linux/net/ipv4/tcp_cong.c

537 lines
13 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* Pluggable TCP congestion control support and newReno
* congestion control.
* Based on ideas from I/O scheduler support and Web100.
*
* Copyright (C) 2005 Stephen Hemminger <shemminger@osdl.org>
*/
#define pr_fmt(fmt) "TCP: " fmt
#include <linux/module.h>
#include <linux/mm.h>
#include <linux/types.h>
#include <linux/list.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
#include <linux/gfp.h>
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
#include <linux/jhash.h>
#include <net/tcp.h>
#include <trace/events/tcp.h>
static DEFINE_SPINLOCK(tcp_cong_list_lock);
static LIST_HEAD(tcp_cong_list);
/* Simple linear search, don't expect many entries! */
bpf: tcp: Support tcp_congestion_ops in bpf This patch makes "struct tcp_congestion_ops" to be the first user of BPF STRUCT_OPS. It allows implementing a tcp_congestion_ops in bpf. The BPF implemented tcp_congestion_ops can be used like regular kernel tcp-cc through sysctl and setsockopt. e.g. [root@arch-fb-vm1 bpf]# sysctl -a | egrep congestion net.ipv4.tcp_allowed_congestion_control = reno cubic bpf_cubic net.ipv4.tcp_available_congestion_control = reno bic cubic bpf_cubic net.ipv4.tcp_congestion_control = bpf_cubic There has been attempt to move the TCP CC to the user space (e.g. CCP in TCP). The common arguments are faster turn around, get away from long-tail kernel versions in production...etc, which are legit points. BPF has been the continuous effort to join both kernel and userspace upsides together (e.g. XDP to gain the performance advantage without bypassing the kernel). The recent BPF advancements (in particular BTF-aware verifier, BPF trampoline, BPF CO-RE...) made implementing kernel struct ops (e.g. tcp cc) possible in BPF. It allows a faster turnaround for testing algorithm in the production while leveraging the existing (and continue growing) BPF feature/framework instead of building one specifically for userspace TCP CC. This patch allows write access to a few fields in tcp-sock (in bpf_tcp_ca_btf_struct_access()). The optional "get_info" is unsupported now. It can be added later. One possible way is to output the info with a btf-id to describe the content. Signed-off-by: Martin KaFai Lau <kafai@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Andrii Nakryiko <andriin@fb.com> Acked-by: Yonghong Song <yhs@fb.com> Link: https://lore.kernel.org/bpf/20200109003508.3856115-1-kafai@fb.com
2020-01-09 00:35:08 +00:00
struct tcp_congestion_ops *tcp_ca_find(const char *name)
{
struct tcp_congestion_ops *e;
list_for_each_entry_rcu(e, &tcp_cong_list, list) {
if (strcmp(e->name, name) == 0)
return e;
}
return NULL;
}
void tcp_set_ca_state(struct sock *sk, const u8 ca_state)
{
struct inet_connection_sock *icsk = inet_csk(sk);
trace_tcp_cong_state_set(sk, ca_state);
if (icsk->icsk_ca_ops->set_state)
icsk->icsk_ca_ops->set_state(sk, ca_state);
icsk->icsk_ca_state = ca_state;
}
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
/* Must be called with rcu lock held */
static struct tcp_congestion_ops *tcp_ca_find_autoload(struct net *net,
const char *name)
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
{
struct tcp_congestion_ops *ca = tcp_ca_find(name);
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
#ifdef CONFIG_MODULES
if (!ca && capable(CAP_NET_ADMIN)) {
rcu_read_unlock();
request_module("tcp_%s", name);
rcu_read_lock();
ca = tcp_ca_find(name);
}
#endif
return ca;
}
/* Simple linear search, not much in here. */
struct tcp_congestion_ops *tcp_ca_find_key(u32 key)
{
struct tcp_congestion_ops *e;
list_for_each_entry_rcu(e, &tcp_cong_list, list) {
if (e->key == key)
return e;
}
return NULL;
}
int tcp_validate_congestion_control(struct tcp_congestion_ops *ca)
{
/* all algorithms must implement these */
if (!ca->ssthresh || !ca->undo_cwnd ||
!(ca->cong_avoid || ca->cong_control)) {
pr_err("%s does not implement required ops\n", ca->name);
return -EINVAL;
}
return 0;
}
/* Attach new congestion control algorithm to the list
* of available options.
*/
int tcp_register_congestion_control(struct tcp_congestion_ops *ca)
{
int ret;
ret = tcp_validate_congestion_control(ca);
if (ret)
return ret;
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
ca->key = jhash(ca->name, sizeof(ca->name), strlen(ca->name));
spin_lock(&tcp_cong_list_lock);
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
if (ca->key == TCP_CA_UNSPEC || tcp_ca_find_key(ca->key)) {
pr_notice("%s already registered or non-unique key\n",
ca->name);
ret = -EEXIST;
} else {
list_add_tail_rcu(&ca->list, &tcp_cong_list);
pr_debug("%s registered\n", ca->name);
}
spin_unlock(&tcp_cong_list_lock);
return ret;
}
EXPORT_SYMBOL_GPL(tcp_register_congestion_control);
/*
* Remove congestion control algorithm, called from
* the module's remove function. Module ref counts are used
* to ensure that this can't be done till all sockets using
* that method are closed.
*/
void tcp_unregister_congestion_control(struct tcp_congestion_ops *ca)
{
spin_lock(&tcp_cong_list_lock);
list_del_rcu(&ca->list);
spin_unlock(&tcp_cong_list_lock);
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
/* Wait for outstanding readers to complete before the
* module gets removed entirely.
*
* A try_module_get() should fail by now as our module is
* in "going" state since no refs are held anymore and
* module_exit() handler being called.
*/
synchronize_rcu();
}
EXPORT_SYMBOL_GPL(tcp_unregister_congestion_control);
/* Replace a registered old ca with a new one.
*
* The new ca must have the same name as the old one, that has been
* registered.
*/
int tcp_update_congestion_control(struct tcp_congestion_ops *ca, struct tcp_congestion_ops *old_ca)
{
struct tcp_congestion_ops *existing;
int ret = 0;
ca->key = jhash(ca->name, sizeof(ca->name), strlen(ca->name));
spin_lock(&tcp_cong_list_lock);
existing = tcp_ca_find_key(old_ca->key);
if (ca->key == TCP_CA_UNSPEC || !existing || strcmp(existing->name, ca->name)) {
pr_notice("%s not registered or non-unique key\n",
ca->name);
ret = -EINVAL;
} else if (existing != old_ca) {
pr_notice("invalid old congestion control algorithm to replace\n");
ret = -EINVAL;
} else {
/* Add the new one before removing the old one to keep
* one implementation available all the time.
*/
list_add_tail_rcu(&ca->list, &tcp_cong_list);
list_del_rcu(&existing->list);
pr_debug("%s updated\n", ca->name);
}
spin_unlock(&tcp_cong_list_lock);
/* Wait for outstanding readers to complete before the
* module or struct_ops gets removed entirely.
*/
if (!ret)
synchronize_rcu();
return ret;
}
u32 tcp_ca_get_key_by_name(struct net *net, const char *name, bool *ecn_ca)
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
{
const struct tcp_congestion_ops *ca;
u32 key = TCP_CA_UNSPEC;
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
might_sleep();
rcu_read_lock();
ca = tcp_ca_find_autoload(net, name);
if (ca) {
key = ca->key;
*ecn_ca = ca->flags & TCP_CONG_NEEDS_ECN;
}
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
rcu_read_unlock();
return key;
}
char *tcp_ca_get_name_by_key(u32 key, char *buffer)
{
const struct tcp_congestion_ops *ca;
char *ret = NULL;
rcu_read_lock();
ca = tcp_ca_find_key(key);
if (ca)
ret = strncpy(buffer, ca->name,
TCP_CA_NAME_MAX);
rcu_read_unlock();
return ret;
}
/* Assign choice of congestion control. */
void tcp_assign_congestion_control(struct sock *sk)
{
struct net *net = sock_net(sk);
struct inet_connection_sock *icsk = inet_csk(sk);
const struct tcp_congestion_ops *ca;
rcu_read_lock();
ca = rcu_dereference(net->ipv4.tcp_congestion_control);
bpf: tcp: Support tcp_congestion_ops in bpf This patch makes "struct tcp_congestion_ops" to be the first user of BPF STRUCT_OPS. It allows implementing a tcp_congestion_ops in bpf. The BPF implemented tcp_congestion_ops can be used like regular kernel tcp-cc through sysctl and setsockopt. e.g. [root@arch-fb-vm1 bpf]# sysctl -a | egrep congestion net.ipv4.tcp_allowed_congestion_control = reno cubic bpf_cubic net.ipv4.tcp_available_congestion_control = reno bic cubic bpf_cubic net.ipv4.tcp_congestion_control = bpf_cubic There has been attempt to move the TCP CC to the user space (e.g. CCP in TCP). The common arguments are faster turn around, get away from long-tail kernel versions in production...etc, which are legit points. BPF has been the continuous effort to join both kernel and userspace upsides together (e.g. XDP to gain the performance advantage without bypassing the kernel). The recent BPF advancements (in particular BTF-aware verifier, BPF trampoline, BPF CO-RE...) made implementing kernel struct ops (e.g. tcp cc) possible in BPF. It allows a faster turnaround for testing algorithm in the production while leveraging the existing (and continue growing) BPF feature/framework instead of building one specifically for userspace TCP CC. This patch allows write access to a few fields in tcp-sock (in bpf_tcp_ca_btf_struct_access()). The optional "get_info" is unsupported now. It can be added later. One possible way is to output the info with a btf-id to describe the content. Signed-off-by: Martin KaFai Lau <kafai@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Andrii Nakryiko <andriin@fb.com> Acked-by: Yonghong Song <yhs@fb.com> Link: https://lore.kernel.org/bpf/20200109003508.3856115-1-kafai@fb.com
2020-01-09 00:35:08 +00:00
if (unlikely(!bpf_try_module_get(ca, ca->owner)))
ca = &tcp_reno;
icsk->icsk_ca_ops = ca;
rcu_read_unlock();
memset(icsk->icsk_ca_priv, 0, sizeof(icsk->icsk_ca_priv));
if (ca->flags & TCP_CONG_NEEDS_ECN)
INET_ECN_xmit(sk);
else
INET_ECN_dontxmit(sk);
}
void tcp_init_congestion_control(struct sock *sk)
{
struct inet_connection_sock *icsk = inet_csk(sk);
tcp_sk(sk)->prior_ssthresh = 0;
if (icsk->icsk_ca_ops->init)
icsk->icsk_ca_ops->init(sk);
if (tcp_ca_needs_ecn(sk))
INET_ECN_xmit(sk);
else
INET_ECN_dontxmit(sk);
icsk->icsk_ca_initialized = 1;
}
static void tcp_reinit_congestion_control(struct sock *sk,
const struct tcp_congestion_ops *ca)
{
struct inet_connection_sock *icsk = inet_csk(sk);
tcp_cleanup_congestion_control(sk);
icsk->icsk_ca_ops = ca;
tcp: fix child sockets to use system default congestion control if not set Linux 3.17 and earlier are explicitly engineered so that if the app doesn't specifically request a CC module on a listener before the SYN arrives, then the child gets the system default CC when the connection is established. See tcp_init_congestion_control() in 3.17 or earlier, which says "if no choice made yet assign the current value set as default". The change ("net: tcp: assign tcp cong_ops when tcp sk is created") altered these semantics, so that children got their parent listener's congestion control even if the system default had changed after the listener was created. This commit returns to those original semantics from 3.17 and earlier, since they are the original semantics from 2007 in 4d4d3d1e8 ("[TCP]: Congestion control initialization."), and some Linux congestion control workflows depend on that. In summary, if a listener socket specifically sets TCP_CONGESTION to "x", or the route locks the CC module to "x", then the child gets "x". Otherwise the child gets current system default from net.ipv4.tcp_congestion_control. That's the behavior in 3.17 and earlier, and this commit restores that. Fixes: 55d8694fa82c ("net: tcp: assign tcp cong_ops when tcp sk is created") Cc: Florian Westphal <fw@strlen.de> Cc: Daniel Borkmann <dborkman@redhat.com> Cc: Glenn Judd <glenn.judd@morganstanley.com> Cc: Stephen Hemminger <stephen@networkplumber.org> Signed-off-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Yuchung Cheng <ycheng@google.com> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-29 17:47:07 +00:00
icsk->icsk_ca_setsockopt = 1;
memset(icsk->icsk_ca_priv, 0, sizeof(icsk->icsk_ca_priv));
if (ca->flags & TCP_CONG_NEEDS_ECN)
INET_ECN_xmit(sk);
else
INET_ECN_dontxmit(sk);
tcp: make sure listeners don't initialize congestion-control state syzkaller found its way into setsockopt with TCP_CONGESTION "cdg". tcp_cdg_init() does a kcalloc to store the gradients. As sk_clone_lock just copies all the memory, the allocated pointer will be copied as well, if the app called setsockopt(..., TCP_CONGESTION) on the listener. If now the socket will be destroyed before the congestion-control has properly been initialized (through a call to tcp_init_transfer), we will end up freeing memory that does not belong to that particular socket, opening the door to a double-free: [ 11.413102] ================================================================== [ 11.414181] BUG: KASAN: double-free or invalid-free in tcp_cleanup_congestion_control+0x58/0xd0 [ 11.415329] [ 11.415560] CPU: 3 PID: 4884 Comm: syz-executor.5 Not tainted 5.8.0-rc2 #80 [ 11.416544] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014 [ 11.418148] Call Trace: [ 11.418534] <IRQ> [ 11.418834] dump_stack+0x7d/0xb0 [ 11.419297] print_address_description.constprop.0+0x1a/0x210 [ 11.422079] kasan_report_invalid_free+0x51/0x80 [ 11.423433] __kasan_slab_free+0x15e/0x170 [ 11.424761] kfree+0x8c/0x230 [ 11.425157] tcp_cleanup_congestion_control+0x58/0xd0 [ 11.425872] tcp_v4_destroy_sock+0x57/0x5a0 [ 11.426493] inet_csk_destroy_sock+0x153/0x2c0 [ 11.427093] tcp_v4_syn_recv_sock+0xb29/0x1100 [ 11.427731] tcp_get_cookie_sock+0xc3/0x4a0 [ 11.429457] cookie_v4_check+0x13d0/0x2500 [ 11.433189] tcp_v4_do_rcv+0x60e/0x780 [ 11.433727] tcp_v4_rcv+0x2869/0x2e10 [ 11.437143] ip_protocol_deliver_rcu+0x23/0x190 [ 11.437810] ip_local_deliver+0x294/0x350 [ 11.439566] __netif_receive_skb_one_core+0x15d/0x1a0 [ 11.441995] process_backlog+0x1b1/0x6b0 [ 11.443148] net_rx_action+0x37e/0xc40 [ 11.445361] __do_softirq+0x18c/0x61a [ 11.445881] asm_call_on_stack+0x12/0x20 [ 11.446409] </IRQ> [ 11.446716] do_softirq_own_stack+0x34/0x40 [ 11.447259] do_softirq.part.0+0x26/0x30 [ 11.447827] __local_bh_enable_ip+0x46/0x50 [ 11.448406] ip_finish_output2+0x60f/0x1bc0 [ 11.450109] __ip_queue_xmit+0x71c/0x1b60 [ 11.451861] __tcp_transmit_skb+0x1727/0x3bb0 [ 11.453789] tcp_rcv_state_process+0x3070/0x4d3a [ 11.456810] tcp_v4_do_rcv+0x2ad/0x780 [ 11.457995] __release_sock+0x14b/0x2c0 [ 11.458529] release_sock+0x4a/0x170 [ 11.459005] __inet_stream_connect+0x467/0xc80 [ 11.461435] inet_stream_connect+0x4e/0xa0 [ 11.462043] __sys_connect+0x204/0x270 [ 11.465515] __x64_sys_connect+0x6a/0xb0 [ 11.466088] do_syscall_64+0x3e/0x70 [ 11.466617] entry_SYSCALL_64_after_hwframe+0x44/0xa9 [ 11.467341] RIP: 0033:0x7f56046dc469 [ 11.467844] Code: Bad RIP value. [ 11.468282] RSP: 002b:00007f5604dccdd8 EFLAGS: 00000246 ORIG_RAX: 000000000000002a [ 11.469326] RAX: ffffffffffffffda RBX: 000000000068bf00 RCX: 00007f56046dc469 [ 11.470379] RDX: 0000000000000010 RSI: 0000000020000000 RDI: 0000000000000004 [ 11.471311] RBP: 00000000ffffffff R08: 0000000000000000 R09: 0000000000000000 [ 11.472286] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000 [ 11.473341] R13: 000000000041427c R14: 00007f5604dcd5c0 R15: 0000000000000003 [ 11.474321] [ 11.474527] Allocated by task 4884: [ 11.475031] save_stack+0x1b/0x40 [ 11.475548] __kasan_kmalloc.constprop.0+0xc2/0xd0 [ 11.476182] tcp_cdg_init+0xf0/0x150 [ 11.476744] tcp_init_congestion_control+0x9b/0x3a0 [ 11.477435] tcp_set_congestion_control+0x270/0x32f [ 11.478088] do_tcp_setsockopt.isra.0+0x521/0x1a00 [ 11.478744] __sys_setsockopt+0xff/0x1e0 [ 11.479259] __x64_sys_setsockopt+0xb5/0x150 [ 11.479895] do_syscall_64+0x3e/0x70 [ 11.480395] entry_SYSCALL_64_after_hwframe+0x44/0xa9 [ 11.481097] [ 11.481321] Freed by task 4872: [ 11.481783] save_stack+0x1b/0x40 [ 11.482230] __kasan_slab_free+0x12c/0x170 [ 11.482839] kfree+0x8c/0x230 [ 11.483240] tcp_cleanup_congestion_control+0x58/0xd0 [ 11.483948] tcp_v4_destroy_sock+0x57/0x5a0 [ 11.484502] inet_csk_destroy_sock+0x153/0x2c0 [ 11.485144] tcp_close+0x932/0xfe0 [ 11.485642] inet_release+0xc1/0x1c0 [ 11.486131] __sock_release+0xc0/0x270 [ 11.486697] sock_close+0xc/0x10 [ 11.487145] __fput+0x277/0x780 [ 11.487632] task_work_run+0xeb/0x180 [ 11.488118] __prepare_exit_to_usermode+0x15a/0x160 [ 11.488834] do_syscall_64+0x4a/0x70 [ 11.489326] entry_SYSCALL_64_after_hwframe+0x44/0xa9 Wei Wang fixed a part of these CDG-malloc issues with commit c12014440750 ("tcp: memset ca_priv data to 0 properly"). This patch here fixes the listener-scenario: We make sure that listeners setting the congestion-control through setsockopt won't initialize it (thus CDG never allocates on listeners). For those who use AF_UNSPEC to reuse a socket, tcp_disconnect() is changed to cleanup afterwards. (The issue can be reproduced at least down to v4.4.x.) Cc: Wei Wang <weiwan@google.com> Cc: Eric Dumazet <edumazet@google.com> Fixes: 2b0a8c9eee81 ("tcp: add CDG congestion control") Signed-off-by: Christoph Paasch <cpaasch@apple.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-07-08 23:18:34 +00:00
if (!((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
tcp_init_congestion_control(sk);
}
/* Manage refcounts on socket close. */
void tcp_cleanup_congestion_control(struct sock *sk)
{
struct inet_connection_sock *icsk = inet_csk(sk);
if (icsk->icsk_ca_ops->release)
icsk->icsk_ca_ops->release(sk);
bpf: tcp: Support tcp_congestion_ops in bpf This patch makes "struct tcp_congestion_ops" to be the first user of BPF STRUCT_OPS. It allows implementing a tcp_congestion_ops in bpf. The BPF implemented tcp_congestion_ops can be used like regular kernel tcp-cc through sysctl and setsockopt. e.g. [root@arch-fb-vm1 bpf]# sysctl -a | egrep congestion net.ipv4.tcp_allowed_congestion_control = reno cubic bpf_cubic net.ipv4.tcp_available_congestion_control = reno bic cubic bpf_cubic net.ipv4.tcp_congestion_control = bpf_cubic There has been attempt to move the TCP CC to the user space (e.g. CCP in TCP). The common arguments are faster turn around, get away from long-tail kernel versions in production...etc, which are legit points. BPF has been the continuous effort to join both kernel and userspace upsides together (e.g. XDP to gain the performance advantage without bypassing the kernel). The recent BPF advancements (in particular BTF-aware verifier, BPF trampoline, BPF CO-RE...) made implementing kernel struct ops (e.g. tcp cc) possible in BPF. It allows a faster turnaround for testing algorithm in the production while leveraging the existing (and continue growing) BPF feature/framework instead of building one specifically for userspace TCP CC. This patch allows write access to a few fields in tcp-sock (in bpf_tcp_ca_btf_struct_access()). The optional "get_info" is unsupported now. It can be added later. One possible way is to output the info with a btf-id to describe the content. Signed-off-by: Martin KaFai Lau <kafai@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Andrii Nakryiko <andriin@fb.com> Acked-by: Yonghong Song <yhs@fb.com> Link: https://lore.kernel.org/bpf/20200109003508.3856115-1-kafai@fb.com
2020-01-09 00:35:08 +00:00
bpf_module_put(icsk->icsk_ca_ops, icsk->icsk_ca_ops->owner);
}
/* Used by sysctl to change default congestion control */
int tcp_set_default_congestion_control(struct net *net, const char *name)
{
struct tcp_congestion_ops *ca;
const struct tcp_congestion_ops *prev;
int ret;
rcu_read_lock();
ca = tcp_ca_find_autoload(net, name);
if (!ca) {
ret = -ENOENT;
bpf: tcp: Support tcp_congestion_ops in bpf This patch makes "struct tcp_congestion_ops" to be the first user of BPF STRUCT_OPS. It allows implementing a tcp_congestion_ops in bpf. The BPF implemented tcp_congestion_ops can be used like regular kernel tcp-cc through sysctl and setsockopt. e.g. [root@arch-fb-vm1 bpf]# sysctl -a | egrep congestion net.ipv4.tcp_allowed_congestion_control = reno cubic bpf_cubic net.ipv4.tcp_available_congestion_control = reno bic cubic bpf_cubic net.ipv4.tcp_congestion_control = bpf_cubic There has been attempt to move the TCP CC to the user space (e.g. CCP in TCP). The common arguments are faster turn around, get away from long-tail kernel versions in production...etc, which are legit points. BPF has been the continuous effort to join both kernel and userspace upsides together (e.g. XDP to gain the performance advantage without bypassing the kernel). The recent BPF advancements (in particular BTF-aware verifier, BPF trampoline, BPF CO-RE...) made implementing kernel struct ops (e.g. tcp cc) possible in BPF. It allows a faster turnaround for testing algorithm in the production while leveraging the existing (and continue growing) BPF feature/framework instead of building one specifically for userspace TCP CC. This patch allows write access to a few fields in tcp-sock (in bpf_tcp_ca_btf_struct_access()). The optional "get_info" is unsupported now. It can be added later. One possible way is to output the info with a btf-id to describe the content. Signed-off-by: Martin KaFai Lau <kafai@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Andrii Nakryiko <andriin@fb.com> Acked-by: Yonghong Song <yhs@fb.com> Link: https://lore.kernel.org/bpf/20200109003508.3856115-1-kafai@fb.com
2020-01-09 00:35:08 +00:00
} else if (!bpf_try_module_get(ca, ca->owner)) {
ret = -EBUSY;
} else if (!net_eq(net, &init_net) &&
!(ca->flags & TCP_CONG_NON_RESTRICTED)) {
/* Only init netns can set default to a restricted algorithm */
ret = -EPERM;
} else {
prev = xchg(&net->ipv4.tcp_congestion_control, ca);
if (prev)
bpf: tcp: Support tcp_congestion_ops in bpf This patch makes "struct tcp_congestion_ops" to be the first user of BPF STRUCT_OPS. It allows implementing a tcp_congestion_ops in bpf. The BPF implemented tcp_congestion_ops can be used like regular kernel tcp-cc through sysctl and setsockopt. e.g. [root@arch-fb-vm1 bpf]# sysctl -a | egrep congestion net.ipv4.tcp_allowed_congestion_control = reno cubic bpf_cubic net.ipv4.tcp_available_congestion_control = reno bic cubic bpf_cubic net.ipv4.tcp_congestion_control = bpf_cubic There has been attempt to move the TCP CC to the user space (e.g. CCP in TCP). The common arguments are faster turn around, get away from long-tail kernel versions in production...etc, which are legit points. BPF has been the continuous effort to join both kernel and userspace upsides together (e.g. XDP to gain the performance advantage without bypassing the kernel). The recent BPF advancements (in particular BTF-aware verifier, BPF trampoline, BPF CO-RE...) made implementing kernel struct ops (e.g. tcp cc) possible in BPF. It allows a faster turnaround for testing algorithm in the production while leveraging the existing (and continue growing) BPF feature/framework instead of building one specifically for userspace TCP CC. This patch allows write access to a few fields in tcp-sock (in bpf_tcp_ca_btf_struct_access()). The optional "get_info" is unsupported now. It can be added later. One possible way is to output the info with a btf-id to describe the content. Signed-off-by: Martin KaFai Lau <kafai@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Andrii Nakryiko <andriin@fb.com> Acked-by: Yonghong Song <yhs@fb.com> Link: https://lore.kernel.org/bpf/20200109003508.3856115-1-kafai@fb.com
2020-01-09 00:35:08 +00:00
bpf_module_put(prev, prev->owner);
ca->flags |= TCP_CONG_NON_RESTRICTED;
ret = 0;
}
rcu_read_unlock();
return ret;
}
/* Set default value from kernel configuration at bootup */
static int __init tcp_congestion_default(void)
{
return tcp_set_default_congestion_control(&init_net,
CONFIG_DEFAULT_TCP_CONG);
}
late_initcall(tcp_congestion_default);
/* Build string with list of available congestion control values */
void tcp_get_available_congestion_control(char *buf, size_t maxlen)
{
struct tcp_congestion_ops *ca;
size_t offs = 0;
rcu_read_lock();
list_for_each_entry_rcu(ca, &tcp_cong_list, list) {
offs += snprintf(buf + offs, maxlen - offs,
"%s%s",
offs == 0 ? "" : " ", ca->name);
if (WARN_ON_ONCE(offs >= maxlen))
break;
}
rcu_read_unlock();
}
/* Get current default congestion control */
void tcp_get_default_congestion_control(struct net *net, char *name)
{
const struct tcp_congestion_ops *ca;
rcu_read_lock();
ca = rcu_dereference(net->ipv4.tcp_congestion_control);
strncpy(name, ca->name, TCP_CA_NAME_MAX);
rcu_read_unlock();
}
/* Built list of non-restricted congestion control values */
void tcp_get_allowed_congestion_control(char *buf, size_t maxlen)
{
struct tcp_congestion_ops *ca;
size_t offs = 0;
*buf = '\0';
rcu_read_lock();
list_for_each_entry_rcu(ca, &tcp_cong_list, list) {
if (!(ca->flags & TCP_CONG_NON_RESTRICTED))
continue;
offs += snprintf(buf + offs, maxlen - offs,
"%s%s",
offs == 0 ? "" : " ", ca->name);
if (WARN_ON_ONCE(offs >= maxlen))
break;
}
rcu_read_unlock();
}
/* Change list of non-restricted congestion control */
int tcp_set_allowed_congestion_control(char *val)
{
struct tcp_congestion_ops *ca;
char *saved_clone, *clone, *name;
int ret = 0;
saved_clone = clone = kstrdup(val, GFP_USER);
if (!clone)
return -ENOMEM;
spin_lock(&tcp_cong_list_lock);
/* pass 1 check for bad entries */
while ((name = strsep(&clone, " ")) && *name) {
ca = tcp_ca_find(name);
if (!ca) {
ret = -ENOENT;
goto out;
}
}
/* pass 2 clear old values */
list_for_each_entry_rcu(ca, &tcp_cong_list, list)
ca->flags &= ~TCP_CONG_NON_RESTRICTED;
/* pass 3 mark as allowed */
while ((name = strsep(&val, " ")) && *name) {
ca = tcp_ca_find(name);
WARN_ON(!ca);
if (ca)
ca->flags |= TCP_CONG_NON_RESTRICTED;
}
out:
spin_unlock(&tcp_cong_list_lock);
kfree(saved_clone);
return ret;
}
/* Change congestion control for socket. If load is false, then it is the
* responsibility of the caller to call tcp_init_congestion_control or
* tcp_reinit_congestion_control (if the current congestion control was
* already initialized.
*/
int tcp_set_congestion_control(struct sock *sk, const char *name, bool load,
bool cap_net_admin)
{
struct inet_connection_sock *icsk = inet_csk(sk);
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
const struct tcp_congestion_ops *ca;
int err = 0;
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
if (icsk->icsk_ca_dst_locked)
return -EPERM;
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
rcu_read_lock();
if (!load)
ca = tcp_ca_find(name);
else
ca = tcp_ca_find_autoload(sock_net(sk), name);
net: tcp: add key management to congestion control This patch adds necessary infrastructure to the congestion control framework for later per route congestion control support. For a per route congestion control possibility, our aim is to store a unique u32 key identifier into dst metrics, which can then be mapped into a tcp_congestion_ops struct. We argue that having a RTAX key entry is the most simple, generic and easy way to manage, and also keeps the memory footprint of dst entries lower on 64 bit than with storing a pointer directly, for example. Having a unique key id also allows for decoupling actual TCP congestion control module management from the FIB layer, i.e. we don't have to care about expensive module refcounting inside the FIB at this point. We first thought of using an IDR store for the realization, which takes over dynamic assignment of unused key space and also performs the key to pointer mapping in RCU. While doing so, we stumbled upon the issue that due to the nature of dynamic key distribution, it just so happens, arguably in very rare occasions, that excessive module loads and unloads can lead to a possible reuse of previously used key space. Thus, previously stale keys in the dst metric are now being reassigned to a different congestion control algorithm, which might lead to unexpected behaviour. One way to resolve this would have been to walk FIBs on the actually rare occasion of a module unload and reset the metric keys for each FIB in each netns, but that's just very costly. Therefore, we argue a better solution is to reuse the unique congestion control algorithm name member and map that into u32 key space through jhash. For that, we split the flags attribute (as it currently uses 2 bits only anyway) into two u32 attributes, flags and key, so that we can keep the cacheline boundary of 2 cachelines on x86_64 and cache the precalculated key at registration time for the fast path. On average we might expect 2 - 4 modules being loaded worst case perhaps 15, so a key collision possibility is extremely low, and guaranteed collision-free on LE/BE for all in-tree modules. Overall this results in much simpler code, and all without the overhead of an IDR. Due to the deterministic nature, modules can now be unloaded, the congestion control algorithm for a specific but unloaded key will fall back to the default one, and on module reload time it will switch back to the expected algorithm transparently. Joint work with Florian Westphal. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-05 22:57:46 +00:00
/* No change asking for existing value */
tcp: fix child sockets to use system default congestion control if not set Linux 3.17 and earlier are explicitly engineered so that if the app doesn't specifically request a CC module on a listener before the SYN arrives, then the child gets the system default CC when the connection is established. See tcp_init_congestion_control() in 3.17 or earlier, which says "if no choice made yet assign the current value set as default". The change ("net: tcp: assign tcp cong_ops when tcp sk is created") altered these semantics, so that children got their parent listener's congestion control even if the system default had changed after the listener was created. This commit returns to those original semantics from 3.17 and earlier, since they are the original semantics from 2007 in 4d4d3d1e8 ("[TCP]: Congestion control initialization."), and some Linux congestion control workflows depend on that. In summary, if a listener socket specifically sets TCP_CONGESTION to "x", or the route locks the CC module to "x", then the child gets "x". Otherwise the child gets current system default from net.ipv4.tcp_congestion_control. That's the behavior in 3.17 and earlier, and this commit restores that. Fixes: 55d8694fa82c ("net: tcp: assign tcp cong_ops when tcp sk is created") Cc: Florian Westphal <fw@strlen.de> Cc: Daniel Borkmann <dborkman@redhat.com> Cc: Glenn Judd <glenn.judd@morganstanley.com> Cc: Stephen Hemminger <stephen@networkplumber.org> Signed-off-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Yuchung Cheng <ycheng@google.com> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-29 17:47:07 +00:00
if (ca == icsk->icsk_ca_ops) {
icsk->icsk_ca_setsockopt = 1;
goto out;
tcp: fix child sockets to use system default congestion control if not set Linux 3.17 and earlier are explicitly engineered so that if the app doesn't specifically request a CC module on a listener before the SYN arrives, then the child gets the system default CC when the connection is established. See tcp_init_congestion_control() in 3.17 or earlier, which says "if no choice made yet assign the current value set as default". The change ("net: tcp: assign tcp cong_ops when tcp sk is created") altered these semantics, so that children got their parent listener's congestion control even if the system default had changed after the listener was created. This commit returns to those original semantics from 3.17 and earlier, since they are the original semantics from 2007 in 4d4d3d1e8 ("[TCP]: Congestion control initialization."), and some Linux congestion control workflows depend on that. In summary, if a listener socket specifically sets TCP_CONGESTION to "x", or the route locks the CC module to "x", then the child gets "x". Otherwise the child gets current system default from net.ipv4.tcp_congestion_control. That's the behavior in 3.17 and earlier, and this commit restores that. Fixes: 55d8694fa82c ("net: tcp: assign tcp cong_ops when tcp sk is created") Cc: Florian Westphal <fw@strlen.de> Cc: Daniel Borkmann <dborkman@redhat.com> Cc: Glenn Judd <glenn.judd@morganstanley.com> Cc: Stephen Hemminger <stephen@networkplumber.org> Signed-off-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Yuchung Cheng <ycheng@google.com> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-29 17:47:07 +00:00
}
if (!ca)
err = -ENOENT;
else if (!((ca->flags & TCP_CONG_NON_RESTRICTED) || cap_net_admin))
err = -EPERM;
else if (!bpf_try_module_get(ca, ca->owner))
err = -EBUSY;
else
tcp_reinit_congestion_control(sk, ca);
out:
rcu_read_unlock();
return err;
}
/* Slow start is used when congestion window is no greater than the slow start
* threshold. We base on RFC2581 and also handle stretch ACKs properly.
* We do not implement RFC3465 Appropriate Byte Counting (ABC) per se but
* something better;) a packet is only considered (s)acked in its entirety to
* defend the ACK attacks described in the RFC. Slow start processes a stretch
* ACK of degree N as if N acks of degree 1 are received back to back except
* ABC caps N to 2. Slow start exits when cwnd grows over ssthresh and
* returns the leftover acks to adjust cwnd in congestion avoidance mode.
*/
__bpf_kfunc u32 tcp_slow_start(struct tcp_sock *tp, u32 acked)
{
u32 cwnd = min(tcp_snd_cwnd(tp) + acked, tp->snd_ssthresh);
acked -= cwnd - tcp_snd_cwnd(tp);
tcp_snd_cwnd_set(tp, min(cwnd, tp->snd_cwnd_clamp));
return acked;
}
EXPORT_SYMBOL_GPL(tcp_slow_start);
/* In theory this is tp->snd_cwnd += 1 / tp->snd_cwnd (or alternative w),
* for every packet that was ACKed.
*/
__bpf_kfunc void tcp_cong_avoid_ai(struct tcp_sock *tp, u32 w, u32 acked)
{
/* If credits accumulated at a higher w, apply them gently now. */
if (tp->snd_cwnd_cnt >= w) {
tp->snd_cwnd_cnt = 0;
tcp_snd_cwnd_set(tp, tcp_snd_cwnd(tp) + 1);
}
tp->snd_cwnd_cnt += acked;
if (tp->snd_cwnd_cnt >= w) {
u32 delta = tp->snd_cwnd_cnt / w;
tp->snd_cwnd_cnt -= delta * w;
tcp_snd_cwnd_set(tp, tcp_snd_cwnd(tp) + delta);
}
tcp_snd_cwnd_set(tp, min(tcp_snd_cwnd(tp), tp->snd_cwnd_clamp));
}
EXPORT_SYMBOL_GPL(tcp_cong_avoid_ai);
/*
* TCP Reno congestion control
* This is special case used for fallback as well.
*/
/* This is Jacobson's slow start and congestion avoidance.
* SIGCOMM '88, p. 328.
*/
__bpf_kfunc void tcp_reno_cong_avoid(struct sock *sk, u32 ack, u32 acked)
{
struct tcp_sock *tp = tcp_sk(sk);
if (!tcp_is_cwnd_limited(sk))
return;
/* In "safe" area, increase. */
if (tcp_in_slow_start(tp)) {
acked = tcp_slow_start(tp, acked);
if (!acked)
return;
}
/* In dangerous area, increase slowly. */
tcp_cong_avoid_ai(tp, tcp_snd_cwnd(tp), acked);
}
EXPORT_SYMBOL_GPL(tcp_reno_cong_avoid);
/* Slow start threshold is half the congestion window (min 2) */
__bpf_kfunc u32 tcp_reno_ssthresh(struct sock *sk)
{
const struct tcp_sock *tp = tcp_sk(sk);
return max(tcp_snd_cwnd(tp) >> 1U, 2U);
}
EXPORT_SYMBOL_GPL(tcp_reno_ssthresh);
__bpf_kfunc u32 tcp_reno_undo_cwnd(struct sock *sk)
{
const struct tcp_sock *tp = tcp_sk(sk);
return max(tcp_snd_cwnd(tp), tp->prior_cwnd);
}
EXPORT_SYMBOL_GPL(tcp_reno_undo_cwnd);
struct tcp_congestion_ops tcp_reno = {
.flags = TCP_CONG_NON_RESTRICTED,
.name = "reno",
.owner = THIS_MODULE,
.ssthresh = tcp_reno_ssthresh,
.cong_avoid = tcp_reno_cong_avoid,
.undo_cwnd = tcp_reno_undo_cwnd,
};