maxval", USHRT_MAX, BOND_VALFLAG_MAX},
|
|
|
|
{ NULL, -1, 0},
|
|
|
|
};
|
|
|
|
|
2014-01-22 13:53:19 +00:00
|
|
|
static struct bond_opt_value bond_xmit_hashtype_tbl[] = {
|
|
|
|
{ "layer2", BOND_XMIT_POLICY_LAYER2, BOND_VALFLAG_DEFAULT},
|
|
|
|
{ "layer3+4", BOND_XMIT_POLICY_LAYER34, 0},
|
|
|
|
{ "layer2+3", BOND_XMIT_POLICY_LAYER23, 0},
|
|
|
|
{ "encap2+3", BOND_XMIT_POLICY_ENCAP23, 0},
|
|
|
|
{ "encap3+4", BOND_XMIT_POLICY_ENCAP34, 0},
|
|
|
|
{ NULL, -1, 0},
|
|
|
|
};
|
|
|
|
|
2014-01-22 13:53:20 +00:00
|
|
|
static struct bond_opt_value bond_arp_validate_tbl[] = {
|
|
|
|
{ "none", BOND_ARP_VALIDATE_NONE, BOND_VALFLAG_DEFAULT},
|
|
|
|
{ "active", BOND_ARP_VALIDATE_ACTIVE, 0},
|
|
|
|
{ "backup", BOND_ARP_VALIDATE_BACKUP, 0},
|
|
|
|
{ "all", BOND_ARP_VALIDATE_ALL, 0},
|
|
|
|
{ NULL, -1, 0},
|
|
|
|
};
|
|
|
|
|
2014-01-22 13:53:21 +00:00
|
|
|
static struct bond_opt_value bond_arp_all_targets_tbl[] = {
|
|
|
|
{ "any", BOND_ARP_TARGETS_ANY, BOND_VALFLAG_DEFAULT},
|
|
|
|
{ "all", BOND_ARP_TARGETS_ALL, 0},
|
|
|
|
{ NULL, -1, 0},
|
|
|
|
};
|
|
|
|
|
2014-01-22 13:53:22 +00:00
|
|
|
static struct bond_opt_value bond_fail_over_mac_tbl[] = {
|
|
|
|
{ "none", BOND_FOM_NONE, BOND_VALFLAG_DEFAULT},
|
|
|
|
{ "active", BOND_FOM_ACTIVE, 0},
|
|
|
|
{ "follow", BOND_FOM_FOLLOW, 0},
|
|
|
|
{ NULL, -1, 0},
|
|
|
|
};
|
|
|
|
|
2014-01-22 13:53:23 +00:00
|
|
|
static struct bond_opt_value bond_intmax_tbl[] = {
|
|
|
|
{ "off", 0, BOND_VALFLAG_DEFAULT},
|
|
|
|
{ "maxval", INT_MAX, BOND_VALFLAG_MAX},
|
|
|
|
};
|
|
|
|
|
bonding: add infrastructure for an option API
This patch adds the necessary basic infrastructure to support
centralized and unified option manipulation API for the bonding. The new
structure bond_option will be used to describe each option with its
dependencies on modes which will be checked automatically thus removing a
lot of duplicated code. Also automatic range checking is added for
some options. Currently the option setting function requires RTNL to
be acquired prior to calling it, since many options already rely on RTNL
it seemed like the best choice to protect all against common race
conditions.
In order to add an option the following steps need to be done:
1. Add an entry BOND_OPT_<option> to bond_options.h so it gets a unique id
and a bit corresponding to the id
2. Add a bond_option entry to the bond_opts[] array in bond_options.c which
describes the option, its dependencies and its manipulation function
3. Add code to export the option through sysfs and/or as a module parameter
(the sysfs export will be made automatically in the future)
The options can have different flags set, currently the following are
supported:
BOND_OPTFLAG_NOSLAVES - require that the bond device has no slaves prior
to setting the option
BOND_OPTFLAG_IFDOWN - require that the bond device is down prior to
setting the option
BOND_OPTFLAG_RAWVAL - don't parse the value but return it raw for the
option to parse
There's a new value structure to describe different types of values
which can have the following flags:
BOND_VALFLAG_DEFAULT - marks the default option (permanent string alias
to this option is "default")
BOND_VALFLAG_MIN - the minimum value that this option can have
BOND_VALFLAG_MAX - the maximum value that this option can have
An example would be nice here, so if we have an option which can have
the values "off"(2), "special"(4, default) and supports a range, say
16 - 32, it should be defined as follows:
"off", 2,
"special", 4, BOND_VALFLAG_DEFAULT,
"rangemin", 16, BOND_VALFLAG_MIN,
"rangemax", 32, BOND_VALFLAG_MAX
So we have the valid intervals: [2, 2], [4, 4], [16, 32]
Also the valid strings: "off" = 2, "special" and "default" = 4
"rangemin" = 16, "rangemax" = 32
BOND_VALFLAG_(MIN|MAX) can be used to specify a valid range for an
option, if MIN is omitted then 0 is considered as a minimum. If an
exact match is found in the values[] table it will be returned,
otherwise the range is tried (if available).
The option parameter passing is done by using a special structure called
bond_opt_value which can take either a string or a value to parse. One
of the bond_opt_init(val|str) macros should be used depending on which
one does the user want to parse (string or value). Then a call to
__bond_opt_set should be done under RTNL.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-01-22 13:53:16 +00:00
|
|
|
static struct bond_option bond_opts[] = {
|
2014-01-22 13:53:17 +00:00
|
|
|
[BOND_OPT_MODE] = {
|
|
|
|
.id = BOND_OPT_MODE,
|
|
|
|
.name = "mode",
|
|
|
|
.desc = "bond device mode",
|
|
|
|
.flags = BOND_OPTFLAG_NOSLAVES | BOND_OPTFLAG_IFDOWN,
|
|
|
|
.values = bond_mode_tbl,
|
|
|
|
.set = bond_option_mode_set
|
|
|
|
},
|
2014-01-22 13:53:18 +00:00
|
|
|
[BOND_OPT_PACKETS_PER_SLAVE] = {
|
|
|
|
.id = BOND_OPT_PACKETS_PER_SLAVE,
|
|
|
|
.name = "packets_per_slave",
|
|
|
|
.desc = "Packets to send per slave in RR mode",
|
|
|
|
.unsuppmodes = BOND_MODE_ALL_EX(BIT(BOND_MODE_ROUNDROBIN)),
|
|
|
|
.values = bond_pps_tbl,
|
|
|
|
.set = bond_option_pps_set
|
|
|
|
},
|
2014-01-22 13:53:19 +00:00
|
|
|
[BOND_OPT_XMIT_HASH] = {
|
|
|
|
.id = BOND_OPT_XMIT_HASH,
|
|
|
|
.name = "xmit_hash_policy",
|
|
|
|
.desc = "balance-xor and 802.3ad hashing method",
|
|
|
|
.values = bond_xmit_hashtype_tbl,
|
|
|
|
.set = bond_option_xmit_hash_policy_set
|
|
|
|
},
|
2014-01-22 13:53:20 +00:00
|
|
|
[BOND_OPT_ARP_VALIDATE] = {
|
|
|
|
.id = BOND_OPT_ARP_VALIDATE,
|
|
|
|
.name = "arp_validate",
|
|
|
|
.desc = "validate src/dst of ARP probes",
|
|
|
|
.unsuppmodes = BOND_MODE_ALL_EX(BIT(BOND_MODE_ACTIVEBACKUP)),
|
|
|
|
.values = bond_arp_validate_tbl,
|
|
|
|
.set = bond_option_arp_validate_set
|
|
|
|
},
|
2014-01-22 13:53:21 +00:00
|
|
|
[BOND_OPT_ARP_ALL_TARGETS] = {
|
|
|
|
.id = BOND_OPT_ARP_ALL_TARGETS,
|
|
|
|
.name = "arp_all_targets",
|
|
|
|
.desc = "fail on any/all arp targets timeout",
|
|
|
|
.values = bond_arp_all_targets_tbl,
|
|
|
|
.set = bond_option_arp_all_targets_set
|
|
|
|
},
|
2014-01-22 13:53:22 +00:00
|
|
|
[BOND_OPT_FAIL_OVER_MAC] = {
|
|
|
|
.id = BOND_OPT_FAIL_OVER_MAC,
|
|
|
|
.name = "fail_over_mac",
|
|
|
|
.desc = "For active-backup, do not set all slaves to the same MAC",
|
|
|
|
.flags = BOND_OPTFLAG_NOSLAVES,
|
|
|
|
.values = bond_fail_over_mac_tbl,
|
|
|
|
.set = bond_option_fail_over_mac_set
|
|
|
|
},
|
2014-01-22 13:53:23 +00:00
|
|
|
[BOND_OPT_ARP_INTERVAL] = {
|
|
|
|
.id = BOND_OPT_ARP_INTERVAL,
|
|
|
|
.name = "arp_interval",
|
|
|
|
.desc = "arp interval in milliseconds",
|
|
|
|
.unsuppmodes = BIT(BOND_MODE_8023AD) | BIT(BOND_MODE_TLB) |
|
|
|
|
BIT(BOND_MODE_ALB),
|
|
|
|
.values = bond_intmax_tbl,
|
|
|
|
.set = bond_option_arp_interval_set
|
|
|
|
},
|
bonding: add infrastructure for an option API
This patch adds the necessary basic infrastructure to support
centralized and unified option manipulation API for the bonding. The new
structure bond_option will be used to describe each option with its
dependencies on modes which will be checked automatically thus removing a
lot of duplicated code. Also automatic range checking is added for
some options. Currently the option setting function requires RTNL to
be acquired prior to calling it, since many options already rely on RTNL
it seemed like the best choice to protect all against common race
conditions.
In order to add an option the following steps need to be done:
1. Add an entry BOND_OPT_<option> to bond_options.h so it gets a unique id
and a bit corresponding to the id
2. Add a bond_option entry to the bond_opts[] array in bond_options.c which
describes the option, its dependencies and its manipulation function
3. Add code to export the option through sysfs and/or as a module parameter
(the sysfs export will be made automatically in the future)
The options can have different flags set, currently the following are
supported:
BOND_OPTFLAG_NOSLAVES - require that the bond device has no slaves prior
to setting the option
BOND_OPTFLAG_IFDOWN - require that the bond device is down prior to
setting the option
BOND_OPTFLAG_RAWVAL - don't parse the value but return it raw for the
option to parse
There's a new value structure to describe different types of values
which can have the following flags:
BOND_VALFLAG_DEFAULT - marks the default option (permanent string alias
to this option is "default")
BOND_VALFLAG_MIN - the minimum value that this option can have
BOND_VALFLAG_MAX - the maximum value that this option can have
An example would be nice here, so if we have an option which can have
the values "off"(2), "special"(4, default) and supports a range, say
16 - 32, it should be defined as follows:
"off", 2,
"special", 4, BOND_VALFLAG_DEFAULT,
"rangemin", 16, BOND_VALFLAG_MIN,
"rangemax", 32, BOND_VALFLAG_MAX
So we have the valid intervals: [2, 2], [4, 4], [16, 32]
Also the valid strings: "off" = 2, "special" and "default" = 4
"rangemin" = 16, "rangemax" = 32
BOND_VALFLAG_(MIN|MAX) can be used to specify a valid range for an
option, if MIN is omitted then 0 is considered as a minimum. If an
exact match is found in the values[] table it will be returned,
otherwise the range is tried (if available).
The option parameter passing is done by using a special structure called
bond_opt_value which can take either a string or a value to parse. One
of the bond_opt_init(val|str) macros should be used depending on which
one does the user want to parse (string or value). Then a call to
__bond_opt_set should be done under RTNL.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-01-22 13:53:16 +00:00
|
|
|
{ }
|
|
|
|
};
|
|
|
|
|
|
|
|
/* Searches for a value in opt's values[] table */
|
|
|
|
struct bond_opt_value *bond_opt_get_val(unsigned int option, u64 val)
|
|
|
|
{
|
|
|
|
struct bond_option *opt;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
opt = bond_opt_get(option);
|
|
|
|
if (WARN_ON(!opt))
|
|
|
|
return NULL;
|
|
|
|
for (i = 0; opt->values && opt->values[i].string; i++)
|
|
|
|
if (opt->values[i].value == val)
|
|
|
|
return &opt->values[i];
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Searches for a value in opt's values[] table which matches the flagmask */
|
|
|
|
static struct bond_opt_value *bond_opt_get_flags(const struct bond_option *opt,
|
|
|
|
u32 flagmask)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; opt->values && opt->values[i].string; i++)
|
|
|
|
if (opt->values[i].flags & flagmask)
|
|
|
|
return &opt->values[i];
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* If maxval is missing then there's no range to check. In case minval is
|
|
|
|
* missing then it's considered to be 0.
|
|
|
|
*/
|
|
|
|
static bool bond_opt_check_range(const struct bond_option *opt, u64 val)
|
|
|
|
{
|
|
|
|
struct bond_opt_value *minval, *maxval;
|
|
|
|
|
|
|
|
minval = bond_opt_get_flags(opt, BOND_VALFLAG_MIN);
|
|
|
|
maxval = bond_opt_get_flags(opt, BOND_VALFLAG_MAX);
|
|
|
|
if (!maxval || (minval && val < minval->value) || val > maxval->value)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* bond_opt_parse - parse option value
|
|
|
|
* @opt: the option to parse against
|
|
|
|
* @val: value to parse
|
|
|
|
*
|
|
|
|
* This function tries to extract the value from @val and check if it's
|
|
|
|
* a possible match for the option and returns NULL if a match isn't found,
|
|
|
|
* or the struct_opt_value that matched. It also strips the new line from
|
|
|
|
* @val->string if it's present.
|
|
|
|
*/
|
|
|
|
struct bond_opt_value *bond_opt_parse(const struct bond_option *opt,
|
|
|
|
struct bond_opt_value *val)
|
|
|
|
{
|
|
|
|
char *p, valstr[BOND_OPT_MAX_NAMELEN + 1] = { 0, };
|
|
|
|
struct bond_opt_value *tbl, *ret = NULL;
|
|
|
|
bool checkval;
|
|
|
|
int i, rv;
|
|
|
|
|
|
|
|
/* No parsing if the option wants a raw val */
|
|
|
|
if (opt->flags & BOND_OPTFLAG_RAWVAL)
|
|
|
|
return val;
|
|
|
|
|
|
|
|
tbl = opt->values;
|
|
|
|
if (!tbl)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
/* ULLONG_MAX is used to bypass string processing */
|
|
|
|
checkval = val->value != ULLONG_MAX;
|
|
|
|
if (!checkval) {
|
|
|
|
if (!val->string)
|
|
|
|
goto out;
|
|
|
|
p = strchr(val->string, '\n');
|
|
|
|
if (p)
|
|
|
|
*p = '\0';
|
|
|
|
for (p = val->string; *p; p++)
|
|
|
|
if (!(isdigit(*p) || isspace(*p)))
|
|
|
|
break;
|
|
|
|
/* The following code extracts the string to match or the value
|
|
|
|
* and sets checkval appropriately
|
|
|
|
*/
|
|
|
|
if (*p) {
|
|
|
|
rv = sscanf(val->string, "%32s", valstr);
|
|
|
|
} else {
|
|
|
|
rv = sscanf(val->string, "%llu", &val->value);
|
|
|
|
checkval = true;
|
|
|
|
}
|
|
|
|
if (!rv)
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; tbl[i].string; i++) {
|
|
|
|
/* Check for exact match */
|
|
|
|
if (checkval) {
|
|
|
|
if (val->value == tbl[i].value)
|
|
|
|
ret = &tbl[i];
|
|
|
|
} else {
|
|
|
|
if (!strcmp(valstr, "default") &&
|
|
|
|
(tbl[i].flags & BOND_VALFLAG_DEFAULT))
|
|
|
|
ret = &tbl[i];
|
|
|
|
|
|
|
|
if (!strcmp(valstr, tbl[i].string))
|
|
|
|
ret = &tbl[i];
|
|
|
|
}
|
|
|
|
/* Found an exact match */
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
/* Possible range match */
|
|
|
|
if (checkval && bond_opt_check_range(opt, val->value))
|
|
|
|
ret = val;
|
|
|
|
out:
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Check opt's dependencies against bond mode and currently set options */
|
|
|
|
static int bond_opt_check_deps(struct bonding *bond,
|
|
|
|
const struct bond_option *opt)
|
|
|
|
{
|
|
|
|
struct bond_params *params = &bond->params;
|
|
|
|
|
|
|
|
if (test_bit(params->mode, &opt->unsuppmodes))
|
|
|
|
return -EACCES;
|
|
|
|
if ((opt->flags & BOND_OPTFLAG_NOSLAVES) && bond_has_slaves(bond))
|
|
|
|
return -ENOTEMPTY;
|
|
|
|
if ((opt->flags & BOND_OPTFLAG_IFDOWN) && (bond->dev->flags & IFF_UP))
|
|
|
|
return -EBUSY;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void bond_opt_dep_print(struct bonding *bond,
|
|
|
|
const struct bond_option *opt)
|
|
|
|
{
|
2014-01-22 13:53:17 +00:00
|
|
|
struct bond_opt_value *modeval;
|
bonding: add infrastructure for an option API
This patch adds the necessary basic infrastructure to support
centralized and unified option manipulation API for the bonding. The new
structure bond_option will be used to describe each option with its
dependencies on modes which will be checked automatically thus removing a
lot of duplicated code. Also automatic range checking is added for
some options. Currently the option setting function requires RTNL to
be acquired prior to calling it, since many options already rely on RTNL
it seemed like the best choice to protect all against common race
conditions.
In order to add an option the following steps need to be done:
1. Add an entry BOND_OPT_<option> to bond_options.h so it gets a unique id
and a bit corresponding to the id
2. Add a bond_option entry to the bond_opts[] array in bond_options.c which
describes the option, its dependencies and its manipulation function
3. Add code to export the option through sysfs and/or as a module parameter
(the sysfs export will be made automatically in the future)
The options can have different flags set, currently the following are
supported:
BOND_OPTFLAG_NOSLAVES - require that the bond device has no slaves prior
to setting the option
BOND_OPTFLAG_IFDOWN - require that the bond device is down prior to
setting the option
BOND_OPTFLAG_RAWVAL - don't parse the value but return it raw for the
option to parse
There's a new value structure to describe different types of values
which can have the following flags:
BOND_VALFLAG_DEFAULT - marks the default option (permanent string alias
to this option is "default")
BOND_VALFLAG_MIN - the minimum value that this option can have
BOND_VALFLAG_MAX - the maximum value that this option can have
An example would be nice here, so if we have an option which can have
the values "off"(2), "special"(4, default) and supports a range, say
16 - 32, it should be defined as follows:
"off", 2,
"special", 4, BOND_VALFLAG_DEFAULT,
"rangemin", 16, BOND_VALFLAG_MIN,
"rangemax", 32, BOND_VALFLAG_MAX
So we have the valid intervals: [2, 2], [4, 4], [16, 32]
Also the valid strings: "off" = 2, "special" and "default" = 4
"rangemin" = 16, "rangemax" = 32
BOND_VALFLAG_(MIN|MAX) can be used to specify a valid range for an
option, if MIN is omitted then 0 is considered as a minimum. If an
exact match is found in the values[] table it will be returned,
otherwise the range is tried (if available).
The option parameter passing is done by using a special structure called
bond_opt_value which can take either a string or a value to parse. One
of the bond_opt_init(val|str) macros should be used depending on which
one does the user want to parse (string or value). Then a call to
__bond_opt_set should be done under RTNL.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-01-22 13:53:16 +00:00
|
|
|
struct bond_params *params;
|
|
|
|
|
|
|
|
params = &bond->params;
|
2014-01-22 13:53:17 +00:00
|
|
|
modeval = bond_opt_get_val(BOND_OPT_MODE, params->mode);
|
bonding: add infrastructure for an option API
This patch adds the necessary basic infrastructure to support
centralized and unified option manipulation API for the bonding. The new
structure bond_option will be used to describe each option with its
dependencies on modes which will be checked automatically thus removing a
lot of duplicated code. Also automatic range checking is added for
some options. Currently the option setting function requires RTNL to
be acquired prior to calling it, since many options already rely on RTNL
it seemed like the best choice to protect all against common race
conditions.
In order to add an option the following steps need to be done:
1. Add an entry BOND_OPT_<option> to bond_options.h so it gets a unique id
and a bit corresponding to the id
2. Add a bond_option entry to the bond_opts[] array in bond_options.c which
describes the option, its dependencies and its manipulation function
3. Add code to export the option through sysfs and/or as a module parameter
(the sysfs export will be made automatically in the future)
The options can have different flags set, currently the following are
supported:
BOND_OPTFLAG_NOSLAVES - require that the bond device has no slaves prior
to setting the option
BOND_OPTFLAG_IFDOWN - require that the bond device is down prior to
setting the option
BOND_OPTFLAG_RAWVAL - don't parse the value but return it raw for the
option to parse
There's a new value structure to describe different types of values
which can have the following flags:
BOND_VALFLAG_DEFAULT - marks the default option (permanent string alias
to this option is "default")
BOND_VALFLAG_MIN - the minimum value that this option can have
BOND_VALFLAG_MAX - the maximum value that this option can have
An example would be nice here, so if we have an option which can have
the values "off"(2), "special"(4, default) and supports a range, say
16 - 32, it should be defined as follows:
"off", 2,
"special", 4, BOND_VALFLAG_DEFAULT,
"rangemin", 16, BOND_VALFLAG_MIN,
"rangemax", 32, BOND_VALFLAG_MAX
So we have the valid intervals: [2, 2], [4, 4], [16, 32]
Also the valid strings: "off" = 2, "special" and "default" = 4
"rangemin" = 16, "rangemax" = 32
BOND_VALFLAG_(MIN|MAX) can be used to specify a valid range for an
option, if MIN is omitted then 0 is considered as a minimum. If an
exact match is found in the values[] table it will be returned,
otherwise the range is tried (if available).
The option parameter passing is done by using a special structure called
bond_opt_value which can take either a string or a value to parse. One
of the bond_opt_init(val|str) macros should be used depending on which
one does the user want to parse (string or value). Then a call to
__bond_opt_set should be done under RTNL.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-01-22 13:53:16 +00:00
|
|
|
if (test_bit(params->mode, &opt->unsuppmodes))
|
2014-01-22 13:53:17 +00:00
|
|
|
pr_err("%s: option %s: mode dependency failed, not supported in mode %s(%llu)\n",
|
|
|
|
bond->dev->name, opt->name,
|
|
|
|
modeval->string, modeval->value);
|
bonding: add infrastructure for an option API
This patch adds the necessary basic infrastructure to support
centralized and unified option manipulation API for the bonding. The new
structure bond_option will be used to describe each option with its
dependencies on modes which will be checked automatically thus removing a
lot of duplicated code. Also automatic range checking is added for
some options. Currently the option setting function requires RTNL to
be acquired prior to calling it, since many options already rely on RTNL
it seemed like the best choice to protect all against common race
conditions.
In order to add an option the following steps need to be done:
1. Add an entry BOND_OPT_<option> to bond_options.h so it gets a unique id
and a bit corresponding to the id
2. Add a bond_option entry to the bond_opts[] array in bond_options.c which
describes the option, its dependencies and its manipulation function
3. Add code to export the option through sysfs and/or as a module parameter
(the sysfs export will be made automatically in the future)
The options can have different flags set, currently the following are
supported:
BOND_OPTFLAG_NOSLAVES - require that the bond device has no slaves prior
to setting the option
BOND_OPTFLAG_IFDOWN - require that the bond device is down prior to
setting the option
BOND_OPTFLAG_RAWVAL - don't parse the value but return it raw for the
option to parse
There's a new value structure to describe different types of values
which can have the following flags:
BOND_VALFLAG_DEFAULT - marks the default option (permanent string alias
to this option is "default")
BOND_VALFLAG_MIN - the minimum value that this option can have
BOND_VALFLAG_MAX - the maximum value that this option can have
An example would be nice here, so if we have an option which can have
the values "off"(2), "special"(4, default) and supports a range, say
16 - 32, it should be defined as follows:
"off", 2,
"special", 4, BOND_VALFLAG_DEFAULT,
"rangemin", 16, BOND_VALFLAG_MIN,
"rangemax", 32, BOND_VALFLAG_MAX
So we have the valid intervals: [2, 2], [4, 4], [16, 32]
Also the valid strings: "off" = 2, "special" and "default" = 4
"rangemin" = 16, "rangemax" = 32
BOND_VALFLAG_(MIN|MAX) can be used to specify a valid range for an
option, if MIN is omitted then 0 is considered as a minimum. If an
exact match is found in the values[] table it will be returned,
otherwise the range is tried (if available).
The option parameter passing is done by using a special structure called
bond_opt_value which can take either a string or a value to parse. One
of the bond_opt_init(val|str) macros should be used depending on which
one does the user want to parse (string or value). Then a call to
__bond_opt_set should be done under RTNL.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-01-22 13:53:16 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static void bond_opt_error_interpret(struct bonding *bond,
|
|
|
|
const struct bond_option *opt,
|
|
|
|
int error, struct bond_opt_value *val)
|
|
|
|
{
|
|
|
|
struct bond_opt_value *minval, *maxval;
|
|
|
|
char *p;
|
|
|
|
|
|
|
|
switch (error) {
|
|
|
|
case -EINVAL:
|
|
|
|
if (val) {
|
|
|
|
if (val->string) {
|
|
|
|
/* sometimes RAWVAL opts may have new lines */
|
|
|
|
p = strchr(val->string, '\n');
|
|
|
|
if (p)
|
|
|
|
*p = '\0';
|
|
|
|
pr_err("%s: option %s: invalid value (%s).\n",
|
|
|
|
bond->dev->name, opt->name, val->string);
|
|
|
|
} else {
|
|
|
|
pr_err("%s: option %s: invalid value (%llu).\n",
|
|
|
|
bond->dev->name, opt->name, val->value);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
minval = bond_opt_get_flags(opt, BOND_VALFLAG_MIN);
|
|
|
|
maxval = bond_opt_get_flags(opt, BOND_VALFLAG_MAX);
|
|
|
|
if (!maxval)
|
|
|
|
break;
|
|
|
|
pr_err("%s: option %s: allowed values %llu - %llu.\n",
|
|
|
|
bond->dev->name, opt->name, minval ? minval->value : 0,
|
|
|
|
maxval->value);
|
|
|
|
break;
|
|
|
|
case -EACCES:
|
|
|
|
bond_opt_dep_print(bond, opt);
|
|
|
|
break;
|
|
|
|
case -ENOTEMPTY:
|
|
|
|
pr_err("%s: option %s: unable to set because the bond device has slaves.\n",
|
|
|
|
bond->dev->name, opt->name);
|
|
|
|
break;
|
|
|
|
case -EBUSY:
|
|
|
|
pr_err("%s: option %s: unable to set because the bond device is up.\n",
|
|
|
|
bond->dev->name, opt->name);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* __bond_opt_set - set a bonding option
|
|
|
|
* @bond: target bond device
|
|
|
|
* @option: option to set
|
|
|
|
* @val: value to set it to
|
|
|
|
*
|
|
|
|
* This function is used to change the bond's option value, it can be
|
|
|
|
* used for both enabling/changing an option and for disabling it. RTNL lock
|
|
|
|
* must be obtained before calling this function.
|
|
|
|
*/
|
|
|
|
int __bond_opt_set(struct bonding *bond,
|
|
|
|
unsigned int option, struct bond_opt_value *val)
|
|
|
|
{
|
|
|
|
struct bond_opt_value *retval = NULL;
|
|
|
|
const struct bond_option *opt;
|
|
|
|
int ret = -ENOENT;
|
|
|
|
|
|
|
|
ASSERT_RTNL();
|
|
|
|
|
|
|
|
opt = bond_opt_get(option);
|
|
|
|
if (WARN_ON(!val) || WARN_ON(!opt))
|
|
|
|
goto out;
|
|
|
|
ret = bond_opt_check_deps(bond, opt);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
retval = bond_opt_parse(opt, val);
|
|
|
|
if (!retval) {
|
|
|
|
ret = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
ret = opt->set(bond, retval);
|
|
|
|
out:
|
|
|
|
if (ret)
|
|
|
|
bond_opt_error_interpret(bond, opt, ret, val);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* bond_opt_tryset_rtnl - try to acquire rtnl and call __bond_opt_set
|
|
|
|
* @bond: target bond device
|
|
|
|
* @option: option to set
|
|
|
|
* @buf: value to set it to
|
|
|
|
*
|
|
|
|
* This function tries to acquire RTNL without blocking and if successful
|
|
|
|
* calls __bond_opt_set. It is mainly used for sysfs option manipulation.
|
|
|
|
*/
|
|
|
|
int bond_opt_tryset_rtnl(struct bonding *bond, unsigned int option, char *buf)
|
|
|
|
{
|
|
|
|
struct bond_opt_value optval;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (!rtnl_trylock())
|
|
|
|
return restart_syscall();
|
|
|
|
bond_opt_initstr(&optval, buf);
|
|
|
|
ret = __bond_opt_set(bond, option, &optval);
|
|
|
|
rtnl_unlock();
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* bond_opt_get - get a pointer to an option
|
|
|
|
* @option: option for which to return a pointer
|
|
|
|
*
|
|
|
|
* This function checks if option is valid and if so returns a pointer
|
|
|
|
* to its entry in the bond_opts[] option array.
|
|
|
|
*/
|
|
|
|
struct bond_option *bond_opt_get(unsigned int option)
|
|
|
|
{
|
|
|
|
if (!BOND_OPT_VALID(option))
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
return &bond_opts[option];
|
|
|
|
}
|
|
|
|
|
2014-01-22 13:53:17 +00:00
|
|
|
int bond_option_mode_set(struct bonding *bond, struct bond_opt_value *newval)
|
2013-10-18 15:43:34 +00:00
|
|
|
{
|
2014-01-22 13:53:17 +00:00
|
|
|
if (BOND_NO_USES_ARP(newval->value) && bond->params.arp_interval) {
|
bonding: disable arp and enable mii monitoring when bond change to no uses arp mode
Because the ARP monitoring is not support for 802.3ad, but I still
could change the mode to 802.3ad from ab mode while ARP monitoring
is running, it is incorrect.
So add a check for 802.3ad in bonding_store_mode to fix the problem,
and make a new macro BOND_NO_USES_ARP() to simplify the code.
v2: according to the Dan Williams's suggestion, bond mode is the most
important bond option, it should override any of the other sub-options.
So when the mode is changed, the conficting values should be cleared
or reset, otherwise the user has to duplicate more operations to modify
the logic. I disable the arp and enable mii monitoring when the bond mode
is changed to AB, TB and 8023AD if the arp interval is true.
v3: according to the Nik's suggestion, the default value of miimon should need
a name, there is several place to use it, and the bond_store_arp_interval()
could use micro BOND_NO_USES_ARP to make the code more simpify.
Suggested-by: Dan Williams <dcbw@redhat.com>
Suggested-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: Ding Tianhong <dingtianhong@huawei.com>
Reviewed-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-11-22 14:28:43 +00:00
|
|
|
pr_info("%s: %s mode is incompatible with arp monitoring, start mii monitoring\n",
|
2014-01-22 13:53:17 +00:00
|
|
|
bond->dev->name, newval->string);
|
bonding: disable arp and enable mii monitoring when bond change to no uses arp mode
Because the ARP monitoring is not support for 802.3ad, but I still
could change the mode to 802.3ad from ab mode while ARP monitoring
is running, it is incorrect.
So add a check for 802.3ad in bonding_store_mode to fix the problem,
and make a new macro BOND_NO_USES_ARP() to simplify the code.
v2: according to the Dan Williams's suggestion, bond mode is the most
important bond option, it should override any of the other sub-options.
So when the mode is changed, the conficting values should be cleared
or reset, otherwise the user has to duplicate more operations to modify
the logic. I disable the arp and enable mii monitoring when the bond mode
is changed to AB, TB and 8023AD if the arp interval is true.
v3: according to the Nik's suggestion, the default value of miimon should need
a name, there is several place to use it, and the bond_store_arp_interval()
could use micro BOND_NO_USES_ARP to make the code more simpify.
Suggested-by: Dan Williams <dcbw@redhat.com>
Suggested-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: Ding Tianhong <dingtianhong@huawei.com>
Reviewed-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-11-22 14:28:43 +00:00
|
|
|
/* disable arp monitoring */
|
|
|
|
bond->params.arp_interval = 0;
|
|
|
|
/* set miimon to default value */
|
|
|
|
bond->params.miimon = BOND_DEFAULT_MIIMON;
|
|
|
|
pr_info("%s: Setting MII monitoring interval to %d.\n",
|
|
|
|
bond->dev->name, bond->params.miimon);
|
2013-10-18 15:43:34 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* don't cache arp_validate between modes */
|
|
|
|
bond->params.arp_validate = BOND_ARP_VALIDATE_NONE;
|
2014-01-22 13:53:17 +00:00
|
|
|
bond->params.mode = newval->value;
|
|
|
|
|
2013-10-18 15:43:34 +00:00
|
|
|
return 0;
|
|
|
|
}
|
2013-10-18 15:43:35 +00:00
|
|
|
|
2013-10-18 15:43:37 +00:00
|
|
|
static struct net_device *__bond_option_active_slave_get(struct bonding *bond,
|
|
|
|
struct slave *slave)
|
|
|
|
{
|
|
|
|
return USES_PRIMARY(bond->params.mode) && slave ? slave->dev : NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct net_device *bond_option_active_slave_get_rcu(struct bonding *bond)
|
|
|
|
{
|
|
|
|
struct slave *slave = rcu_dereference(bond->curr_active_slave);
|
|
|
|
|
|
|
|
return __bond_option_active_slave_get(bond, slave);
|
|
|
|
}
|
|
|
|
|
|
|
|
struct net_device *bond_option_active_slave_get(struct bonding *bond)
|
|
|
|
{
|
|
|
|
return __bond_option_active_slave_get(bond, bond->curr_active_slave);
|
|
|
|
}
|
|
|
|
|
2013-10-18 15:43:35 +00:00
|
|
|
int bond_option_active_slave_set(struct bonding *bond,
|
|
|
|
struct net_device *slave_dev)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
if (slave_dev) {
|
|
|
|
if (!netif_is_bond_slave(slave_dev)) {
|
|
|
|
pr_err("Device %s is not bonding slave.\n",
|
|
|
|
slave_dev->name);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (bond->dev != netdev_master_upper_dev_get(slave_dev)) {
|
|
|
|
pr_err("%s: Device %s is not our slave.\n",
|
|
|
|
bond->dev->name, slave_dev->name);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!USES_PRIMARY(bond->params.mode)) {
|
|
|
|
pr_err("%s: Unable to change active slave; %s is in mode %d\n",
|
|
|
|
bond->dev->name, bond->dev->name, bond->params.mode);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
block_netpoll_tx();
|
|
|
|
write_lock_bh(&bond->curr_slave_lock);
|
|
|
|
|
|
|
|
/* check to see if we are clearing active */
|
|
|
|
if (!slave_dev) {
|
|
|
|
pr_info("%s: Clearing current active slave.\n",
|
|
|
|
bond->dev->name);
|
|
|
|
rcu_assign_pointer(bond->curr_active_slave, NULL);
|
|
|
|
bond_select_active_slave(bond);
|
|
|
|
} else {
|
|
|
|
struct slave *old_active = bond->curr_active_slave;
|
|
|
|
struct slave *new_active = bond_slave_get_rtnl(slave_dev);
|
|
|
|
|
|
|
|
BUG_ON(!new_active);
|
|
|
|
|
|
|
|
if (new_active == old_active) {
|
|
|
|
/* do nothing */
|
|
|
|
pr_info("%s: %s is already the current active slave.\n",
|
|
|
|
bond->dev->name, new_active->dev->name);
|
|
|
|
} else {
|
|
|
|
if (old_active && (new_active->link == BOND_LINK_UP) &&
|
|
|
|
IS_UP(new_active->dev)) {
|
|
|
|
pr_info("%s: Setting %s as active slave.\n",
|
|
|
|
bond->dev->name, new_active->dev->name);
|
|
|
|
bond_change_active_slave(bond, new_active);
|
|
|
|
} else {
|
|
|
|
pr_err("%s: Could not set %s as active slave; either %s is down or the link is down.\n",
|
|
|
|
bond->dev->name, new_active->dev->name,
|
|
|
|
new_active->dev->name);
|
|
|
|
ret = -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
write_unlock_bh(&bond->curr_slave_lock);
|
|
|
|
unblock_netpoll_tx();
|
|
|
|
return ret;
|
|
|
|
}
|
2013-12-12 22:09:55 +00:00
|
|
|
|
|
|
|
int bond_option_miimon_set(struct bonding *bond, int miimon)
|
|
|
|
{
|
|
|
|
if (miimon < 0) {
|
|
|
|
pr_err("%s: Invalid miimon value %d not in range %d-%d; rejected.\n",
|
|
|
|
bond->dev->name, miimon, 0, INT_MAX);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
pr_info("%s: Setting MII monitoring interval to %d.\n",
|
|
|
|
bond->dev->name, miimon);
|
|
|
|
bond->params.miimon = miimon;
|
|
|
|
if (bond->params.updelay)
|
|
|
|
pr_info("%s: Note: Updating updelay (to %d) since it is a multiple of the miimon value.\n",
|
|
|
|
bond->dev->name,
|
|
|
|
bond->params.updelay * bond->params.miimon);
|
|
|
|
if (bond->params.downdelay)
|
|
|
|
pr_info("%s: Note: Updating downdelay (to %d) since it is a multiple of the miimon value.\n",
|
|
|
|
bond->dev->name,
|
|
|
|
bond->params.downdelay * bond->params.miimon);
|
|
|
|
if (miimon && bond->params.arp_interval) {
|
|
|
|
pr_info("%s: MII monitoring cannot be used with ARP monitoring. Disabling ARP monitoring...\n",
|
|
|
|
bond->dev->name);
|
|
|
|
bond->params.arp_interval = 0;
|
|
|
|
if (bond->params.arp_validate)
|
|
|
|
bond->params.arp_validate = BOND_ARP_VALIDATE_NONE;
|
|
|
|
}
|
|
|
|
if (bond->dev->flags & IFF_UP) {
|
|
|
|
/* If the interface is up, we may need to fire off
|
|
|
|
* the MII timer. If the interface is down, the
|
|
|
|
* timer will get fired off when the open function
|
|
|
|
* is called.
|
|
|
|
*/
|
|
|
|
if (!miimon) {
|
|
|
|
cancel_delayed_work_sync(&bond->mii_work);
|
|
|
|
} else {
|
|
|
|
cancel_delayed_work_sync(&bond->arp_work);
|
|
|
|
queue_delayed_work(bond->wq, &bond->mii_work, 0);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-12 22:10:02 +00:00
|
|
|
|
|
|
|
int bond_option_updelay_set(struct bonding *bond, int updelay)
|
|
|
|
{
|
|
|
|
if (!(bond->params.miimon)) {
|
|
|
|
pr_err("%s: Unable to set up delay as MII monitoring is disabled\n",
|
|
|
|
bond->dev->name);
|
|
|
|
return -EPERM;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (updelay < 0) {
|
|
|
|
pr_err("%s: Invalid up delay value %d not in range %d-%d; rejected.\n",
|
|
|
|
bond->dev->name, updelay, 0, INT_MAX);
|
|
|
|
return -EINVAL;
|
|
|
|
} else {
|
|
|
|
if ((updelay % bond->params.miimon) != 0) {
|
|
|
|
pr_warn("%s: Warning: up delay (%d) is not a multiple of miimon (%d), updelay rounded to %d ms\n",
|
|
|
|
bond->dev->name, updelay,
|
|
|
|
bond->params.miimon,
|
|
|
|
(updelay / bond->params.miimon) *
|
|
|
|
bond->params.miimon);
|
|
|
|
}
|
|
|
|
bond->params.updelay = updelay / bond->params.miimon;
|
|
|
|
pr_info("%s: Setting up delay to %d.\n",
|
|
|
|
bond->dev->name,
|
|
|
|
bond->params.updelay * bond->params.miimon);
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-12 22:10:09 +00:00
|
|
|
|
|
|
|
int bond_option_downdelay_set(struct bonding *bond, int downdelay)
|
|
|
|
{
|
|
|
|
if (!(bond->params.miimon)) {
|
|
|
|
pr_err("%s: Unable to set down delay as MII monitoring is disabled\n",
|
|
|
|
bond->dev->name);
|
|
|
|
return -EPERM;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (downdelay < 0) {
|
|
|
|
pr_err("%s: Invalid down delay value %d not in range %d-%d; rejected.\n",
|
|
|
|
bond->dev->name, downdelay, 0, INT_MAX);
|
|
|
|
return -EINVAL;
|
|
|
|
} else {
|
|
|
|
if ((downdelay % bond->params.miimon) != 0) {
|
|
|
|
pr_warn("%s: Warning: down delay (%d) is not a multiple of miimon (%d), delay rounded to %d ms\n",
|
|
|
|
bond->dev->name, downdelay,
|
|
|
|
bond->params.miimon,
|
|
|
|
(downdelay / bond->params.miimon) *
|
|
|
|
bond->params.miimon);
|
|
|
|
}
|
|
|
|
bond->params.downdelay = downdelay / bond->params.miimon;
|
|
|
|
pr_info("%s: Setting down delay to %d.\n",
|
|
|
|
bond->dev->name,
|
|
|
|
bond->params.downdelay * bond->params.miimon);
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-12 22:10:16 +00:00
|
|
|
|
|
|
|
int bond_option_use_carrier_set(struct bonding *bond, int use_carrier)
|
|
|
|
{
|
|
|
|
if ((use_carrier == 0) || (use_carrier == 1)) {
|
|
|
|
bond->params.use_carrier = use_carrier;
|
|
|
|
pr_info("%s: Setting use_carrier to %d.\n",
|
|
|
|
bond->dev->name, use_carrier);
|
|
|
|
} else {
|
|
|
|
pr_info("%s: Ignoring invalid use_carrier value %d.\n",
|
|
|
|
bond->dev->name, use_carrier);
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-12 22:10:24 +00:00
|
|
|
|
2014-01-22 13:53:23 +00:00
|
|
|
int bond_option_arp_interval_set(struct bonding *bond,
|
|
|
|
struct bond_opt_value *newval)
|
2013-12-12 22:10:24 +00:00
|
|
|
{
|
2014-01-22 13:53:23 +00:00
|
|
|
pr_info("%s: Setting ARP monitoring interval to %llu.\n",
|
|
|
|
bond->dev->name, newval->value);
|
|
|
|
bond->params.arp_interval = newval->value;
|
|
|
|
if (newval->value) {
|
2013-12-12 22:10:24 +00:00
|
|
|
if (bond->params.miimon) {
|
|
|
|
pr_info("%s: ARP monitoring cannot be used with MII monitoring. %s Disabling MII monitoring.\n",
|
|
|
|
bond->dev->name, bond->dev->name);
|
|
|
|
bond->params.miimon = 0;
|
|
|
|
}
|
|
|
|
if (!bond->params.arp_targets[0])
|
|
|
|
pr_info("%s: ARP monitoring has been set up, but no ARP targets have been specified.\n",
|
|
|
|
bond->dev->name);
|
|
|
|
}
|
|
|
|
if (bond->dev->flags & IFF_UP) {
|
|
|
|
/* If the interface is up, we may need to fire off
|
|
|
|
* the ARP timer. If the interface is down, the
|
|
|
|
* timer will get fired off when the open function
|
|
|
|
* is called.
|
|
|
|
*/
|
2014-01-22 13:53:23 +00:00
|
|
|
if (!newval->value) {
|
2013-12-12 22:10:24 +00:00
|
|
|
if (bond->params.arp_validate)
|
|
|
|
bond->recv_probe = NULL;
|
|
|
|
cancel_delayed_work_sync(&bond->arp_work);
|
|
|
|
} else {
|
|
|
|
/* arp_validate can be set only in active-backup mode */
|
|
|
|
if (bond->params.arp_validate)
|
|
|
|
bond->recv_probe = bond_arp_rcv;
|
|
|
|
cancel_delayed_work_sync(&bond->mii_work);
|
|
|
|
queue_delayed_work(bond->wq, &bond->arp_work, 0);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-12 22:10:31 +00:00
|
|
|
|
|
|
|
static void _bond_options_arp_ip_target_set(struct bonding *bond, int slot,
|
|
|
|
__be32 target,
|
|
|
|
unsigned long last_rx)
|
|
|
|
{
|
|
|
|
__be32 *targets = bond->params.arp_targets;
|
|
|
|
struct list_head *iter;
|
|
|
|
struct slave *slave;
|
|
|
|
|
|
|
|
if (slot >= 0 && slot < BOND_MAX_ARP_TARGETS) {
|
|
|
|
bond_for_each_slave(bond, slave, iter)
|
|
|
|
slave->target_last_arp_rx[slot] = last_rx;
|
|
|
|
targets[slot] = target;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static int _bond_option_arp_ip_target_add(struct bonding *bond, __be32 target)
|
|
|
|
{
|
|
|
|
__be32 *targets = bond->params.arp_targets;
|
|
|
|
int ind;
|
|
|
|
|
|
|
|
if (IS_IP_TARGET_UNUSABLE_ADDRESS(target)) {
|
|
|
|
pr_err("%s: invalid ARP target %pI4 specified for addition\n",
|
|
|
|
bond->dev->name, &target);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (bond_get_targets_ip(targets, target) != -1) { /* dup */
|
|
|
|
pr_err("%s: ARP target %pI4 is already present\n",
|
|
|
|
bond->dev->name, &target);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
ind = bond_get_targets_ip(targets, 0); /* first free slot */
|
|
|
|
if (ind == -1) {
|
|
|
|
pr_err("%s: ARP target table is full!\n",
|
|
|
|
bond->dev->name);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
pr_info("%s: adding ARP target %pI4.\n", bond->dev->name, &target);
|
|
|
|
|
|
|
|
_bond_options_arp_ip_target_set(bond, ind, target, jiffies);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int bond_option_arp_ip_target_add(struct bonding *bond, __be32 target)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* not to race with bond_arp_rcv */
|
|
|
|
write_lock_bh(&bond->lock);
|
|
|
|
ret = _bond_option_arp_ip_target_add(bond, target);
|
|
|
|
write_unlock_bh(&bond->lock);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
int bond_option_arp_ip_target_rem(struct bonding *bond, __be32 target)
|
|
|
|
{
|
|
|
|
__be32 *targets = bond->params.arp_targets;
|
|
|
|
struct list_head *iter;
|
|
|
|
struct slave *slave;
|
|
|
|
unsigned long *targets_rx;
|
|
|
|
int ind, i;
|
|
|
|
|
|
|
|
if (IS_IP_TARGET_UNUSABLE_ADDRESS(target)) {
|
|
|
|
pr_err("%s: invalid ARP target %pI4 specified for removal\n",
|
|
|
|
bond->dev->name, &target);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
ind = bond_get_targets_ip(targets, target);
|
|
|
|
if (ind == -1) {
|
|
|
|
pr_err("%s: unable to remove nonexistent ARP target %pI4.\n",
|
|
|
|
bond->dev->name, &target);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (ind == 0 && !targets[1] && bond->params.arp_interval)
|
|
|
|
pr_warn("%s: removing last arp target with arp_interval on\n",
|
|
|
|
bond->dev->name);
|
|
|
|
|
|
|
|
pr_info("%s: removing ARP target %pI4.\n", bond->dev->name,
|
|
|
|
&target);
|
|
|
|
|
|
|
|
/* not to race with bond_arp_rcv */
|
|
|
|
write_lock_bh(&bond->lock);
|
|
|
|
|
|
|
|
bond_for_each_slave(bond, slave, iter) {
|
|
|
|
targets_rx = slave->target_last_arp_rx;
|
|
|
|
for (i = ind; (i < BOND_MAX_ARP_TARGETS-1) && targets[i+1]; i++)
|
|
|
|
targets_rx[i] = targets_rx[i+1];
|
|
|
|
targets_rx[i] = 0;
|
|
|
|
}
|
|
|
|
for (i = ind; (i < BOND_MAX_ARP_TARGETS-1) && targets[i+1]; i++)
|
|
|
|
targets[i] = targets[i+1];
|
|
|
|
targets[i] = 0;
|
|
|
|
|
|
|
|
write_unlock_bh(&bond->lock);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int bond_option_arp_ip_targets_set(struct bonding *bond, __be32 *targets,
|
|
|
|
int count)
|
|
|
|
{
|
|
|
|
int i, ret = 0;
|
|
|
|
|
|
|
|
/* not to race with bond_arp_rcv */
|
|
|
|
write_lock_bh(&bond->lock);
|
|
|
|
|
|
|
|
/* clear table */
|
|
|
|
for (i = 0; i < BOND_MAX_ARP_TARGETS; i++)
|
|
|
|
_bond_options_arp_ip_target_set(bond, i, 0, 0);
|
|
|
|
|
|
|
|
if (count == 0 && bond->params.arp_interval)
|
|
|
|
pr_warn("%s: removing last arp target with arp_interval on\n",
|
|
|
|
bond->dev->name);
|
|
|
|
|
|
|
|
for (i = 0; i < count; i++) {
|
|
|
|
ret = _bond_option_arp_ip_target_add(bond, targets[i]);
|
|
|
|
if (ret)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
write_unlock_bh(&bond->lock);
|
|
|
|
return ret;
|
|
|
|
}
|
2013-12-12 22:10:38 +00:00
|
|
|
|
2014-01-22 13:53:20 +00:00
|
|
|
int bond_option_arp_validate_set(struct bonding *bond,
|
|
|
|
struct bond_opt_value *newval)
|
2013-12-12 22:10:38 +00:00
|
|
|
{
|
2014-01-22 13:53:20 +00:00
|
|
|
pr_info("%s: setting arp_validate to %s (%llu).\n",
|
|
|
|
bond->dev->name, newval->string, newval->value);
|
2013-12-12 22:10:38 +00:00
|
|
|
|
|
|
|
if (bond->dev->flags & IFF_UP) {
|
2014-01-22 13:53:20 +00:00
|
|
|
if (!newval->value)
|
2013-12-12 22:10:38 +00:00
|
|
|
bond->recv_probe = NULL;
|
|
|
|
else if (bond->params.arp_interval)
|
|
|
|
bond->recv_probe = bond_arp_rcv;
|
|
|
|
}
|
2014-01-22 13:53:20 +00:00
|
|
|
bond->params.arp_validate = newval->value;
|
2013-12-12 22:10:38 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-12 22:10:45 +00:00
|
|
|
|
2014-01-22 13:53:21 +00:00
|
|
|
int bond_option_arp_all_targets_set(struct bonding *bond,
|
|
|
|
struct bond_opt_value *newval)
|
2013-12-12 22:10:45 +00:00
|
|
|
{
|
2014-01-22 13:53:21 +00:00
|
|
|
pr_info("%s: setting arp_all_targets to %s (%llu).\n",
|
|
|
|
bond->dev->name, newval->string, newval->value);
|
|
|
|
bond->params.arp_all_targets = newval->value;
|
2013-12-12 22:10:45 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-16 00:41:51 +00:00
|
|
|
|
|
|
|
int bond_option_primary_set(struct bonding *bond, const char *primary)
|
|
|
|
{
|
|
|
|
struct list_head *iter;
|
|
|
|
struct slave *slave;
|
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
block_netpoll_tx();
|
|
|
|
read_lock(&bond->lock);
|
|
|
|
write_lock_bh(&bond->curr_slave_lock);
|
|
|
|
|
|
|
|
if (!USES_PRIMARY(bond->params.mode)) {
|
|
|
|
pr_err("%s: Unable to set primary slave; %s is in mode %d\n",
|
|
|
|
bond->dev->name, bond->dev->name, bond->params.mode);
|
|
|
|
err = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* check to see if we are clearing primary */
|
|
|
|
if (!strlen(primary)) {
|
|
|
|
pr_info("%s: Setting primary slave to None.\n",
|
|
|
|
bond->dev->name);
|
|
|
|
bond->primary_slave = NULL;
|
|
|
|
memset(bond->params.primary, 0, sizeof(bond->params.primary));
|
|
|
|
bond_select_active_slave(bond);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
bond_for_each_slave(bond, slave, iter) {
|
|
|
|
if (strncmp(slave->dev->name, primary, IFNAMSIZ) == 0) {
|
|
|
|
pr_info("%s: Setting %s as primary slave.\n",
|
|
|
|
bond->dev->name, slave->dev->name);
|
|
|
|
bond->primary_slave = slave;
|
|
|
|
strcpy(bond->params.primary, slave->dev->name);
|
|
|
|
bond_select_active_slave(bond);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
strncpy(bond->params.primary, primary, IFNAMSIZ);
|
|
|
|
bond->params.primary[IFNAMSIZ - 1] = 0;
|
|
|
|
|
|
|
|
pr_info("%s: Recording %s as primary, but it has not been enslaved to %s yet.\n",
|
|
|
|
bond->dev->name, primary, bond->dev->name);
|
|
|
|
|
|
|
|
out:
|
|
|
|
write_unlock_bh(&bond->curr_slave_lock);
|
|
|
|
read_unlock(&bond->lock);
|
|
|
|
unblock_netpoll_tx();
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
2013-12-16 00:41:58 +00:00
|
|
|
|
|
|
|
int bond_option_primary_reselect_set(struct bonding *bond, int primary_reselect)
|
|
|
|
{
|
2014-01-03 22:28:18 +00:00
|
|
|
if (bond_parm_tbl_lookup(primary_reselect, pri_reselect_tbl) < 0) {
|
|
|
|
pr_err("%s: Ignoring invalid primary_reselect value %d.\n",
|
|
|
|
bond->dev->name, primary_reselect);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2013-12-16 00:41:58 +00:00
|
|
|
bond->params.primary_reselect = primary_reselect;
|
|
|
|
pr_info("%s: setting primary_reselect to %s (%d).\n",
|
|
|
|
bond->dev->name, pri_reselect_tbl[primary_reselect].modename,
|
|
|
|
primary_reselect);
|
|
|
|
|
|
|
|
block_netpoll_tx();
|
|
|
|
write_lock_bh(&bond->curr_slave_lock);
|
|
|
|
bond_select_active_slave(bond);
|
|
|
|
write_unlock_bh(&bond->curr_slave_lock);
|
|
|
|
unblock_netpoll_tx();
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-16 00:42:05 +00:00
|
|
|
|
2014-01-22 13:53:22 +00:00
|
|
|
int bond_option_fail_over_mac_set(struct bonding *bond,
|
|
|
|
struct bond_opt_value *newval)
|
2013-12-16 00:42:05 +00:00
|
|
|
{
|
2014-01-22 13:53:22 +00:00
|
|
|
pr_info("%s: Setting fail_over_mac to %s (%llu).\n",
|
|
|
|
bond->dev->name, newval->string, newval->value);
|
|
|
|
bond->params.fail_over_mac = newval->value;
|
2013-12-16 00:42:05 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-16 00:42:12 +00:00
|
|
|
|
2014-01-22 13:53:19 +00:00
|
|
|
int bond_option_xmit_hash_policy_set(struct bonding *bond,
|
|
|
|
struct bond_opt_value *newval)
|
2013-12-16 00:42:12 +00:00
|
|
|
{
|
2014-01-22 13:53:19 +00:00
|
|
|
pr_info("%s: setting xmit hash policy to %s (%llu).\n",
|
|
|
|
bond->dev->name, newval->string, newval->value);
|
|
|
|
bond->params.xmit_policy = newval->value;
|
2013-12-16 00:42:12 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-16 00:42:19 +00:00
|
|
|
|
|
|
|
int bond_option_resend_igmp_set(struct bonding *bond, int resend_igmp)
|
|
|
|
{
|
|
|
|
if (resend_igmp < 0 || resend_igmp > 255) {
|
|
|
|
pr_err("%s: Invalid resend_igmp value %d not in range 0-255; rejected.\n",
|
|
|
|
bond->dev->name, resend_igmp);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
bond->params.resend_igmp = resend_igmp;
|
|
|
|
pr_info("%s: Setting resend_igmp to %d.\n",
|
|
|
|
bond->dev->name, resend_igmp);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-18 05:30:09 +00:00
|
|
|
|
|
|
|
int bond_option_num_peer_notif_set(struct bonding *bond, int num_peer_notif)
|
|
|
|
{
|
|
|
|
bond->params.num_peer_notif = num_peer_notif;
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-18 05:30:16 +00:00
|
|
|
|
|
|
|
int bond_option_all_slaves_active_set(struct bonding *bond,
|
|
|
|
int all_slaves_active)
|
|
|
|
{
|
|
|
|
struct list_head *iter;
|
|
|
|
struct slave *slave;
|
|
|
|
|
|
|
|
if (all_slaves_active == bond->params.all_slaves_active)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if ((all_slaves_active == 0) || (all_slaves_active == 1)) {
|
|
|
|
bond->params.all_slaves_active = all_slaves_active;
|
|
|
|
} else {
|
|
|
|
pr_info("%s: Ignoring invalid all_slaves_active value %d.\n",
|
|
|
|
bond->dev->name, all_slaves_active);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
bond_for_each_slave(bond, slave, iter) {
|
|
|
|
if (!bond_is_active_slave(slave)) {
|
|
|
|
if (all_slaves_active)
|
|
|
|
slave->inactive = 0;
|
|
|
|
else
|
|
|
|
slave->inactive = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-18 05:30:23 +00:00
|
|
|
|
|
|
|
int bond_option_min_links_set(struct bonding *bond, int min_links)
|
|
|
|
{
|
|
|
|
pr_info("%s: Setting min links value to %u\n",
|
|
|
|
bond->dev->name, min_links);
|
|
|
|
bond->params.min_links = min_links;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-18 05:30:30 +00:00
|
|
|
|
|
|
|
int bond_option_lp_interval_set(struct bonding *bond, int lp_interval)
|
|
|
|
{
|
|
|
|
if (lp_interval <= 0) {
|
|
|
|
pr_err("%s: lp_interval must be between 1 and %d\n",
|
|
|
|
bond->dev->name, INT_MAX);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
bond->params.lp_interval = lp_interval;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2013-12-18 05:30:37 +00:00
|
|
|
|
2014-01-22 13:53:18 +00:00
|
|
|
int bond_option_pps_set(struct bonding *bond, struct bond_opt_value *newval)
|
2013-12-18 05:30:37 +00:00
|
|
|
{
|
2014-01-22 13:53:18 +00:00
|
|
|
bond->params.packets_per_slave = newval->value;
|
|
|
|
if (newval->value > 0) {
|
reciprocal_divide: update/correction of the algorithm
Jakub Zawadzki noticed that some divisions by reciprocal_divide()
were not correct [1][2], which he could also show with BPF code
after divisions are transformed into reciprocal_value() for runtime
invariance which can be passed to reciprocal_divide() later on;
reverse in BPF dump ended up with a different, off-by-one K in
some situations.
This has been fixed by Eric Dumazet in commit aee636c4809fa5
("bpf: do not use reciprocal divide"). This follow-up patch
improves reciprocal_value() and reciprocal_divide() to work in
all cases by using Granlund and Montgomery method, so that also
future use is safe and without any non-obvious side-effects.
Known problems with the old implementation were that division by 1
always returned 0 and some off-by-ones when the dividend and divisor
where very large. This seemed to not be problematic with its
current users, as far as we can tell. Eric Dumazet checked for
the slab usage, we cannot surely say so in the case of flex_array.
Still, in order to fix that, we propose an extension from the
original implementation from commit 6a2d7a955d8d resp. [3][4],
by using the algorithm proposed in "Division by Invariant Integers
Using Multiplication" [5], Torbjörn Granlund and Peter L.
Montgomery, that is, pseudocode for q = n/d where q, n, d is in
u32 universe:
1) Initialization:
int l = ceil(log_2 d)
uword m' = floor((1<<32)*((1<<l)-d)/d)+1
int sh_1 = min(l,1)
int sh_2 = max(l-1,0)
2) For q = n/d, all uword:
uword t = (n*m')>>32
q = (t+((n-t)>>sh_1))>>sh_2
The assembler implementation from Agner Fog [6] also helped a lot
while implementing. We have tested the implementation on x86_64,
ppc64, i686, s390x; on x86_64/haswell we're still half the latency
compared to normal divide.
Joint work with Daniel Borkmann.
[1] http://www.wireshark.org/~darkjames/reciprocal-buggy.c
[2] http://www.wireshark.org/~darkjames/set-and-dump-filter-k-bug.c
[3] https://gmplib.org/~tege/division-paper.pdf
[4] http://homepage.cs.uiowa.edu/~jones/bcd/divide.html
[5] http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1.2556
[6] http://www.agner.org/optimize/asmlib.zip
Reported-by: Jakub Zawadzki <darkjames-ws@darkjames.pl>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Austin S Hemmelgarn <ahferroin7@gmail.com>
Cc: linux-kernel@vger.kernel.org
Cc: Jesse Gross <jesse@nicira.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Matt Mackall <mpm@selenic.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Andy Gospodarek <andy@greyhouse.net>
Cc: Veaceslav Falico <vfalico@redhat.com>
Cc: Jay Vosburgh <fubar@us.ibm.com>
Cc: Jakub Zawadzki <darkjames-ws@darkjames.pl>
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-01-22 01:29:41 +00:00
|
|
|
bond->params.reciprocal_packets_per_slave =
|
2014-01-22 13:53:18 +00:00
|
|
|
reciprocal_value(newval->value);
|
reciprocal_divide: update/correction of the algorithm
Jakub Zawadzki noticed that some divisions by reciprocal_divide()
were not correct [1][2], which he could also show with BPF code
after divisions are transformed into reciprocal_value() for runtime
invariance which can be passed to reciprocal_divide() later on;
reverse in BPF dump ended up with a different, off-by-one K in
some situations.
This has been fixed by Eric Dumazet in commit aee636c4809fa5
("bpf: do not use reciprocal divide"). This follow-up patch
improves reciprocal_value() and reciprocal_divide() to work in
all cases by using Granlund and Montgomery method, so that also
future use is safe and without any non-obvious side-effects.
Known problems with the old implementation were that division by 1
always returned 0 and some off-by-ones when the dividend and divisor
where very large. This seemed to not be problematic with its
current users, as far as we can tell. Eric Dumazet checked for
the slab usage, we cannot surely say so in the case of flex_array.
Still, in order to fix that, we propose an extension from the
original implementation from commit 6a2d7a955d8d resp. [3][4],
by using the algorithm proposed in "Division by Invariant Integers
Using Multiplication" [5], Torbjörn Granlund and Peter L.
Montgomery, that is, pseudocode for q = n/d where q, n, d is in
u32 universe:
1) Initialization:
int l = ceil(log_2 d)
uword m' = floor((1<<32)*((1<<l)-d)/d)+1
int sh_1 = min(l,1)
int sh_2 = max(l-1,0)
2) For q = n/d, all uword:
uword t = (n*m')>>32
q = (t+((n-t)>>sh_1))>>sh_2
The assembler implementation from Agner Fog [6] also helped a lot
while implementing. We have tested the implementation on x86_64,
ppc64, i686, s390x; on x86_64/haswell we're still half the latency
compared to normal divide.
Joint work with Daniel Borkmann.
[1] http://www.wireshark.org/~darkjames/reciprocal-buggy.c
[2] http://www.wireshark.org/~darkjames/set-and-dump-filter-k-bug.c
[3] https://gmplib.org/~tege/division-paper.pdf
[4] http://homepage.cs.uiowa.edu/~jones/bcd/divide.html
[5] http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1.2556
[6] http://www.agner.org/optimize/asmlib.zip
Reported-by: Jakub Zawadzki <darkjames-ws@darkjames.pl>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Austin S Hemmelgarn <ahferroin7@gmail.com>
Cc: linux-kernel@vger.kernel.org
Cc: Jesse Gross <jesse@nicira.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Matt Mackall <mpm@selenic.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Andy Gospodarek <andy@greyhouse.net>
Cc: Veaceslav Falico <vfalico@redhat.com>
Cc: Jay Vosburgh <fubar@us.ibm.com>
Cc: Jakub Zawadzki <darkjames-ws@darkjames.pl>
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-01-22 01:29:41 +00:00
|
|
|
} else {
|
|
|
|
/* reciprocal_packets_per_slave is unused if
|
|
|
|
* packets_per_slave is 0 or 1, just initialize it
|
|
|
|
*/
|
|
|
|
bond->params.reciprocal_packets_per_slave =
|
|
|
|
(struct reciprocal_value) { 0 };
|
|
|
|
}
|
2013-12-18 05:30:37 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2014-01-03 22:18:41 +00:00
|
|
|
|
|
|
|
int bond_option_lacp_rate_set(struct bonding *bond, int lacp_rate)
|
|
|
|
{
|
2014-01-03 22:28:18 +00:00
|
|
|
if (bond_parm_tbl_lookup(lacp_rate, bond_lacp_tbl) < 0) {
|
|
|
|
pr_err("%s: Ignoring invalid LACP rate value %d.\n",
|
|
|
|
bond->dev->name, lacp_rate);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2014-01-03 22:18:41 +00:00
|
|
|
if (bond->dev->flags & IFF_UP) {
|
|
|
|
pr_err("%s: Unable to update LACP rate because interface is up.\n",
|
|
|
|
bond->dev->name);
|
|
|
|
return -EPERM;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (bond->params.mode != BOND_MODE_8023AD) {
|
|
|
|
pr_err("%s: Unable to update LACP rate because bond is not in 802.3ad mode.\n",
|
|
|
|
bond->dev->name);
|
|
|
|
return -EPERM;
|
|
|
|
}
|
|
|
|
|
2014-01-03 22:28:18 +00:00
|
|
|
bond->params.lacp_fast = lacp_rate;
|
|
|
|
bond_3ad_update_lacp_rate(bond);
|
|
|
|
pr_info("%s: Setting LACP rate to %s (%d).\n",
|
|
|
|
bond->dev->name, bond_lacp_tbl[lacp_rate].modename,
|
|
|
|
lacp_rate);
|
2014-01-03 22:18:41 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2014-01-03 22:18:49 +00:00
|
|
|
|
|
|
|
int bond_option_ad_select_set(struct bonding *bond, int ad_select)
|
|
|
|
{
|
|
|
|
if (bond_parm_tbl_lookup(ad_select, ad_select_tbl) < 0) {
|
|
|
|
pr_err("%s: Ignoring invalid ad_select value %d.\n",
|
|
|
|
bond->dev->name, ad_select);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2014-01-03 22:28:18 +00:00
|
|
|
if (bond->dev->flags & IFF_UP) {
|
|
|
|
pr_err("%s: Unable to update ad_select because interface is up.\n",
|
|
|
|
bond->dev->name);
|
|
|
|
return -EPERM;
|
|
|
|
}
|
|
|
|
|
2014-01-03 22:18:49 +00:00
|
|
|
bond->params.ad_select = ad_select;
|
|
|
|
pr_info("%s: Setting ad_select to %s (%d).\n",
|
|
|
|
bond->dev->name, ad_select_tbl[ad_select].modename,
|
|
|
|
ad_select);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|