wireguard patches for 5.20-rc1
Subsystems (5)
Jason A. Donenfeld
maintainercover wireguard patches for 5.20-rc1
Hi Jakub,
I had planned to send these out eventually as net.git patches, but as you emailed earlier, I figure there's no harm in just doing this now for net-next.git. Please apply the following small fixes:
1) Rather than using msleep() in order to approximate ktime_get_coarse_ boottime_ns(), instead use an hrtimer, rounded heuristically.
2) An update in selftest config fragments, from Lukas.
3) Linus noticed that a debugging WARN_ON() to detect (impossible) stack corruption would still allow the corruption to happen, making it harder to get the report about the corruption subsequently.
4) Support for User Mode Linux in the test suite. This depends on some UML patches that are slated for 5.20. Richard hasn't sent his pull in, but they're in his tree, so I assume it'll happen.
Thanks,
Jason
Jason A. Donenfeld (3):
wireguard: ratelimiter: use hrtimer in selftest
wireguard: allowedips: don't corrupt stack when detecting overflow
wireguard: selftests: support UML
Lukas Bulwahn (1):
wireguard: selftests: update config fragments
drivers/net/wireguard/allowedips.c | 9 ++++--- drivers/net/wireguard/selftest/allowedips.c | 6 ++--- drivers/net/wireguard/selftest/ratelimiter.c | 25 +++++++++++-------- kernel/time/hrtimer.c | 1 + .../testing/selftests/wireguard/qemu/Makefile | 17 ++++++++++++- .../selftests/wireguard/qemu/arch/um.config | 3 +++ .../selftests/wireguard/qemu/debug.config | 5 ---- .../selftests/wireguard/qemu/kernel.config | 1 - 8 files changed, 43 insertions(+), 24 deletions(-) create mode 100644 tools/testing/selftests/wireguard/qemu/arch/um.config
Signature · 1 line
2.35.1
Jason A. Donenfeld
maintainerPATCH 1/4 wireguard: ratelimiter: use hrtimer in selftest
Using msleep() is problematic because it's compared against ratelimiter.c's ktime_get_coarse_boottime_ns(), which means on systems with slow jiffies (such as UML's forced HZ=100), the result is inaccurate. So switch to using schedule_hrtimeout().
However, hrtimer gives us access only to the traditional posix timers, and none of the _COARSE variants. So now, rather than being too imprecise like jiffies, it's too precise.
One solution would be to give it a large "range" value, but this will still fire early on a loaded system. A better solution is to align the timeout to the actual coarse timer, and then round up to the nearest tick, plus change.
So add the timeout to the current coarse time, and then schedule_hrtimer() until the absolute computed time.
This should hopefully reduce flakes in CI as well. Note that we keep the retry loop in case the entire function is running behind, because the test could still be scheduled out, by either the kernel or by the hypervisor's kernel, in which case restarting the test and hoping to not be scheduled out still helps.
--- drivers/net/wireguard/selftest/ratelimiter.c | 25 +++++++++++--------- kernel/time/hrtimer.c | 1 + 2 files changed, 15 insertions(+), 11 deletions(-)
drivers/net/wireguard/selftest/ratelimiter.c +14 -11
@@ -6,28 +6,29 @@#ifdef DEBUG #include <linux/jiffies.h> +#include <linux/hrtimer.h> static const struct {bool result; -unsigned int msec_to_sleep_before; +u64 nsec_to_sleep_before; } expected_results[]__initconst = { [0 ...PACKETS_BURSTABLE -1 ] = { true,0 }, [PACKETS_BURSTABLE ] = { false,0 }, - [PACKETS_BURSTABLE +1 ] = { true,MSEC_PER_SEC /PACKETS_PER_SECOND }, + [PACKETS_BURSTABLE +1 ] = { true,NSEC_PER_SEC /PACKETS_PER_SECOND }, [PACKETS_BURSTABLE +2 ] = { false,0 }, - [PACKETS_BURSTABLE +3 ] = { true, (MSEC_PER_SEC /PACKETS_PER_SECOND ) *2 }, + [PACKETS_BURSTABLE +3 ] = { true, (NSEC_PER_SEC /PACKETS_PER_SECOND ) *2 }, [PACKETS_BURSTABLE +4 ] = { true,0 }, [PACKETS_BURSTABLE +5 ] = { false,0 } };static __init unsigned int maximum_jiffies_at_index (int index) { -unsigned int total_msecs =2 *MSEC_PER_SEC /PACKETS_PER_SECOND /3 ; +u64 total_nsecs =2 *NSEC_PER_SEC /PACKETS_PER_SECOND /3 ;int i;for (i =0 ; i <= index; ++i) - total_msecs += expected_results[i].msec_to_sleep_before; -return msecs_to_jiffies (total_msecs); + total_nsecs += expected_results[i].nsec_to_sleep_before; +return nsecs_to_jiffies (total_nsecs); } static __init int timings_test(struct sk_buff *skb4, struct iphdr *hdr4,
@@ -42,8 +43,12 @@ static __init int timings_test(struct sk_buff *skb4, struct iphdr *hdr4, loop_start_time = jiffies;for (i =0 ; i <ARRAY_SIZE (expected_results); ++i) { -if (expected_results[i].msec_to_sleep_before) -msleep (expected_results[i].msec_to_sleep_before); +if (expected_results[i].nsec_to_sleep_before) { +ktime_t timeout =ktime_add (ktime_add_ns (ktime_get_coarse_boottime (),TICK_NSEC *4 /3 ), +ns_to_ktime (expected_results[i].nsec_to_sleep_before)); +set_current_state (TASK_UNINTERRUPTIBLE ); +schedule_hrtimeout_range_clock (&timeout,0 ,HRTIMER_MODE_ABS ,CLOCK_BOOTTIME ); + }if (time_is_before_jiffies (loop_start_time +maximum_jiffies_at_index (i)))
@@ -127,7 +132,7 @@ bool __init wg_ratelimiter_selftest(void)if (IS_ENABLED (CONFIG_KASAN ) ||IS_ENABLED (CONFIG_UBSAN ))return true; -BUILD_BUG_ON (MSEC_PER_SEC %PACKETS_PER_SECOND !=0 ); +BUILD_BUG_ON (NSEC_PER_SEC %PACKETS_PER_SECOND !=0 );if (wg_ratelimiter_init ()) goto out;
@@ -176,7 +181,6 @@ bool __init wg_ratelimiter_selftest(void) test += test_count; goto err; } -msleep (500 );continue ; }else if (ret <0 ) { test += test_count;
@@ -195,7 +199,6 @@ bool __init wg_ratelimiter_selftest(void) test += test_count; goto err; } -msleep (50 );continue ; } test += test_count;
kernel/time/hrtimer.c +1 -0
@@ -2311,6 +2311,7 @@ schedule_hrtimeout_range_clock(ktime_t *expires, u64 delta,return !t.task ?0 : -EINTR ; } +EXPORT_SYMBOL_GPL (schedule_hrtimeout_range_clock);/** * schedule_hrtimeout_range - sleep until timeout
Signature · 1 line
2.35.1
Jason A. Donenfeld
maintainerPATCH 2/4 wireguard: selftests: update config fragments
In reply to Jason A. DonenfeldFrom: Lukas Bulwahn <[email protected]>
The kernel.config and debug.config fragments in wireguard selftests mention some config symbols that have been reworked:
Commit c5665868183f ("mm: kmemleak: use the memory pool for early allocations") removes the config DEBUG_KMEMLEAK_EARLY_LOG_SIZE and since then, the config's feature is available without further configuration.
Commit 4675ff05de2d ("kmemcheck: rip it out") removes kmemcheck and the corresponding arch config HAVE_ARCH_KMEMCHECK. There is no need for this config.
Commit 3bf195ae6037 ("netfilter: nat: merge nf_nat_ipv4,6 into nat core") removes the config NF_NAT_IPV4 and since then, the config's feature is available without further configuration.
Commit 41a2901e7d22 ("rcu: Remove SPARSE_RCU_POINTER Kconfig option") removes the config SPARSE_RCU_POINTER and since then, the config's feature is enabled by default.
Commit dfb4357da6dd ("time: Remove CONFIG_TIMER_STATS") removes the feature and config CONFIG_TIMER_STATS without any replacement.
Commit 3ca17b1f3628 ("lib/ubsan: remove null-pointer checks") removes the check and config UBSAN_NULL without any replacement.
Adjust the config fragments to those changes in configs.
--- tools/testing/selftests/wireguard/qemu/debug.config | 5 ----- tools/testing/selftests/wireguard/qemu/kernel.config | 1 - 2 files changed, 6 deletions(-)
tools/testing/selftests/wireguard/qemu/debug.config +0 -5
@@ -18,15 +18,12 @@ CONFIG_DEBUG_VM=y CONFIG_DEBUG_MEMORY_INIT=y CONFIG_HAVE_DEBUG_STACKOVERFLOW=y CONFIG_DEBUG_STACKOVERFLOW=y -CONFIG_HAVE_ARCH_KMEMCHECK=y CONFIG_HAVE_ARCH_KASAN=y CONFIG_KASAN=y CONFIG_KASAN_INLINE=y CONFIG_UBSAN=y CONFIG_UBSAN_SANITIZE_ALL=y -CONFIG_UBSAN_NULL=y CONFIG_DEBUG_KMEMLEAK=y -CONFIG_DEBUG_KMEMLEAK_EARLY_LOG_SIZE=8192 CONFIG_DEBUG_STACK_USAGE=y CONFIG_DEBUG_SHIRQ=y CONFIG_WQ_WATCHDOG=y
@@ -35,7 +32,6 @@ CONFIG_SCHED_INFO=y CONFIG_SCHEDSTATS=y CONFIG_SCHED_STACK_END_CHECK=y CONFIG_DEBUG_TIMEKEEPING=y -CONFIG_TIMER_STATS=y CONFIG_DEBUG_PREEMPT=y CONFIG_DEBUG_RT_MUTEXES=y CONFIG_DEBUG_SPINLOCK=y
@@ -49,7 +45,6 @@ CONFIG_DEBUG_BUGVERBOSE=y CONFIG_DEBUG_LIST=y CONFIG_DEBUG_PLIST=y CONFIG_PROVE_RCU=y -CONFIG_SPARSE_RCU_POINTER=y CONFIG_RCU_CPU_STALL_TIMEOUT=21 CONFIG_RCU_TRACE=y CONFIG_RCU_EQS_DEBUG=y
tools/testing/selftests/wireguard/qemu/kernel.config +0 -1
@@ -19,7 +19,6 @@ CONFIG_NETFILTER_XTABLES=y CONFIG_NETFILTER_XT_NAT=y CONFIG_NETFILTER_XT_MATCH_LENGTH=y CONFIG_NETFILTER_XT_MARK=y -CONFIG_NF_NAT_IPV4=y CONFIG_IP_NF_IPTABLES=y CONFIG_IP_NF_FILTER=y CONFIG_IP_NF_MANGLE=y
Signature · 1 line
2.35.1
Jason A. Donenfeld
maintainerPATCH 3/4 wireguard: allowedips: don't corrupt stack when detecting overflow
In reply to Jason A. DonenfeldIn case push_rcu() and related functions are buggy, there's a WARN_ON(len >= 128), which the selftest tries to hit by being tricky. In case it is hit, we shouldn't corrupt the kernel's stack, though; otherwise it may be hard to even receive the report that it's buggy. So conditionalize the stack write based on that WARN_ON()'s return value.
Note that this never *actually* happens anyway. The WARN_ON() in the first place is bounded by IS_ENABLED(DEBUG), and isn't expected to ever actually hit. This is just a debugging sanity check.
Additionally, hoist the constant 128 into a named enum, MAX_ALLOWEDIPS_BITS, so that it's clear why this value is chosen.
--- drivers/net/wireguard/allowedips.c | 9 ++++++--- drivers/net/wireguard/selftest/allowedips.c | 6 +++--- 2 files changed, 9 insertions(+), 6 deletions(-)
drivers/net/wireguard/allowedips.c +6 -3
@@ -6,6 +6,8 @@#include "allowedips.h" #include "peer.h" +enum {MAX_ALLOWEDIPS_BITS =128 }; +static struct kmem_cache *node_cache;static void swap_endian (u8 *dst,const u8 *src,u8 bits)
@@ -40,7 +42,8 @@ static void push_rcu(struct allowedips_node **stack, struct allowedips_node __rcu *p, unsigned int *len) { if (rcu_access_pointer(p)) { - WARN_ON(IS_ENABLED(DEBUG) && *len >= 128); + if (WARN_ON(IS_ENABLED(DEBUG) && *len >= MAX_ALLOWEDIPS_BITS)) + return; stack[(*len)++] =rcu_dereference_raw (p); } }
@@ -52,7 +55,7 @@ static void node_free_rcu(struct rcu_head *rcu)static void root_free_rcu (struct rcu_head *rcu) { -struct allowedips_node *node, *stack[128 ] = { +struct allowedips_node *node, *stack[MAX_ALLOWEDIPS_BITS ] = { container_of(rcu, struct allowedips_node, rcu) };unsigned int len =1 ;
@@ -65,7 +68,7 @@ static void root_free_rcu(struct rcu_head *rcu)static void root_remove_peer_lists (struct allowedips_node *root) { -struct allowedips_node *node, *stack[128 ] = { root }; +struct allowedips_node *node, *stack[MAX_ALLOWEDIPS_BITS ] = { root };unsigned int len =1 ;while (len >0 && (node = stack[--len])) {
drivers/net/wireguard/selftest/allowedips.c +3 -3
@@ -593,10 +593,10 @@ bool __init wg_allowedips_selftest(void)wg_allowedips_remove_by_peer (&t, a, &mutex);test_negative (4 , a,192 ,168 ,0 ,1 ); -/* These will hit the WARN_ON(len >= 128) in free_node if something -* goes wrong. +/* These will hit the WARN_ON(len >= MAX_ALLOWEDIPS_BITS) in free_node +* if something goes wrong. */ -for (i =0 ; i <128 ; ++i) { +for (i =0 ; i <MAX_ALLOWEDIPS_BITS ; ++i) { part =cpu_to_be64 (~(1LLU << (i %64 )));memset (&ip,0xff ,16 );memcpy ((u8 *)&ip + (i <64 ) *8 , &part,8 );
Signature · 1 line
2.35.1
Jason A. Donenfeld
maintainerPATCH 4/4 wireguard: selftests: support UML
In reply to Jason A. DonenfeldThis shoud open up various possibilities like time travel execution, and is also just another platform to help shake out bugs.
--- tools/testing/selftests/wireguard/qemu/Makefile | 17 ++++++++++++++++- .../selftests/wireguard/qemu/arch/um.config | 3 +++ 2 files changed, 19 insertions(+), 1 deletion(-) create mode 100644 tools/testing/selftests/wireguard/qemu/arch/um.config
tools/testing/selftests/wireguard/qemu/Makefile +16 -1
@@ -248,8 +248,13 @@ QEMU_MACHINE := -cpu host,accel=kvm -machine s390-ccw-virtio -append $(KERNEL_CM else QEMU_MACHINE := -cpu max -machine s390-ccw-virtio -append $(KERNEL_CMDLINE) endif +else ifeq ($(ARCH),um) +CHOST := $(HOST_ARCH)-linux-musl +KERNEL_BZIMAGE := $(KERNEL_BUILD_PATH)/vmlinux +KERNEL_ARCH := um +KERNEL_CMDLINE := $(shell sed -n 's/CONFIG_CMDLINE=\(.*\)/\1/p' arch/um.config) else -$(error I only build: x86_64, i686, arm, armeb, aarch64, aarch64_be, mips, mipsel, mips64, mips64el, powerpc64, powerpc64le, powerpc, m68k, riscv64, riscv32, s390x) +$(error I only build: x86_64, i686, arm, armeb, aarch64, aarch64_be, mips, mipsel, mips64, mips64el, powerpc64, powerpc64le, powerpc, m68k, riscv64, riscv32, s390x, um) endif TOOLCHAIN_FILENAME := $(CHOST)-cross.tgz
@@ -262,7 +267,9 @@ $(eval $(call file_download,$(TOOLCHAIN_FILENAME),$(TOOLCHAIN_DIR),,$(DISTFILES_ STRIP := $(CHOST)-strip CROSS_COMPILE_FLAG := --build=$(CBUILD) --host=$(CHOST) $(info Building for $(CHOST) using $(CBUILD)) +ifneq ($(ARCH),um) export CROSS_COMPILE := $(CHOST)- +endif export PATH := $(TOOLCHAIN_PATH)/bin:$(PATH) export CC := $(CHOST)-gcc CCACHE_PATH := $(shell which ccache 2>/dev/null)
@@ -279,6 +286,7 @@ comma := , build: $(KERNEL_BZIMAGE) qemu: $(KERNEL_BZIMAGE) rm -f $(BUILD_PATH)/result +ifneq ($(ARCH),um) timeout --foreground 20m qemu-system-$(QEMU_ARCH) \ -nodefaults \ -nographic \
@@ -291,6 +299,13 @@ qemu: $(KERNEL_BZIMAGE) -no-reboot \ -monitor none \ -kernel $< +else + timeout --foreground 20m $< \ + $(KERNEL_CMDLINE) \ + mem=$$(grep -q CONFIG_DEBUG_KMEMLEAK=y $(KERNEL_BUILD_PATH)/.config && echo 1G || echo 256M) \ + noreboot \ + con1=fd:51 51>$(BUILD_PATH)/result </dev/null 2>&1 | cat +endif grep -Fq success $(BUILD_PATH)/result $(BUILD_PATH)/init-cpio-spec.txt: $(TOOLCHAIN_PATH)/.installed $(BUILD_PATH)/init
tools/testing/selftests/wireguard/qemu/arch/um.config new file +3 -0
@@ -0,0 +1,3 @@ +CONFIG_64BIT=y +CONFIG_CMDLINE="wg.success=tty1 panic_on_warn=1" +CONFIG_FRAME_WARN=1280
Signature · 1 line
2.35.1
Jason A. Donenfeld
maintainerQuoted 1 line
[PATCH next-next 0/4] wireguard patches for 5.20-rc1
Grrr... obviously I meant net-next not next-next.
Jason