- Linux-kselftest-mirror - lists.linaro.org

[PATCH AUTOSEL 5.12 106/116] selftests: mlxsw: Increase the tolerance of backlog buildup

by Sasha Levin

From: Petr Machata <petrm(a)nvidia.com> [ Upstream commit dda7f4fa55839baeb72ae040aeaf9ccf89d3e416 ] The intention behind this test is to make sure that qdisc limit is correctly projected to the HW. However, first, due to rounding in the qdisc, and then in the driver, the number cannot actually be accurate. And second, the approach to testing this is to oversubscribe the port with traffic generated on the same switch. The actual backlog size therefore fluctuates. In practice, this test proved to be noisier than the rest, and spuriously fails every now and then. Increase the tolerance to 10 % to avoid these issues. Signed-off-by: Petr Machata <petrm(a)nvidia.com> Acked-by: Jiri Pirko <jiri(a)nvidia.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- tools/testing/selftests/drivers/net/mlxsw/sch_red_core.sh | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/drivers/net/mlxsw/sch_red_core.sh b/tools/testing/selftests/drivers/net/mlxsw/sch_red_core.sh index b0cb1aaffdda..33ddd01689be 100644 --- a/tools/testing/selftests/drivers/net/mlxsw/sch_red_core.sh +++ b/tools/testing/selftests/drivers/net/mlxsw/sch_red_core.sh @@ -507,8 +507,8 @@ do_red_test() check_err $? "backlog $backlog / $limit Got $pct% marked packets, expected == 0." local diff=$((limit - backlog)) pct=$((100 * diff / limit)) - ((0 <= pct && pct <= 5)) - check_err $? "backlog $backlog / $limit expected <= 5% distance" + ((0 <= pct && pct <= 10)) + check_err $? "backlog $backlog / $limit expected <= 10% distance" log_test "TC $((vlan - 10)): RED backlog > limit" stop_traffic -- 2.30.2

4 years, 2 months

1
0
0 0

[PATCH AUTOSEL 5.12 078/116] selftests: Set CC to clang in lib.mk if LLVM is set

by Sasha Levin

From: Yonghong Song <yhs(a)fb.com> [ Upstream commit 26e6dd1072763cd5696b75994c03982dde952ad9 ] selftests/bpf/Makefile includes lib.mk. With the following command make -j60 LLVM=1 LLVM_IAS=1 <=== compile kernel make -j60 -C tools/testing/selftests/bpf LLVM=1 LLVM_IAS=1 V=1 some files are still compiled with gcc. This patch fixed lib.mk issue which sets CC to gcc in all cases. Signed-off-by: Yonghong Song <yhs(a)fb.com> Signed-off-by: Alexei Starovoitov <ast(a)kernel.org> Acked-by: Andrii Nakryiko <andrii(a)kernel.org> Link: https://lore.kernel.org/bpf/20210413153413.3027426-1-yhs@fb.com Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- tools/testing/selftests/lib.mk | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/tools/testing/selftests/lib.mk b/tools/testing/selftests/lib.mk index a5ce26d548e4..9a41d8bb9ff1 100644 --- a/tools/testing/selftests/lib.mk +++ b/tools/testing/selftests/lib.mk @@ -1,6 +1,10 @@ # This mimics the top-level Makefile. We do it explicitly here so that this # Makefile can operate with or without the kbuild infrastructure. +ifneq ($(LLVM),) +CC := clang +else CC := $(CROSS_COMPILE)gcc +endif ifeq (0,$(MAKELEVEL)) ifeq ($(OUTPUT),) -- 2.30.2

4 years, 2 months

1
0
0 0

[PATCH AUTOSEL 5.12 043/116] selftests: mptcp: launch mptcp_connect with timeout

by Sasha Levin

From: Matthieu Baerts <matthieu.baerts(a)tessares.net> [ Upstream commit 5888a61cb4e00695075bbacfd86f3fa73af00413 ] 'mptcp_connect' already has a timeout for poll() but in some cases, it is not enough. With "timeout" tool, we will force the command to fail if it doesn't finish on time. Thanks to that, the script will continue and display details about the current state before marking the test as failed. Displaying this state is very important to be able to understand the issue. Best to have our CI reporting the issue than just "the test hanged". Note that in mptcp_connect.sh, we were using a long timeout to validate the fact we cannot create a socket if a sysctl is set. We don't need this timeout. In diag.sh, we want to send signals to mptcp_connect instances that have been started in the netns. But we cannot send this signal to 'timeout' otherwise that will stop the timeout and messages telling us SIGUSR1 has been received will be printed. Instead of trying to find the right PID and storing them in an array, we can simply use the output of 'ip netns pids' which is all the PIDs we want to send signal to. Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/160 Signed-off-by: Matthieu Baerts <matthieu.baerts(a)tessares.net> Signed-off-by: Mat Martineau <mathew.j.martineau(a)linux.intel.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- tools/testing/selftests/net/mptcp/diag.sh | 55 ++++++++++++------- .../selftests/net/mptcp/mptcp_connect.sh | 15 +++-- .../testing/selftests/net/mptcp/mptcp_join.sh | 22 ++++++-- .../selftests/net/mptcp/simult_flows.sh | 13 ++++- 4 files changed, 72 insertions(+), 33 deletions(-) diff --git a/tools/testing/selftests/net/mptcp/diag.sh b/tools/testing/selftests/net/mptcp/diag.sh index 39edce4f541c..2674ba20d524 100755 --- a/tools/testing/selftests/net/mptcp/diag.sh +++ b/tools/testing/selftests/net/mptcp/diag.sh @@ -5,8 +5,9 @@ rndh=$(printf %x $sec)-$(mktemp -u XXXXXX) ns="ns1-$rndh" ksft_skip=4 test_cnt=1 +timeout_poll=100 +timeout_test=$((timeout_poll * 2 + 1)) ret=0 -pids=() flush_pids() { @@ -14,18 +15,14 @@ flush_pids() # give it some time sleep 1.1 - for pid in ${pids[@]}; do - [ -d /proc/$pid ] && kill -SIGUSR1 $pid >/dev/null 2>&1 - done - pids=() + ip netns pids "${ns}" | xargs --no-run-if-empty kill -SIGUSR1 &>/dev/null } cleanup() { + ip netns pids "${ns}" | xargs --no-run-if-empty kill -SIGKILL &>/dev/null + ip netns del $ns - for pid in ${pids[@]}; do - [ -d /proc/$pid ] && kill -9 $pid >/dev/null 2>&1 - done } ip -Version > /dev/null 2>&1 @@ -79,39 +76,57 @@ trap cleanup EXIT ip netns add $ns ip -n $ns link set dev lo up -echo "a" | ip netns exec $ns ./mptcp_connect -p 10000 -l 0.0.0.0 -t 100 >/dev/null & +echo "a" | \ + timeout ${timeout_test} \ + ip netns exec $ns \ + ./mptcp_connect -p 10000 -l -t ${timeout_poll} \ + 0.0.0.0 >/dev/null & sleep 0.1 -pids[0]=$! chk_msk_nr 0 "no msk on netns creation" -echo "b" | ip netns exec $ns ./mptcp_connect -p 10000 127.0.0.1 -j -t 100 >/dev/null & +echo "b" | \ + timeout ${timeout_test} \ + ip netns exec $ns \ + ./mptcp_connect -p 10000 -j -t ${timeout_poll} \ + 127.0.0.1 >/dev/null & sleep 0.1 -pids[1]=$! chk_msk_nr 2 "after MPC handshake " chk_msk_remote_key_nr 2 "....chk remote_key" chk_msk_fallback_nr 0 "....chk no fallback" flush_pids -echo "a" | ip netns exec $ns ./mptcp_connect -p 10001 -s TCP -l 0.0.0.0 -t 100 >/dev/null & -pids[0]=$! +echo "a" | \ + timeout ${timeout_test} \ + ip netns exec $ns \ + ./mptcp_connect -p 10001 -l -s TCP -t ${timeout_poll} \ + 0.0.0.0 >/dev/null & sleep 0.1 -echo "b" | ip netns exec $ns ./mptcp_connect -p 10001 127.0.0.1 -j -t 100 >/dev/null & -pids[1]=$! +echo "b" | \ + timeout ${timeout_test} \ + ip netns exec $ns \ + ./mptcp_connect -p 10001 -j -t ${timeout_poll} \ + 127.0.0.1 >/dev/null & sleep 0.1 chk_msk_fallback_nr 1 "check fallback" flush_pids NR_CLIENTS=100 for I in `seq 1 $NR_CLIENTS`; do - echo "a" | ip netns exec $ns ./mptcp_connect -p $((I+10001)) -l 0.0.0.0 -t 100 -w 10 >/dev/null & - pids[$((I*2))]=$! + echo "a" | \ + timeout ${timeout_test} \ + ip netns exec $ns \ + ./mptcp_connect -p $((I+10001)) -l -w 10 \ + -t ${timeout_poll} 0.0.0.0 >/dev/null & done sleep 0.1 for I in `seq 1 $NR_CLIENTS`; do - echo "b" | ip netns exec $ns ./mptcp_connect -p $((I+10001)) 127.0.0.1 -t 100 -w 10 >/dev/null & - pids[$((I*2 + 1))]=$! + echo "b" | \ + timeout ${timeout_test} \ + ip netns exec $ns \ + ./mptcp_connect -p $((I+10001)) -w 10 \ + -t ${timeout_poll} 127.0.0.1 >/dev/null & done sleep 1.5 diff --git a/tools/testing/selftests/net/mptcp/mptcp_connect.sh b/tools/testing/selftests/net/mptcp/mptcp_connect.sh index 10a030b53b23..65b3b983efc2 100755 --- a/tools/testing/selftests/net/mptcp/mptcp_connect.sh +++ b/tools/testing/selftests/net/mptcp/mptcp_connect.sh @@ -11,7 +11,8 @@ cin="" cout="" ksft_skip=4 capture=false -timeout=30 +timeout_poll=30 +timeout_test=$((timeout_poll * 2 + 1)) ipv6=true ethtool_random_on=true tc_delay="$((RANDOM%50))" @@ -273,7 +274,7 @@ check_mptcp_disabled() ip netns exec ${disabled_ns} sysctl -q net.mptcp.enabled=0 local err=0 - LANG=C ip netns exec ${disabled_ns} ./mptcp_connect -t $timeout -p 10000 -s MPTCP 127.0.0.1 < "$cin" 2>&1 | \ + LANG=C ip netns exec ${disabled_ns} ./mptcp_connect -p 10000 -s MPTCP 127.0.0.1 < "$cin" 2>&1 | \ grep -q "^socket: Protocol not available$" && err=1 ip netns delete ${disabled_ns} @@ -430,14 +431,20 @@ do_transfer() local stat_cookietx_last=$(get_mib_counter "${listener_ns}" "TcpExtSyncookiesSent") local stat_cookierx_last=$(get_mib_counter "${listener_ns}" "TcpExtSyncookiesRecv") - ip netns exec ${listener_ns} ./mptcp_connect -t $timeout -l -p $port -s ${srv_proto} $extra_args $local_addr < "$sin" > "$sout" & + timeout ${timeout_test} \ + ip netns exec ${listener_ns} \ + ./mptcp_connect -t ${timeout_poll} -l -p $port -s ${srv_proto} \ + $extra_args $local_addr < "$sin" > "$sout" & local spid=$! wait_local_port_listen "${listener_ns}" "${port}" local start start=$(date +%s%3N) - ip netns exec ${connector_ns} ./mptcp_connect -t $timeout -p $port -s ${cl_proto} $extra_args $connect_addr < "$cin" > "$cout" & + timeout ${timeout_test} \ + ip netns exec ${connector_ns} \ + ./mptcp_connect -t ${timeout_poll} -p $port -s ${cl_proto} \ + $extra_args $connect_addr < "$cin" > "$cout" & local cpid=$! wait $cpid diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh index ad32240fbfda..43ed99de7734 100755 --- a/tools/testing/selftests/net/mptcp/mptcp_join.sh +++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh @@ -8,7 +8,8 @@ cin="" cinsent="" cout="" ksft_skip=4 -timeout=30 +timeout_poll=30 +timeout_test=$((timeout_poll * 2 + 1)) mptcp_connect="" capture=0 do_all_tests=1 @@ -245,17 +246,26 @@ do_transfer() local_addr="0.0.0.0" fi - ip netns exec ${listener_ns} $mptcp_connect -t $timeout -l -p $port \ - -s ${srv_proto} ${local_addr} < "$sin" > "$sout" & + timeout ${timeout_test} \ + ip netns exec ${listener_ns} \ + $mptcp_connect -t ${timeout_poll} -l -p $port -s ${srv_proto} \ + ${local_addr} < "$sin" > "$sout" & spid=$! sleep 1 if [ "$test_link_fail" -eq 0 ];then - ip netns exec ${connector_ns} $mptcp_connect -t $timeout -p $port -s ${cl_proto} $connect_addr < "$cin" > "$cout" & + timeout ${timeout_test} \ + ip netns exec ${connector_ns} \ + $mptcp_connect -t ${timeout_poll} -p $port -s ${cl_proto} \ + $connect_addr < "$cin" > "$cout" & else - ( cat "$cin" ; sleep 2; link_failure $listener_ns ; cat "$cin" ) | tee "$cinsent" | \ - ip netns exec ${connector_ns} $mptcp_connect -t $timeout -p $port -s ${cl_proto} $connect_addr > "$cout" & + ( cat "$cin" ; sleep 2; link_failure $listener_ns ; cat "$cin" ) | \ + tee "$cinsent" | \ + timeout ${timeout_test} \ + ip netns exec ${connector_ns} \ + $mptcp_connect -t ${timeout_poll} -p $port -s ${cl_proto} \ + $connect_addr > "$cout" & fi cpid=$! diff --git a/tools/testing/selftests/net/mptcp/simult_flows.sh b/tools/testing/selftests/net/mptcp/simult_flows.sh index f039ee57eb3c..3aeef3bcb101 100755 --- a/tools/testing/selftests/net/mptcp/simult_flows.sh +++ b/tools/testing/selftests/net/mptcp/simult_flows.sh @@ -7,7 +7,8 @@ ns2="ns2-$rndh" ns3="ns3-$rndh" capture=false ksft_skip=4 -timeout=30 +timeout_poll=30 +timeout_test=$((timeout_poll * 2 + 1)) test_cnt=1 ret=0 bail=0 @@ -157,14 +158,20 @@ do_transfer() sleep 1 fi - ip netns exec ${ns3} ./mptcp_connect -jt $timeout -l -p $port 0.0.0.0 < "$sin" > "$sout" & + timeout ${timeout_test} \ + ip netns exec ${ns3} \ + ./mptcp_connect -jt ${timeout_poll} -l -p $port \ + 0.0.0.0 < "$sin" > "$sout" & local spid=$! wait_local_port_listen "${ns3}" "${port}" local start start=$(date +%s%3N) - ip netns exec ${ns1} ./mptcp_connect -jt $timeout -p $port 10.0.3.3 < "$cin" > "$cout" & + timeout ${timeout_test} \ + ip netns exec ${ns1} \ + ./mptcp_connect -jt ${timeout_poll} -p $port \ + 10.0.3.3 < "$cin" > "$cout" & local cpid=$! wait $cpid -- 2.30.2

4 years, 2 months

1
0
0 0

[PATCH AUTOSEL 5.12 029/116] selftests/powerpc: Fix L1D flushing tests for Power10

by Sasha Levin

From: Russell Currey <ruscur(a)russell.cc> [ Upstream commit 3a72c94ebfb1f171eba0715998010678a09ec796 ] The rfi_flush and entry_flush selftests work by using the PM_LD_MISS_L1 perf event to count L1D misses. The value of this event has changed over time: - Power7 uses 0x400f0 - Power8 and Power9 use both 0x400f0 and 0x3e054 - Power10 uses only 0x3e054 Rather than relying on raw values, configure perf to count L1D read misses in the most explicit way available. This fixes the selftests to work on systems without 0x400f0 as PM_LD_MISS_L1, and should change no behaviour for systems that the tests already worked on. The only potential downside is that referring to a specific perf event requires PMU support implemented in the kernel for that platform. Signed-off-by: Russell Currey <ruscur(a)russell.cc> Acked-by: Daniel Axtens <dja(a)axtens.net> Signed-off-by: Michael Ellerman <mpe(a)ellerman.id.au> Link: https://lore.kernel.org/r/20210223070227.2916871-1-ruscur@russell.cc Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- tools/testing/selftests/powerpc/security/entry_flush.c | 2 +- tools/testing/selftests/powerpc/security/flush_utils.h | 4 ++++ tools/testing/selftests/powerpc/security/rfi_flush.c | 2 +- 3 files changed, 6 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/powerpc/security/entry_flush.c b/tools/testing/selftests/powerpc/security/entry_flush.c index 78cf914fa321..68ce377b205e 100644 --- a/tools/testing/selftests/powerpc/security/entry_flush.c +++ b/tools/testing/selftests/powerpc/security/entry_flush.c @@ -53,7 +53,7 @@ int entry_flush_test(void) entry_flush = entry_flush_orig; - fd = perf_event_open_counter(PERF_TYPE_RAW, /* L1d miss */ 0x400f0, -1); + fd = perf_event_open_counter(PERF_TYPE_HW_CACHE, PERF_L1D_READ_MISS_CONFIG, -1); FAIL_IF(fd < 0); p = (char *)memalign(zero_size, CACHELINE_SIZE); diff --git a/tools/testing/selftests/powerpc/security/flush_utils.h b/tools/testing/selftests/powerpc/security/flush_utils.h index 07a5eb301466..7a3d60292916 100644 --- a/tools/testing/selftests/powerpc/security/flush_utils.h +++ b/tools/testing/selftests/powerpc/security/flush_utils.h @@ -9,6 +9,10 @@ #define CACHELINE_SIZE 128 +#define PERF_L1D_READ_MISS_CONFIG ((PERF_COUNT_HW_CACHE_L1D) | \ + (PERF_COUNT_HW_CACHE_OP_READ << 8) | \ + (PERF_COUNT_HW_CACHE_RESULT_MISS << 16)) + void syscall_loop(char *p, unsigned long iterations, unsigned long zero_size); diff --git a/tools/testing/selftests/powerpc/security/rfi_flush.c b/tools/testing/selftests/powerpc/security/rfi_flush.c index 7565fd786640..f73484a6470f 100644 --- a/tools/testing/selftests/powerpc/security/rfi_flush.c +++ b/tools/testing/selftests/powerpc/security/rfi_flush.c @@ -54,7 +54,7 @@ int rfi_flush_test(void) rfi_flush = rfi_flush_orig; - fd = perf_event_open_counter(PERF_TYPE_RAW, /* L1d miss */ 0x400f0, -1); + fd = perf_event_open_counter(PERF_TYPE_HW_CACHE, PERF_L1D_READ_MISS_CONFIG, -1); FAIL_IF(fd < 0); p = (char *)memalign(zero_size, CACHELINE_SIZE); -- 2.30.2

4 years, 2 months

1
0
0 0

[PATCH v3] netfilter: nf_conntrack: Add conntrack helper for ESP/IPsec

by Cole Dishington

Introduce changes to add ESP connection tracking helper to netfilter conntrack. The connection tracking of ESP is based on IPsec SPIs. The underlying motivation for this patch was to allow multiple VPN ESP clients to be distinguished when using NAT. Added config flag CONFIG_NF_CT_PROTO_ESP to enable the ESP/IPsec conntrack helper. Signed-off-by: Cole Dishington <Cole.Dishington(a)alliedtelesis.co.nz> --- Notes: Thanks for your time reviewing! Q. > +static int esp_tuple_to_nlattr(struct sk_buff *skb, > + const struct nf_conntrack_tuple *t) > +{ > + if (nla_put_be16(skb, CTA_PROTO_SRC_ESP_ID, t->src.u.esp.id) || > + nla_put_be16(skb, CTA_PROTO_DST_ESP_ID, t->dst.u.esp.id)) > + goto nla_put_failure; This exposes the 16 bit kernel-generated IDs, right? Should this dump the real on-wire SPIs instead? Or is there are reason why the internal IDs need exposure? A. I think I need to expose the internal esp ids here due to esp_nlattr_to_tuple(). If esp id was changed to real SPIs here I would be unable to lookup the correct tuple (without IP addresses too). changes in v3: - Flush all esp entries for a given netns on nf_conntrack_proto_pernet_fini - Replace _esp_table (and its spinlock) shared over netns with per netns linked lists and bitmap (for esp ids) - Init IPv6 any address with IN6ADDR_ANY_INIT rather than ipv6_addr_set() - Change l3num on hash key from u16 to u8 - Add selftests file for testing tracker with ipv4 and ipv6 - Removed credits .../linux/netfilter/nf_conntrack_proto_esp.h | 23 + .../net/netfilter/ipv4/nf_conntrack_ipv4.h | 3 + include/net/netfilter/nf_conntrack.h | 6 + include/net/netfilter/nf_conntrack_l4proto.h | 16 + include/net/netfilter/nf_conntrack_tuple.h | 3 + include/net/netns/conntrack.h | 17 + .../netfilter/nf_conntrack_tuple_common.h | 3 + .../linux/netfilter/nfnetlink_conntrack.h | 2 + net/netfilter/Kconfig | 10 + net/netfilter/Makefile | 1 + net/netfilter/nf_conntrack_core.c | 23 + net/netfilter/nf_conntrack_netlink.c | 4 +- net/netfilter/nf_conntrack_proto.c | 15 + net/netfilter/nf_conntrack_proto_esp.c | 741 ++++++++++++++++++ net/netfilter/nf_conntrack_standalone.c | 8 + net/netfilter/nf_internals.h | 4 +- .../netfilter/conntrack_esp_related.sh | 268 +++++++ 17 files changed, 1145 insertions(+), 2 deletions(-) create mode 100644 include/linux/netfilter/nf_conntrack_proto_esp.h create mode 100644 net/netfilter/nf_conntrack_proto_esp.c create mode 100755 tools/testing/selftests/netfilter/conntrack_esp_related.sh diff --git a/include/linux/netfilter/nf_conntrack_proto_esp.h b/include/linux/netfilter/nf_conntrack_proto_esp.h new file mode 100644 index 000000000000..96888669edd7 --- /dev/null +++ b/include/linux/netfilter/nf_conntrack_proto_esp.h @@ -0,0 +1,23 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _CONNTRACK_PROTO_ESP_H +#define _CONNTRACK_PROTO_ESP_H +#include <asm/byteorder.h> +#include <net/netfilter/nf_conntrack_tuple.h> + +/* ESP PROTOCOL HEADER */ + +struct esphdr { + __u32 spi; +}; + +struct nf_ct_esp { + __u32 l_spi, r_spi; +}; + +void nf_ct_esp_pernet_flush(struct net *net); + +void destroy_esp_conntrack_entry(struct nf_conn *ct); + +bool esp_pkt_to_tuple(const struct sk_buff *skb, unsigned int dataoff, + struct net *net, struct nf_conntrack_tuple *tuple); +#endif /* _CONNTRACK_PROTO_ESP_H */ diff --git a/include/net/netfilter/ipv4/nf_conntrack_ipv4.h b/include/net/netfilter/ipv4/nf_conntrack_ipv4.h index 2c8c2b023848..1aee91592639 100644 --- a/include/net/netfilter/ipv4/nf_conntrack_ipv4.h +++ b/include/net/netfilter/ipv4/nf_conntrack_ipv4.h @@ -25,5 +25,8 @@ extern const struct nf_conntrack_l4proto nf_conntrack_l4proto_udplite; #ifdef CONFIG_NF_CT_PROTO_GRE extern const struct nf_conntrack_l4proto nf_conntrack_l4proto_gre; #endif +#ifdef CONFIG_NF_CT_PROTO_ESP +extern const struct nf_conntrack_l4proto nf_conntrack_l4proto_esp; +#endif #endif /*_NF_CONNTRACK_IPV4_H*/ diff --git a/include/net/netfilter/nf_conntrack.h b/include/net/netfilter/nf_conntrack.h index 439379ca9ffa..4011be8c5e39 100644 --- a/include/net/netfilter/nf_conntrack.h +++ b/include/net/netfilter/nf_conntrack.h @@ -21,6 +21,7 @@ #include <linux/netfilter/nf_conntrack_dccp.h> #include <linux/netfilter/nf_conntrack_sctp.h> #include <linux/netfilter/nf_conntrack_proto_gre.h> +#include <linux/netfilter/nf_conntrack_proto_esp.h> #include <net/netfilter/nf_conntrack_tuple.h> @@ -36,6 +37,7 @@ union nf_conntrack_proto { struct ip_ct_tcp tcp; struct nf_ct_udp udp; struct nf_ct_gre gre; + struct nf_ct_esp esp; unsigned int tmpl_padto; }; @@ -47,6 +49,10 @@ struct nf_conntrack_net { unsigned int users4; unsigned int users6; unsigned int users_bridge; + +#ifdef CONFIG_NF_CT_PROTO_ESP + DECLARE_BITMAP(esp_id_map, 1024); +#endif }; #include <linux/types.h> diff --git a/include/net/netfilter/nf_conntrack_l4proto.h b/include/net/netfilter/nf_conntrack_l4proto.h index 96f9cf81f46b..f700de0b9059 100644 --- a/include/net/netfilter/nf_conntrack_l4proto.h +++ b/include/net/netfilter/nf_conntrack_l4proto.h @@ -75,6 +75,8 @@ bool nf_conntrack_invert_icmp_tuple(struct nf_conntrack_tuple *tuple, const struct nf_conntrack_tuple *orig); bool nf_conntrack_invert_icmpv6_tuple(struct nf_conntrack_tuple *tuple, const struct nf_conntrack_tuple *orig); +bool nf_conntrack_invert_esp_tuple(struct nf_conntrack_tuple *tuple, + const struct nf_conntrack_tuple *orig); int nf_conntrack_inet_error(struct nf_conn *tmpl, struct sk_buff *skb, unsigned int dataoff, @@ -132,6 +134,11 @@ int nf_conntrack_gre_packet(struct nf_conn *ct, unsigned int dataoff, enum ip_conntrack_info ctinfo, const struct nf_hook_state *state); +int nf_conntrack_esp_packet(struct nf_conn *ct, + struct sk_buff *skb, + unsigned int dataoff, + enum ip_conntrack_info ctinfo, + const struct nf_hook_state *state); void nf_conntrack_generic_init_net(struct net *net); void nf_conntrack_tcp_init_net(struct net *net); @@ -141,6 +148,8 @@ void nf_conntrack_dccp_init_net(struct net *net); void nf_conntrack_sctp_init_net(struct net *net); void nf_conntrack_icmp_init_net(struct net *net); void nf_conntrack_icmpv6_init_net(struct net *net); +int nf_conntrack_esp_init(void); +void nf_conntrack_esp_init_net(struct net *net); /* Existing built-in generic protocol */ extern const struct nf_conntrack_l4proto nf_conntrack_l4proto_generic; @@ -240,4 +249,11 @@ static inline struct nf_gre_net *nf_gre_pernet(struct net *net) } #endif +#ifdef CONFIG_NF_CT_PROTO_ESP +static inline struct nf_esp_net *nf_esp_pernet(struct net *net) +{ + return &net->ct.nf_ct_proto.esp; +} +#endif + #endif /*_NF_CONNTRACK_PROTOCOL_H*/ diff --git a/include/net/netfilter/nf_conntrack_tuple.h b/include/net/netfilter/nf_conntrack_tuple.h index 9334371c94e2..60279ffabe36 100644 --- a/include/net/netfilter/nf_conntrack_tuple.h +++ b/include/net/netfilter/nf_conntrack_tuple.h @@ -62,6 +62,9 @@ struct nf_conntrack_tuple { struct { __be16 key; } gre; + struct { + __be16 id; + } esp; } u; /* The protocol. */ diff --git a/include/net/netns/conntrack.h b/include/net/netns/conntrack.h index 806454e767bf..43cd1e78f790 100644 --- a/include/net/netns/conntrack.h +++ b/include/net/netns/conntrack.h @@ -69,6 +69,20 @@ struct nf_gre_net { }; #endif +#ifdef CONFIG_NF_CT_PROTO_ESP +enum esp_conntrack { + ESP_CT_UNREPLIED, + ESP_CT_REPLIED, + ESP_CT_MAX +}; + +struct nf_esp_net { + spinlock_t id_list_lock; + struct list_head id_list; + unsigned int esp_timeouts[ESP_CT_MAX]; +}; +#endif + struct nf_ip_net { struct nf_generic_net generic; struct nf_tcp_net tcp; @@ -84,6 +98,9 @@ struct nf_ip_net { #ifdef CONFIG_NF_CT_PROTO_GRE struct nf_gre_net gre; #endif +#ifdef CONFIG_NF_CT_PROTO_ESP + struct nf_esp_net esp; +#endif }; struct ct_pcpu { diff --git a/include/uapi/linux/netfilter/nf_conntrack_tuple_common.h b/include/uapi/linux/netfilter/nf_conntrack_tuple_common.h index 64390fac6f7e..78600cb4bfff 100644 --- a/include/uapi/linux/netfilter/nf_conntrack_tuple_common.h +++ b/include/uapi/linux/netfilter/nf_conntrack_tuple_common.h @@ -39,6 +39,9 @@ union nf_conntrack_man_proto { struct { __be16 key; /* GRE key is 32bit, PPtP only uses 16bit */ } gre; + struct { + __be16 id; + } esp; }; #define CTINFO2DIR(ctinfo) ((ctinfo) >= IP_CT_IS_REPLY ? IP_CT_DIR_REPLY : IP_CT_DIR_ORIGINAL) diff --git a/include/uapi/linux/netfilter/nfnetlink_conntrack.h b/include/uapi/linux/netfilter/nfnetlink_conntrack.h index d8484be72fdc..744d8931adeb 100644 --- a/include/uapi/linux/netfilter/nfnetlink_conntrack.h +++ b/include/uapi/linux/netfilter/nfnetlink_conntrack.h @@ -90,6 +90,8 @@ enum ctattr_l4proto { CTA_PROTO_ICMPV6_ID, CTA_PROTO_ICMPV6_TYPE, CTA_PROTO_ICMPV6_CODE, + CTA_PROTO_SRC_ESP_ID, + CTA_PROTO_DST_ESP_ID, __CTA_PROTO_MAX }; #define CTA_PROTO_MAX (__CTA_PROTO_MAX - 1) diff --git a/net/netfilter/Kconfig b/net/netfilter/Kconfig index 1a92063c73a4..7269312d322e 100644 --- a/net/netfilter/Kconfig +++ b/net/netfilter/Kconfig @@ -199,6 +199,16 @@ config NF_CT_PROTO_UDPLITE If unsure, say Y. +config NF_CT_PROTO_ESP + bool "ESP protocol support" + depends on NETFILTER_ADVANCED + help + ESP connection tracking helper. Provides connection tracking for IPsec + clients behind this device based on SPI, especially useful for + distinguishing multiple clients when using NAT. + + If unsure, say N. + config NF_CONNTRACK_AMANDA tristate "Amanda backup protocol support" depends on NETFILTER_ADVANCED diff --git a/net/netfilter/Makefile b/net/netfilter/Makefile index 33da7bf1b68e..0942f2c48ddb 100644 --- a/net/netfilter/Makefile +++ b/net/netfilter/Makefile @@ -14,6 +14,7 @@ nf_conntrack-$(CONFIG_NF_CONNTRACK_LABELS) += nf_conntrack_labels.o nf_conntrack-$(CONFIG_NF_CT_PROTO_DCCP) += nf_conntrack_proto_dccp.o nf_conntrack-$(CONFIG_NF_CT_PROTO_SCTP) += nf_conntrack_proto_sctp.o nf_conntrack-$(CONFIG_NF_CT_PROTO_GRE) += nf_conntrack_proto_gre.o +nf_conntrack-$(CONFIG_NF_CT_PROTO_ESP) += nf_conntrack_proto_esp.o obj-$(CONFIG_NETFILTER) = netfilter.o diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c index ff0168736f6e..3bef361d19ce 100644 --- a/net/netfilter/nf_conntrack_core.c +++ b/net/netfilter/nf_conntrack_core.c @@ -295,6 +295,10 @@ nf_ct_get_tuple(const struct sk_buff *skb, #ifdef CONFIG_NF_CT_PROTO_GRE case IPPROTO_GRE: return gre_pkt_to_tuple(skb, dataoff, net, tuple); +#endif +#ifdef CONFIG_NF_CT_PROTO_ESP + case IPPROTO_ESP: + return esp_pkt_to_tuple(skb, dataoff, net, tuple); #endif case IPPROTO_TCP: case IPPROTO_UDP: /* fallthrough */ @@ -439,6 +443,10 @@ nf_ct_invert_tuple(struct nf_conntrack_tuple *inverse, #if IS_ENABLED(CONFIG_IPV6) case IPPROTO_ICMPV6: return nf_conntrack_invert_icmpv6_tuple(inverse, orig); +#endif +#ifdef CONFIG_NF_CT_PROTO_ESP + case IPPROTO_ESP: + return nf_conntrack_invert_esp_tuple(inverse, orig); #endif } @@ -593,6 +601,13 @@ static void destroy_gre_conntrack(struct nf_conn *ct) #endif } +static void destroy_esp_conntrack(struct nf_conn *ct) +{ +#ifdef CONFIG_NF_CT_PROTO_ESP + destroy_esp_conntrack_entry(ct); +#endif +} + static void destroy_conntrack(struct nf_conntrack *nfct) { @@ -609,6 +624,9 @@ destroy_conntrack(struct nf_conntrack *nfct) if (unlikely(nf_ct_protonum(ct) == IPPROTO_GRE)) destroy_gre_conntrack(ct); + if (unlikely(nf_ct_protonum(ct) == IPPROTO_ESP)) + destroy_esp_conntrack(ct); + local_bh_disable(); /* Expectations will have been removed in clean_from_lists, * except TFTP can create an expectation on the first packet, @@ -1783,6 +1801,11 @@ static int nf_conntrack_handle_packet(struct nf_conn *ct, case IPPROTO_GRE: return nf_conntrack_gre_packet(ct, skb, dataoff, ctinfo, state); +#endif +#ifdef CONFIG_NF_CT_PROTO_ESP + case IPPROTO_ESP: + return nf_conntrack_esp_packet(ct, skb, dataoff, + ctinfo, state); #endif } diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index 1d519b0e51a5..8df33dbbf5a3 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -1382,7 +1382,9 @@ static const struct nla_policy tuple_nla_policy[CTA_TUPLE_MAX+1] = { CTA_FILTER_F_CTA_PROTO_ICMP_ID | \ CTA_FILTER_F_CTA_PROTO_ICMPV6_TYPE | \ CTA_FILTER_F_CTA_PROTO_ICMPV6_CODE | \ - CTA_FILTER_F_CTA_PROTO_ICMPV6_ID) + CTA_FILTER_F_CTA_PROTO_ICMPV6_ID | \ + CTA_FILTER_F_CTA_PROTO_SRC_ESP_ID | \ + CTA_FILTER_F_CTA_PROTO_DST_ESP_ID) static int ctnetlink_parse_tuple_filter(const struct nlattr * const cda[], diff --git a/net/netfilter/nf_conntrack_proto.c b/net/netfilter/nf_conntrack_proto.c index 47e9319d2cf3..e71ddb4e33cc 100644 --- a/net/netfilter/nf_conntrack_proto.c +++ b/net/netfilter/nf_conntrack_proto.c @@ -112,6 +112,9 @@ const struct nf_conntrack_l4proto *nf_ct_l4proto_find(u8 l4proto) #ifdef CONFIG_NF_CT_PROTO_GRE case IPPROTO_GRE: return &nf_conntrack_l4proto_gre; #endif +#ifdef CONFIG_NF_CT_PROTO_ESP + case IPPROTO_ESP: return &nf_conntrack_l4proto_esp; +#endif #if IS_ENABLED(CONFIG_IPV6) case IPPROTO_ICMPV6: return &nf_conntrack_l4proto_icmpv6; #endif /* CONFIG_IPV6 */ @@ -656,6 +659,12 @@ int nf_conntrack_proto_init(void) goto cleanup_sockopt; #endif +#ifdef CONFIG_NF_CT_PROTO_ESP + ret = nf_conntrack_esp_init(); + if (ret < 0) + goto cleanup_sockopt; +#endif + return ret; #if IS_ENABLED(CONFIG_IPV6) @@ -691,6 +700,9 @@ void nf_conntrack_proto_pernet_init(struct net *net) #ifdef CONFIG_NF_CT_PROTO_GRE nf_conntrack_gre_init_net(net); #endif +#ifdef CONFIG_NF_CT_PROTO_ESP + nf_conntrack_esp_init_net(net); +#endif } void nf_conntrack_proto_pernet_fini(struct net *net) @@ -698,6 +710,9 @@ void nf_conntrack_proto_pernet_fini(struct net *net) #ifdef CONFIG_NF_CT_PROTO_GRE nf_ct_gre_keymap_flush(net); #endif +#ifdef CONFIG_NF_CT_PROTO_ESP + nf_ct_esp_pernet_flush(net); +#endif } module_param_call(hashsize, nf_conntrack_set_hashsize, param_get_uint, diff --git a/net/netfilter/nf_conntrack_proto_esp.c b/net/netfilter/nf_conntrack_proto_esp.c new file mode 100644 index 000000000000..1bc0cb879bfd --- /dev/null +++ b/net/netfilter/nf_conntrack_proto_esp.c @@ -0,0 +1,741 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * <:copyright-gpl + * Copyright 2008 Broadcom Corp. All Rights Reserved. + * Copyright (C) 2021 Allied Telesis Labs NZ + * + * This program is free software; you can distribute it and/or modify it + * under the terms of the GNU General Public License (Version 2) as + * published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License + * for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program. + * :> + */ + +#include <linux/module.h> +#include <linux/types.h> +#include <linux/timer.h> +#include <linux/list.h> +#include <linux/seq_file.h> +#include <linux/in.h> +#include <linux/netdevice.h> +#include <linux/skbuff.h> +#include <linux/ip.h> +#include <net/dst.h> +#include <net/netfilter/nf_conntrack.h> +#include <net/netfilter/nf_conntrack_l4proto.h> +#include <net/netfilter/nf_conntrack_helper.h> +#include <net/netfilter/nf_conntrack_core.h> +#include <net/netfilter/nf_conntrack_timeout.h> +#include <linux/netfilter/nf_conntrack_proto_esp.h> +#include <net/netns/hash.h> +#include <linux/rhashtable.h> +#include <net/ipv6.h> + +#include "nf_internals.h" + +/* esp_id of 0 is left for unassigned values */ +#define TEMP_SPI_START 1 +#define TEMP_SPI_MAX (TEMP_SPI_START + 1024 - 1) + +struct _esp_entry { + /* linked list node for per net lookup via esp_id */ + struct list_head net_node; + + /* Hash table nodes for each required lookup + * lnode: net->hash_mix, l_spi, l_ip, r_ip + * rnode: net->hash_mix, r_spi, r_ip + * incmpl_rlist: net->hash_mix, r_ip + */ + struct rhash_head lnode; + struct rhash_head rnode; + struct rhlist_head incmpl_rlist; + + u16 esp_id; + + u16 l3num; + + u32 l_spi; + u32 r_spi; + + union nf_inet_addr l_ip; + union nf_inet_addr r_ip; + + u32 alloc_time_jiffies; + struct net *net; +}; + +struct _esp_hkey { + u8 l3num; + union nf_inet_addr src_ip; + union nf_inet_addr dst_ip; + u32 net_hmix; + u32 spi; +}; + +extern unsigned int nf_conntrack_net_id; + +static struct rhashtable ltable; +static struct rhashtable rtable; +static struct rhltable incmpl_rtable; +static unsigned int esp_timeouts[ESP_CT_MAX] = { + [ESP_CT_UNREPLIED] = 60 * HZ, + [ESP_CT_REPLIED] = 3600 * HZ, +}; + +static void esp_ip_addr_copy(int af, union nf_inet_addr *dst, + const union nf_inet_addr *src) +{ + if (af == AF_INET6) + dst->in6 = src->in6; + else + dst->ip = src->ip; +} + +static int esp_ip_addr_equal(int af, const union nf_inet_addr *a, + const union nf_inet_addr *b) +{ + if (af == AF_INET6) + return ipv6_addr_equal(&a->in6, &b->in6); + return a->ip == b->ip; +} + +static inline struct nf_esp_net *esp_pernet(struct net *net) +{ + return &net->ct.nf_ct_proto.esp; +} + +static inline void calculate_key(const u32 net_hmix, const u32 spi, + const u8 l3num, + const union nf_inet_addr *src_ip, + const union nf_inet_addr *dst_ip, + struct _esp_hkey *key) +{ + key->net_hmix = net_hmix; + key->spi = spi; + key->l3num = l3num; + esp_ip_addr_copy(l3num, &key->src_ip, src_ip); + esp_ip_addr_copy(l3num, &key->dst_ip, dst_ip); +} + +static inline u32 calculate_hash(const void *data, u32 len, u32 seed) +{ + return jhash(data, len, seed); +} + +static int ltable_obj_cmpfn(struct rhashtable_compare_arg *arg, const void *obj) +{ + struct _esp_hkey obj_key = {}; + const struct _esp_hkey *key = (const struct _esp_hkey *)arg->key; + const struct _esp_entry *eobj = (const struct _esp_entry *)obj; + u32 net_hmix = net_hash_mix(eobj->net); + + calculate_key(net_hmix, eobj->l_spi, eobj->l3num, &eobj->l_ip, + &eobj->r_ip, &obj_key); + return memcmp(key, &obj_key, sizeof(struct _esp_hkey)); +} + +static int rtable_obj_cmpfn(struct rhashtable_compare_arg *arg, const void *obj) +{ + const union nf_inet_addr any = { .in6 = IN6ADDR_ANY_INIT }; + struct _esp_hkey obj_key = {}; + const struct _esp_hkey *key = (const struct _esp_hkey *)arg->key; + const struct _esp_entry *eobj = (const struct _esp_entry *)obj; + u32 net_hmix = net_hash_mix(eobj->net); + + calculate_key(net_hmix, eobj->r_spi, eobj->l3num, &any, &eobj->r_ip, + &obj_key); + return memcmp(key, &obj_key, sizeof(struct _esp_hkey)); +} + +static int incmpl_table_obj_cmpfn(struct rhashtable_compare_arg *arg, const void *obj) +{ + const union nf_inet_addr any = { .in6 = IN6ADDR_ANY_INIT }; + struct _esp_hkey obj_key = {}; + const struct _esp_hkey *key = (const struct _esp_hkey *)arg->key; + const struct _esp_entry *eobj = (const struct _esp_entry *)obj; + u32 net_hmix = net_hash_mix(eobj->net); + + calculate_key(net_hmix, 0, eobj->l3num, &any, &eobj->r_ip, &obj_key); + return memcmp(key, &obj_key, sizeof(struct _esp_hkey)); +} + +static u32 ltable_obj_hashfn(const void *data, u32 len, u32 seed) +{ + struct _esp_hkey key = {}; + const struct _esp_entry *eobj = (const struct _esp_entry *)data; + u32 net_hmix = net_hash_mix(eobj->net); + + calculate_key(net_hmix, eobj->l_spi, eobj->l3num, &eobj->l_ip, + &eobj->r_ip, &key); + return calculate_hash(&key, len, seed); +} + +static u32 rtable_obj_hashfn(const void *data, u32 len, u32 seed) +{ + const union nf_inet_addr any = { .in6 = IN6ADDR_ANY_INIT }; + struct _esp_hkey key = {}; + const struct _esp_entry *eobj = (const struct _esp_entry *)data; + u32 net_hmix = net_hash_mix(eobj->net); + + calculate_key(net_hmix, eobj->r_spi, eobj->l3num, &any, &eobj->r_ip, &key); + return calculate_hash(&key, len, seed); +} + +static u32 incmpl_table_obj_hashfn(const void *data, u32 len, u32 seed) +{ + const union nf_inet_addr any = { .in6 = IN6ADDR_ANY_INIT }; + struct _esp_hkey key = {}; + const struct _esp_entry *eobj = (const struct _esp_entry *)data; + u32 net_hmix = net_hash_mix(eobj->net); + + calculate_key(net_hmix, 0, eobj->l3num, &any, &eobj->r_ip, &key); + return calculate_hash(&key, len, seed); +} + +static const struct rhashtable_params ltable_params = { + .key_len = sizeof(struct _esp_hkey), + .head_offset = offsetof(struct _esp_entry, lnode), + .hashfn = calculate_hash, + .obj_hashfn = ltable_obj_hashfn, + .obj_cmpfn = ltable_obj_cmpfn, +}; + +static const struct rhashtable_params rtable_params = { + .key_len = sizeof(struct _esp_hkey), + .head_offset = offsetof(struct _esp_entry, rnode), + .hashfn = calculate_hash, + .obj_hashfn = rtable_obj_hashfn, + .obj_cmpfn = rtable_obj_cmpfn, +}; + +static const struct rhashtable_params incmpl_rtable_params = { + .key_len = sizeof(struct _esp_hkey), + .head_offset = offsetof(struct _esp_entry, incmpl_rlist), + .hashfn = calculate_hash, + .obj_hashfn = incmpl_table_obj_hashfn, + .obj_cmpfn = incmpl_table_obj_cmpfn, +}; + +int nf_conntrack_esp_init(void) +{ + int ret; + + ret = rhashtable_init(&ltable, &ltable_params); + if (ret) + return ret; + + ret = rhashtable_init(&rtable, &rtable_params); + if (ret) + goto err_free_ltable; + + ret = rhltable_init(&incmpl_rtable, &incmpl_rtable_params); + if (ret) + goto err_free_rtable; + + return ret; + +err_free_rtable: + rhashtable_destroy(&rtable); +err_free_ltable: + rhashtable_destroy(&ltable); + + return ret; +} + +void nf_conntrack_esp_init_net(struct net *net) +{ + int i; + struct nf_esp_net *net_esp = esp_pernet(net); + + spin_lock_init(&net_esp->id_list_lock); + INIT_LIST_HEAD(&net_esp->id_list); + + for (i = 0; i < ESP_CT_MAX; i++) + net_esp->esp_timeouts[i] = esp_timeouts[i]; +} + +static struct _esp_entry *find_esp_entry_by_id(struct nf_esp_net *esp_net, int esp_id) +{ + struct list_head *pos, *head; + struct _esp_entry *esp_entry; + + head = &esp_net->id_list; + list_for_each(pos, head) { + esp_entry = list_entry(pos, struct _esp_entry, net_node); + if (esp_entry->esp_id == esp_id) + return esp_entry; + } + return NULL; +} + +static void free_esp_entry(struct nf_conntrack_net *cnet, struct _esp_entry *esp_entry) +{ + if (esp_entry) { + /* Remove from all the hash tables */ + pr_debug("Removing entry %x from all tables", esp_entry->esp_id); + list_del(&esp_entry->net_node); + rhashtable_remove_fast(&ltable, &esp_entry->lnode, ltable_params); + rhashtable_remove_fast(&rtable, &esp_entry->rnode, rtable_params); + rhltable_remove(&incmpl_rtable, &esp_entry->incmpl_rlist, incmpl_rtable_params); + clear_bit(esp_entry->esp_id - TEMP_SPI_START, cnet->esp_id_map); + kfree(esp_entry); + } +} + +/* Free an entry referred to by esp_id. + * + * NOTE: + * Per net linked list locking and unlocking is the responsibility of the calling function. + * Range checking is the responsibility of the calling function. + */ +static void free_esp_entry_by_id(struct net *net, int esp_id) +{ + struct nf_esp_net *esp_net = esp_pernet(net); + struct nf_conntrack_net *cnet = net_generic(net, nf_conntrack_net_id); + struct _esp_entry *esp_entry = find_esp_entry_by_id(esp_net, esp_id); + + free_esp_entry(cnet, esp_entry); +} + +/* Allocate the first available IPSEC table entry. + * NOTE: This function may block on per net list lock. + */ +struct _esp_entry *alloc_esp_entry(struct net *net) +{ + struct nf_conntrack_net *cnet = net_generic(net, nf_conntrack_net_id); + struct nf_esp_net *esp_net = esp_pernet(net); + struct _esp_entry *esp_entry; + int id; + +again: + id = find_first_zero_bit(cnet->esp_id_map, 1024); + if (id >= 1024) + return NULL; + + if (test_and_set_bit(id, cnet->esp_id_map)) + goto again; /* raced */ + + esp_entry = kmalloc(sizeof(*esp_entry), GFP_ATOMIC); + if (!esp_entry) { + clear_bit(id, cnet->esp_id_map); + return NULL; + } + + esp_entry->esp_id = id + TEMP_SPI_START; + esp_entry->alloc_time_jiffies = nfct_time_stamp; + esp_entry->net = net; + + spin_lock(&esp_net->id_list_lock); + list_add(&esp_entry->net_node, &esp_net->id_list); + spin_unlock(&esp_net->id_list_lock); + + return esp_entry; +} + +/* Search for an ESP entry in the initial state based on the IP address of + * the remote peer. + */ +static struct _esp_entry *search_esp_entry_init_remote(struct net *net, + u16 l3num, + const union nf_inet_addr *src_ip) +{ + const union nf_inet_addr any = { .in6 = IN6ADDR_ANY_INIT }; + u32 net_hmix = net_hash_mix(net); + struct _esp_entry *first_esp_entry = NULL; + struct _esp_entry *esp_entry; + struct _esp_hkey key = {}; + struct rhlist_head *pos, *list; + + calculate_key(net_hmix, 0, l3num, &any, src_ip, &key); + list = rhltable_lookup(&incmpl_rtable, (const void *)&key, incmpl_rtable_params); + rhl_for_each_entry_rcu(esp_entry, pos, list, incmpl_rlist) { + if (net_eq(net, esp_entry->net) && + l3num == esp_entry->l3num && + esp_ip_addr_equal(l3num, src_ip, &esp_entry->r_ip)) { + if (!first_esp_entry) { + first_esp_entry = esp_entry; + } else if (first_esp_entry->alloc_time_jiffies - esp_entry->alloc_time_jiffies <= 0) { + /* This entry is older than the last one found so treat this + * as a better match. + */ + first_esp_entry = esp_entry; + } + } + } + + if (first_esp_entry) { + if (first_esp_entry->l3num == AF_INET) { + pr_debug("Matches incmpl_rtable entry %x with l_spi %x r_ip %pI4\n", + first_esp_entry->esp_id, first_esp_entry->l_spi, + &first_esp_entry->r_ip.in); + } else { + pr_debug("Matches incmpl_rtable entry %x with l_spi %x r_ip %pI6\n", + first_esp_entry->esp_id, first_esp_entry->l_spi, + &first_esp_entry->r_ip.in6); + } + } + + return first_esp_entry; +} + +/* Search for an ESP entry by SPI, source and destination IP addresses. + * NOTE: This function may block on per net list lock. + */ +static struct _esp_entry *search_esp_entry_by_spi(struct net *net, const __u32 spi, + u16 l3num, + const union nf_inet_addr *src_ip, + const union nf_inet_addr *dst_ip) +{ + const union nf_inet_addr any = { .in6 = IN6ADDR_ANY_INIT }; + u32 net_hmix = net_hash_mix(net); + struct _esp_entry *esp_entry; + struct _esp_hkey key = {}; + + /* Check for matching established session or repeated initial LAN side */ + /* LAN side first */ + calculate_key(net_hmix, spi, l3num, src_ip, dst_ip, &key); + esp_entry = rhashtable_lookup_fast(&ltable, (const void *)&key, ltable_params); + if (esp_entry) { + /* When r_spi is set this is an established session. When not set it's + * a repeated initial packet from LAN side. But both cases are treated + * the same. + */ + if (esp_entry->l3num == AF_INET) { + pr_debug("Matches ltable entry %x with l_spi %x l_ip %pI4 r_ip %pI4\n", + esp_entry->esp_id, esp_entry->l_spi, + &esp_entry->l_ip.in, &esp_entry->r_ip.in); + } else { + pr_debug("Matches ltable entry %x with l_spi %x l_ip %pI6 r_ip %pI6\n", + esp_entry->esp_id, esp_entry->l_spi, + &esp_entry->l_ip.in6, &esp_entry->r_ip.in6); + } + return esp_entry; + } + + /* Established remote side */ + calculate_key(net_hmix, spi, l3num, &any, src_ip, &key); + esp_entry = rhashtable_lookup_fast(&rtable, (const void *)&key, rtable_params); + if (esp_entry) { + if (esp_entry->l3num == AF_INET) { + pr_debug("Matches rtable entry %x with l_spi %x r_spi %x l_ip %pI4 r_ip %pI4\n", + esp_entry->esp_id, esp_entry->l_spi, esp_entry->r_spi, + &esp_entry->l_ip.in, &esp_entry->r_ip.in); + } else { + pr_debug("Matches rtable entry %x with l_spi %x r_spi %x l_ip %pI6 r_ip %pI6\n", + esp_entry->esp_id, esp_entry->l_spi, esp_entry->r_spi, + &esp_entry->l_ip.in6, &esp_entry->r_ip.in6); + } + return esp_entry; + } + + /* Incomplete remote side, check if packet has a missing r_spi */ + esp_entry = search_esp_entry_init_remote(net, l3num, src_ip); + if (esp_entry) { + int err; + + esp_entry->r_spi = spi; + /* Remove entry from incmpl_rtable and add to rtable */ + rhltable_remove(&incmpl_rtable, &esp_entry->incmpl_rlist, incmpl_rtable_params); + /* Error will not be due to duplicate as established remote side lookup + * above would have found it. Delete entry. + */ + err = rhashtable_insert_fast(&rtable, &esp_entry->rnode, rtable_params); + if (err) { + struct nf_esp_net *esp_net = esp_pernet(net); + + spin_lock(&esp_net->id_list_lock); + free_esp_entry_by_id(net, esp_entry->esp_id); + spin_unlock(&esp_net->id_list_lock); + return NULL; + } + return esp_entry; + } + + if (l3num == AF_INET) { + pr_debug("No entry matches for spi %x src_ip %pI4 dst_ip %pI4\n", + spi, &src_ip->in, &dst_ip->in); + } else { + pr_debug("No entry matches for spi %x src_ip %pI6 dst_ip %pI6\n", + spi, &src_ip->in6, &dst_ip->in6); + } + return NULL; +} + +/* invert esp part of tuple */ +bool nf_conntrack_invert_esp_tuple(struct nf_conntrack_tuple *tuple, + const struct nf_conntrack_tuple *orig) +{ + tuple->dst.u.esp.id = orig->dst.u.esp.id; + tuple->src.u.esp.id = orig->src.u.esp.id; + return true; +} + +/* esp hdr info to tuple */ +bool esp_pkt_to_tuple(const struct sk_buff *skb, unsigned int dataoff, + struct net *net, struct nf_conntrack_tuple *tuple) +{ + struct esphdr _esphdr, *esphdr; + struct _esp_entry *esp_entry; + u32 spi; + + esphdr = skb_header_pointer(skb, dataoff, sizeof(_esphdr), &_esphdr); + if (!esphdr) { + /* try to behave like "nf_conntrack_proto_generic" */ + tuple->src.u.all = 0; + tuple->dst.u.all = 0; + return true; + } + spi = ntohl(esphdr->spi); + + /* Check if esphdr already associated with a pre-existing connection: + * if no, create a new connection, missing the r_spi; + * if yes, check if we have seen the source IP: + * if no, fill in r_spi in the pre-existing connection. + */ + esp_entry = search_esp_entry_by_spi(net, spi, tuple->src.l3num, + &tuple->src.u3, &tuple->dst.u3); + if (!esp_entry) { + struct _esp_hkey key = {}; + const union nf_inet_addr any = { .in6 = IN6ADDR_ANY_INIT }; + u32 net_hmix = net_hash_mix(net); + struct nf_esp_net *esp_net = esp_pernet(net); + struct _esp_entry *esp_entry_old; + int err; + + esp_entry = alloc_esp_entry(net); + if (!esp_entry) { + pr_debug("All esp connection slots in use\n"); + return false; + } + esp_entry->l_spi = spi; + esp_entry->l3num = tuple->src.l3num; + esp_ip_addr_copy(esp_entry->l3num, &esp_entry->l_ip, &tuple->src.u3); + esp_ip_addr_copy(esp_entry->l3num, &esp_entry->r_ip, &tuple->dst.u3); + + /* Add entries to the hash tables */ + + calculate_key(net_hmix, esp_entry->l_spi, esp_entry->l3num, &esp_entry->l_ip, + &esp_entry->r_ip, &key); + esp_entry_old = rhashtable_lookup_get_insert_key(&ltable, &key, &esp_entry->lnode, + ltable_params); + if (esp_entry_old) { + spin_lock(&esp_net->id_list_lock); + + if (IS_ERR(esp_entry_old)) { + free_esp_entry_by_id(net, esp_entry->esp_id); + spin_unlock(&esp_net->id_list_lock); + return false; + } + + free_esp_entry_by_id(net, esp_entry->esp_id); + spin_unlock(&esp_net->id_list_lock); + + /* insertion raced, use existing entry */ + esp_entry = esp_entry_old; + } + /* esp_entry_old == NULL -- insertion successful */ + + calculate_key(net_hmix, 0, esp_entry->l3num, &any, &esp_entry->r_ip, &key); + err = rhltable_insert_key(&incmpl_rtable, (const void *)&key, + &esp_entry->incmpl_rlist, incmpl_rtable_params); + if (err) { + spin_lock(&esp_net->id_list_lock); + free_esp_entry_by_id(net, esp_entry->esp_id); + spin_unlock(&esp_net->id_list_lock); + return false; + } + + if (esp_entry->l3num == AF_INET) { + pr_debug("New entry %x with l_spi %x l_ip %pI4 r_ip %pI4\n", + esp_entry->esp_id, esp_entry->l_spi, + &esp_entry->l_ip.in, &esp_entry->r_ip.in); + } else { + pr_debug("New entry %x with l_spi %x l_ip %pI6 r_ip %pI6\n", + esp_entry->esp_id, esp_entry->l_spi, + &esp_entry->l_ip.in6, &esp_entry->r_ip.in6); + } + } + + tuple->dst.u.esp.id = esp_entry->esp_id; + tuple->src.u.esp.id = esp_entry->esp_id; + return true; +} + +#ifdef CONFIG_NF_CONNTRACK_PROCFS +/* print private data for conntrack */ +static void esp_print_conntrack(struct seq_file *s, struct nf_conn *ct) +{ + seq_printf(s, "l_spi=%x, r_spi=%x ", ct->proto.esp.l_spi, ct->proto.esp.r_spi); +} +#endif + +/* Returns verdict for packet, and may modify conntrack */ +int nf_conntrack_esp_packet(struct nf_conn *ct, struct sk_buff *skb, + unsigned int dataoff, + enum ip_conntrack_info ctinfo, + const struct nf_hook_state *state) +{ + int esp_id; + struct nf_conntrack_tuple *tuple; + unsigned int *timeouts = nf_ct_timeout_lookup(ct); + struct nf_esp_net *esp_net = esp_pernet(nf_ct_net(ct)); + + if (!timeouts) + timeouts = esp_net->esp_timeouts; + + /* If we've seen traffic both ways, this is some kind of ESP + * stream. Extend timeout. + */ + if (test_bit(IPS_SEEN_REPLY_BIT, &ct->status)) { + nf_ct_refresh_acct(ct, ctinfo, skb, timeouts[ESP_CT_REPLIED]); + /* Also, more likely to be important, and not a probe */ + if (!test_and_set_bit(IPS_ASSURED_BIT, &ct->status)) { + /* Was originally IPCT_STATUS but this is no longer an option. + * GRE uses assured for same purpose + */ + nf_conntrack_event_cache(IPCT_ASSURED, ct); + + /* Retrieve SPIs of original and reply from esp_entry. + * Both directions should contain the same esp_entry, + * so just check the first one. + */ + tuple = nf_ct_tuple(ct, IP_CT_DIR_ORIGINAL); + + esp_id = tuple->src.u.esp.id; + if (esp_id >= TEMP_SPI_START && esp_id <= TEMP_SPI_MAX) { + struct _esp_entry *esp_entry; + + spin_lock(&esp_net->id_list_lock); + esp_entry = find_esp_entry_by_id(esp_net, esp_id); + spin_unlock(&esp_net->id_list_lock); + + if (esp_entry) { + ct->proto.esp.l_spi = esp_entry->l_spi; + ct->proto.esp.r_spi = esp_entry->r_spi; + } + } + } + } else { + nf_ct_refresh_acct(ct, ctinfo, skb, timeouts[ESP_CT_UNREPLIED]); + } + + return NF_ACCEPT; +} + +void nf_ct_esp_pernet_flush(struct net *net) +{ + struct nf_conntrack_net *cnet = net_generic(net, nf_conntrack_net_id); + struct nf_esp_net *esp_net = esp_pernet(net); + struct list_head *pos, *tmp, *head = &esp_net->id_list; + struct _esp_entry *esp_entry; + + spin_lock(&esp_net->id_list_lock); + list_for_each_safe(pos, tmp, head) { + esp_entry = list_entry(pos, struct _esp_entry, net_node); + free_esp_entry(cnet, esp_entry); + } + spin_unlock(&esp_net->id_list_lock); +} + +/* Called when a conntrack entry has already been removed from the hashes + * and is about to be deleted from memory + */ +void destroy_esp_conntrack_entry(struct nf_conn *ct) +{ + struct nf_conntrack_tuple *tuple; + enum ip_conntrack_dir dir; + int esp_id; + struct net *net = nf_ct_net(ct); + struct nf_esp_net *esp_net = esp_pernet(net); + + /* Probably all the ESP entries referenced in this connection are the same, + * but the free function handles repeated frees, so best to do them all. + */ + for (dir = IP_CT_DIR_ORIGINAL; dir < IP_CT_DIR_MAX; dir++) { + tuple = nf_ct_tuple(ct, dir); + + spin_lock(&esp_net->id_list_lock); + + esp_id = tuple->src.u.esp.id; + if (esp_id >= TEMP_SPI_START && esp_id <= TEMP_SPI_MAX) + free_esp_entry_by_id(net, esp_id); + tuple->src.u.esp.id = 0; + + esp_id = tuple->dst.u.esp.id; + if (esp_id >= TEMP_SPI_START && esp_id <= TEMP_SPI_MAX) + free_esp_entry_by_id(net, esp_id); + tuple->dst.u.esp.id = 0; + + spin_unlock(&esp_net->id_list_lock); + } +} + +#if IS_ENABLED(CONFIG_NF_CT_NETLINK) + +#include <linux/netfilter/nfnetlink.h> +#include <linux/netfilter/nfnetlink_conntrack.h> + +static int esp_tuple_to_nlattr(struct sk_buff *skb, + const struct nf_conntrack_tuple *t) +{ + if (nla_put_be16(skb, CTA_PROTO_SRC_ESP_ID, t->src.u.esp.id) || + nla_put_be16(skb, CTA_PROTO_DST_ESP_ID, t->dst.u.esp.id)) + goto nla_put_failure; + return 0; + +nla_put_failure: + return -1; +} + +static const struct nla_policy esp_nla_policy[CTA_PROTO_MAX + 1] = { + [CTA_PROTO_SRC_ESP_ID] = { .type = NLA_U16 }, + [CTA_PROTO_DST_ESP_ID] = { .type = NLA_U16 }, +}; + +static int esp_nlattr_to_tuple(struct nlattr *tb[], + struct nf_conntrack_tuple *t, + u32 flags) +{ + if (flags & CTA_FILTER_FLAG(CTA_PROTO_SRC_ESP_ID)) { + if (!tb[CTA_PROTO_SRC_ESP_ID]) + return -EINVAL; + + t->src.u.esp.id = nla_get_be16(tb[CTA_PROTO_SRC_ESP_ID]); + } + + if (flags & CTA_FILTER_FLAG(CTA_PROTO_DST_ESP_ID)) { + if (!tb[CTA_PROTO_DST_ESP_ID]) + return -EINVAL; + + t->dst.u.esp.id = nla_get_be16(tb[CTA_PROTO_DST_ESP_ID]); + } + + return 0; +} + +static unsigned int esp_nlattr_tuple_size(void) +{ + return nla_policy_len(esp_nla_policy, CTA_PROTO_MAX + 1); +} +#endif + +/* protocol helper struct */ +const struct nf_conntrack_l4proto nf_conntrack_l4proto_esp = { + .l4proto = IPPROTO_ESP, +#ifdef CONFIG_NF_CONNTRACK_PROCFS + .print_conntrack = esp_print_conntrack, +#endif +#if IS_ENABLED(CONFIG_NF_CT_NETLINK) + .tuple_to_nlattr = esp_tuple_to_nlattr, + .nlattr_tuple_size = esp_nlattr_tuple_size, + .nlattr_to_tuple = esp_nlattr_to_tuple, + .nla_policy = esp_nla_policy, +#endif +}; diff --git a/net/netfilter/nf_conntrack_standalone.c b/net/netfilter/nf_conntrack_standalone.c index c6c0cb465664..7922ff6cf5a4 100644 --- a/net/netfilter/nf_conntrack_standalone.c +++ b/net/netfilter/nf_conntrack_standalone.c @@ -88,6 +88,14 @@ print_tuple(struct seq_file *s, const struct nf_conntrack_tuple *tuple, ntohs(tuple->src.u.gre.key), ntohs(tuple->dst.u.gre.key)); break; + case IPPROTO_ESP: + /* Both src and dest esp.id should be equal but showing both + * will help find errors. + */ + seq_printf(s, "srcid=0x%x dstid=0x%x ", + ntohs(tuple->src.u.esp.id), + ntohs(tuple->dst.u.esp.id)); + break; default: break; } diff --git a/net/netfilter/nf_internals.h b/net/netfilter/nf_internals.h index 832ae64179f0..4fd8956aec65 100644 --- a/net/netfilter/nf_internals.h +++ b/net/netfilter/nf_internals.h @@ -19,7 +19,9 @@ #define CTA_FILTER_F_CTA_PROTO_ICMPV6_TYPE (1 << 9) #define CTA_FILTER_F_CTA_PROTO_ICMPV6_CODE (1 << 10) #define CTA_FILTER_F_CTA_PROTO_ICMPV6_ID (1 << 11) -#define CTA_FILTER_F_MAX (1 << 12) +#define CTA_FILTER_F_CTA_PROTO_SRC_ESP_ID (1 << 12) +#define CTA_FILTER_F_CTA_PROTO_DST_ESP_ID (1 << 13) +#define CTA_FILTER_F_MAX (1 << 14) #define CTA_FILTER_F_ALL (CTA_FILTER_F_MAX-1) #define CTA_FILTER_FLAG(ctattr) CTA_FILTER_F_ ## ctattr diff --git a/tools/testing/selftests/netfilter/conntrack_esp_related.sh b/tools/testing/selftests/netfilter/conntrack_esp_related.sh new file mode 100755 index 000000000000..88b0f164664f --- /dev/null +++ b/tools/testing/selftests/netfilter/conntrack_esp_related.sh @@ -0,0 +1,268 @@ +#!/bin/bash +# SPDX-License-Identifier: GPL-2.0 +# +# <:copyright-gpl +# Copyright (C) 2021 Allied Telesis Labs NZ +# +# check that related ESP connections are tracked via spi. +# +# Setup is: +# +# nsclient3(veth0) -> (veth2) +# (br0)nsrouter1(veth1) -> (veth1)nsrouter2 -> (veth0)nsclient2 +# nsclient1(veth0) -> (veth0) +# Setup xfrm esp connections for IPv4 and IPv6 and check they are tracked. +# +# In addition, nsrouter1 will perform IP masquerading. If nsrouter1 does not support esp +# connection tracking, it will be unable to tell the difference between packets from nsclient2 to +# either nsclient1 or nsclient3. +# +# ESP connections (for IPv6) need to use tunnel mode, as ICMPv6 computes checksum over encapsulating +# IP header addresses. + +# Kselftest framework requirement - SKIP code is 4. +ksft_skip=4 +ret=0 +ns_all="nsclient1 nsclient3 nsrouter1 nsrouter2 nsclient2" + +conntrack -V > /dev/null 2>&1 +if [ $? -ne 0 ];then + echo "SKIP: Could not run test without conntrack tool" + exit $ksft_skip +fi + +nft --version > /dev/null 2>&1 +if [ $? -ne 0 ];then + echo "SKIP: Could not run test without nft tool" + exit $ksft_skip +fi + +ip -Version > /dev/null 2>&1 +if [ $? -ne 0 ];then + echo "SKIP: Could not run test without ip tool" + exit $ksft_skip +fi + +ipv4() { + echo -n 192.168.$1.$2 +} + +ipv6 () { + echo -n dead:$1::$2 +} + +cleanup() { + for n in $ns_all; do ip netns del $n;done +} + +check_counter() +{ + local ns_name=$1 + local name="unknown" + local expect="packets 0 bytes 0" + local lret=0 + + cnt=$(ip netns exec $ns_name nft list counter inet filter "$name" | grep -q "$expect") + if [ $? -ne 0 ]; then + echo "ERROR: counter $name in $ns_name has unexpected value (expected $expect)" 1>&2 + ip netns exec $ns_name nft list counter inet filter "$name" 1>&2 + lret=1 + fi + return $lret +} + +check_unknown() +{ + for n in nsrouter1 nsrouter2; do + check_counter $n + if [ $? -ne 0 ] ;then + return 1 + fi + done + return 0 +} + +check_conntrack() +{ + local ret=0 + + for p in ipv4 ipv6; do + cnt=$(ip netns exec nsrouter1 conntrack -f $p -L 2>&1) + # Check tracked connection was esp by port (conntrack shows unknown at the moment) + local num=$(echo -e "$cnt" | grep -cE "[a-zA-Z]+ +50") + if [ $? -ne 0 ] || [ "x$num" != "x2" ]; then + echo -e "ERROR: expect to see two conntrack esp flows for $p:\n $cnt" 1>&2 + ret=1 + fi + done + return $ret +} + +for n in $ns_all; do + ip netns add $n + ip -net $n link set lo up +done + +ip link add veth0 netns nsclient1 type veth peer name veth0 netns nsrouter1 +ip link add veth0 netns nsclient3 type veth peer name veth2 netns nsrouter1 +ip link add br0 netns nsrouter1 type bridge +ip -net nsrouter1 link set veth0 master br0 +ip -net nsrouter1 link set veth2 master br0 +ip link add veth1 netns nsrouter1 type veth peer name veth1 netns nsrouter2 +ip link add veth0 netns nsrouter2 type veth peer name veth0 netns nsclient2 + +for n in $ns_all; do + ip -net $n link set veth0 up +done +ip -net nsrouter1 link set veth1 up +ip -net nsrouter1 link set veth2 up +ip -net nsrouter1 link set br0 up +ip -net nsrouter2 link set veth1 up + +for i in 1 2; do + ip -net nsclient$i addr add $(ipv4 $i 2)/24 dev veth0 + ip -net nsclient$i addr add $(ipv6 $i 2)/64 dev veth0 + ip -net nsclient$i route add default via $(ipv4 $i 1) + ip -net nsclient$i -6 route add default via $(ipv6 $i 1) + + ip -net nsrouter$i addr add $(ipv4 3 $i)/24 dev veth1 + ip -net nsrouter$i addr add $(ipv6 3 $i)/64 dev veth1 +done +ip -net nsrouter1 addr add $(ipv4 1 1)/24 dev br0 +ip -net nsrouter1 addr add $(ipv6 1 1)/64 dev br0 +ip -net nsrouter2 addr add $(ipv4 2 1)/24 dev veth0 +ip -net nsrouter2 addr add $(ipv6 2 1)/64 dev veth0 + +ip -net nsclient3 addr add $(ipv4 1 3)/24 dev veth0 +ip -net nsclient3 addr add $(ipv6 1 3)/64 dev veth0 +ip -net nsclient3 route add default via $(ipv4 1 1) +ip -net nsclient3 -6 route add default via $(ipv6 1 1) + +ip -net nsrouter1 route add default via $(ipv4 3 2) +ip -net nsrouter1 -6 route add default via $(ipv6 3 2) +ip -net nsrouter2 route add default via $(ipv4 3 1) +ip -net nsrouter2 -6 route add default via $(ipv6 3 1) + +for i in 1 2; do + ip netns exec nsrouter$i sysctl -q net.ipv4.conf.all.forwarding=1 + ip netns exec nsrouter$i sysctl -q net.ipv6.conf.all.forwarding=1 +done + +for i in 1 2; do + ip netns exec nsrouter$i nft -f - <<-EOF + table inet filter { + counter unknown { } + chain forward { + type filter hook forward priority 0; policy accept; + meta l4proto esp ct state new,established accept + counter name "unknown" accept + } + } + EOF +done + +for i in 1 2; do + ip netns exec nsrouter1 nft -f - <<-EOF + table ip nat { + chain postrouting { + type nat hook postrouting priority 0; policy accept; + oifname "veth1" counter masquerade + } + } + table ip6 nat { + chain postrouting { + type nat hook postrouting priority 0; policy accept; + oifname "veth1" counter masquerade + } + } + EOF +done +sleep 2 + +ip_tunnel() { + ip -net nsclient$2 tunnel add tunnel$1 mode vti${1%4} local $3 remote $4 key 0x$1 + ip -net nsclient$2 link set tunnel$1 up +} + +ip_xfrm() { + ip -net nsclient$2 xfrm state add src $4 dst $5 \ + proto esp spi 0x$1$2$3 mode tunnel mark 0x$1 \ + sel src $6 dst $7 \ + auth-trunc 'hmac(sha256)' \ + 0x0000000000000000000000000000000000000000000000000000000000000$1$2$3 128 \ + enc 'cbc(aes)' \ + 0x0000000000000000000000000000000000000000000000000000000000000$1$2$3 + + ip -net nsclient$2 xfrm state add src $5 dst $4 \ + proto esp spi 0x$1$3$2 mode tunnel mark 0x$1 \ + sel src $7 dst $6 \ + auth-trunc 'hmac(sha256)' \ + 0x0000000000000000000000000000000000000000000000000000000000000$1$3$2 128 \ + enc 'cbc(aes)' \ + 0x0000000000000000000000000000000000000000000000000000000000000$1$3$2 + + ip -net nsclient$2 xfrm policy add src $7 dst $6 dir in mark 0x$1 \ + tmpl src $5 dst $4 proto esp mode tunnel + ip -net nsclient$2 xfrm policy add src $6 dst $7 dir out mark 0x$1 \ + tmpl src $4 dst $5 proto esp mode tunnel +} + +ip_tunnel 4 1 $(ipv4 1 2) $(ipv4 2 2) +ip -net nsclient1 addr add $(ipv4 250 1)/24 dev tunnel4 +ip_xfrm 4 1 2 $(ipv4 1 2) $(ipv4 2 2) $(ipv4 250 1) $(ipv4 250 2) + +ip_tunnel 4 3 $(ipv4 1 3) $(ipv4 2 2) +ip -net nsclient3 addr add $(ipv4 251 1)/24 dev tunnel4 +ip_xfrm 4 3 2 $(ipv4 1 3) $(ipv4 2 2) $(ipv4 251 1) $(ipv4 251 2) + +ip_tunnel 4 2 $(ipv4 2 2) $(ipv4 3 1) +ip -net nsclient2 addr add $(ipv4 250 2)/24 dev tunnel4 +ip -net nsclient2 addr add $(ipv4 251 2)/24 dev tunnel4 +ip_xfrm 4 2 1 $(ipv4 2 2) $(ipv4 3 1) $(ipv4 250 2) $(ipv4 250 1) +ip_xfrm 4 2 3 $(ipv4 2 2) $(ipv4 3 1) $(ipv4 251 2) $(ipv4 251 1) + + +ip_tunnel 6 1 $(ipv6 1 2) $(ipv6 2 2) +ip -net nsclient1 addr add $(ipv6 250 1)/64 dev tunnel6 +ip_xfrm 6 1 2 $(ipv6 1 2) $(ipv6 2 2) $(ipv6 250 1) $(ipv6 250 2) + +ip_tunnel 6 3 $(ipv6 1 3) $(ipv6 2 2) +ip -net nsclient3 addr add $(ipv6 251 1)/64 dev tunnel6 +ip_xfrm 6 3 2 $(ipv6 1 3) $(ipv6 2 2) $(ipv6 251 1) $(ipv6 251 2) + +ip_tunnel 6 2 $(ipv6 2 2) $(ipv6 3 1) +ip -net nsclient2 addr add $(ipv6 250 2)/64 dev tunnel6 +ip -net nsclient2 addr add $(ipv6 251 2)/64 dev tunnel6 +ip_xfrm 6 2 1 $(ipv6 2 2) $(ipv6 3 1) $(ipv6 250 2) $(ipv6 250 1) +ip_xfrm 6 2 3 $(ipv6 2 2) $(ipv6 3 1) $(ipv6 251 2) $(ipv6 251 1) + +test_ping() { + ip netns exec $1 ping -q -c 1 $2 >/dev/null 2>&1 + if [ $? -ne 0 ]; then + echo "ERROR: netns ip routing/connectivity broken from $1 to $2" 1>&2 + fi +} + +test_ping nsclient1 $(ipv4 250 2) +test_ping nsclient3 $(ipv4 251 2) +test_ping nsclient1 $(ipv6 250 2) +test_ping nsclient3 $(ipv6 251 2) + +check_conntrack +if [ $? -ne 0 ]; then + ret=1 +fi + +check_unknown +if [ $? -ne 0 ]; then + ret=1 +fi + +if [ $ret -eq 0 ];then + echo "PASS: ESP connections were tracked via SPIs" +else + echo "ERROR: ESP connections were not tracked" +fi + +cleanup +exit $ret -- 2.31.1

4 years, 2 months

2
2
0 0

[RFC v2 0/4] kunit: tool: add support for QEMU

by Brendan Higgins

TL;DR: Add support to kunit_tool to dispatch tests via QEMU. Also add support to immediately shutdown a kernel after running KUnit tests. Background ---------- KUnit has supported running on all architectures for quite some time; however, kunit_tool - the script commonly used to invoke KUnit tests - has only fully supported KUnit run on UML. Its functionality has been broken up for some time to separate the configure, build, run, and parse phases making it possible to be used in part on other architectures to a small extent. Nevertheless, kunit_tool has not supported running tests on other architectures. What this patchset does ----------------------- This patchset introduces first class support to kunit_tool for KUnit to be run on many popular architectures via QEMU. It does this by adding two new flags: `--arch` and `--cross_compile`. `--arch` allows an architecture to be specified by the name the architecture is given in `arch/`. It uses the specified architecture to select a minimal amount of Kconfigs and QEMU configs needed for the architecture to run in QEMU and provide a console from which KTAP results can be scraped. `--cross_compile` allows a toolchain prefix to be specified to make similar to how `CROSS_COMPILE` is used. Additionally, this patchset revives the previously considered "kunit: tool: add support for QEMU"[1] patchs. The motivation for this new kernel command line flags, `kunit_shutdown`, is to better support running KUnit tests inside of QEMU. For most popular architectures, QEMU can be made to terminate when the Linux kernel that is being run is reboted, halted, or powered off. As Kees pointed out in a previous discussion[2], it is possible to make a kernel initrd that can reboot the kernel immediately, doing this for every architecture would likely be infeasible. Instead, just having an option for the kernel to shutdown when it is done with testing seems a lot simpler, especially since it is an option which would only available in testing configurations of the kernel anyway. What discussion remains for this patchset? ------------------------------------------ The first most obvious thing is settling the debate about `kunit_shutdown`. If I recall correctly, Kees suggested that it might be better to just add a new initrd; however, as I mentioned above, now to support many new architectures, it may be substantially easier to support this option. So I am hoping with this new usecase, the argument for `kunit_shutdown` will be more compelling. The second and likely harder issue is figuring out the best way to configure and provide configs for running KUnit tests via QEMU. I provide a pretty primitive way in this patchset which is not super flexible; for example, for our PPC support we have it set to build big endian, and POWER8 - we currently don't support a way to change that. Nevertheless, having sensible defaults is handy too, so we will probably want to have some support for overriding defaults, while still being able to have defaults. [1] http://patches.linaro.org/patch/208336/ [2] https://lkml.org/lkml/2020/6/26/988 Brendan Higgins (3): Documentation: Add kunit_shutdown to kernel-parameters.txt kunit: tool: add support for QEMU Documentation: kunit: document support for QEMU in kunit_tool David Gow (1): kunit: Add 'kunit_shutdown' option .../admin-guide/kernel-parameters.txt | 8 + Documentation/dev-tools/kunit/usage.rst | 37 +++- lib/kunit/executor.c | 20 ++ tools/testing/kunit/kunit.py | 33 ++- tools/testing/kunit/kunit_config.py | 2 +- tools/testing/kunit/kunit_kernel.py | 209 +++++++++++++++--- tools/testing/kunit/kunit_parser.py | 2 +- tools/testing/kunit/kunit_tool_test.py | 15 +- 8 files changed, 278 insertions(+), 48 deletions(-) base-commit: 7af08140979a6e7e12b78c93b8625c8d25b084e2 -- 2.31.1.498.g6c1eba8ee3d-goog

4 years, 2 months

3
11
0 0

[PATCH] lib/test: convert test_overflow.c to use KUnit

by Daniel Latypov

This largely consists of: * replacing `err |= test_func(...)` calls with `test_func(test, ...)`. * changing pr_warn to KUNIT_*_MSG() macros to mark the failed test while printing a diagnostic message * dropping __init as there are concerns of this causing issues w/ KUnit when running on some other arches. With the exception of now needing to set CONFIG_KUNIT=y, the test can still be run as before, at boot time or as a test module. But now the output of the test will be TAP (Test Anything Protocol), the format used by kselftest instead of just warnings and printing "FAIL!" , see [1]. But now the test can also be run more quickly via kunit.py $ ./tools/testing/kunit/kunit.py run --kunitconfig /dev/stdin <<EOF CONFIG_KUNIT=y CONFIG_TEST_OVERFLOW=y EOF Note: after a mrproper this took > Elapsed time: 38.683s total, 3.598s configuring, 32.141s building, 0.000s running (Normally would be less clunky as you wouldn't use a heredoc). [1] Here's an example failure message after tweaking the test array: TAP version 14 1..1 # Subtest: lib_overflow 1..10 u8 : 18 arithmetic tests # test_u8_overflow: EXPECTATION FAILED at lib/test_overflow.c:250 Expected _got == p->sum, but _got == 1 p->sum == 0 wrong result for 1 + 0 (type u8) ... not ok 1 - test_u8_overflow s8 : 19 arithmetic tests ok 2 - test_s8_overflow ... ok: (s32)(0 << 31) == 0 ok: (s64)(0 << 63) == 0 ok 9 - test_overflow_shift Signed-off-by: Daniel Latypov <dlatypov(a)google.com> --- lib/Kconfig.debug | 4 +- lib/test_overflow.c | 406 ++++++++++++++++++++------------------------ 2 files changed, 185 insertions(+), 225 deletions(-) diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 678c13967580..8e19ba9ac22c 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -2195,7 +2195,9 @@ config TEST_XARRAY tristate "Test the XArray code at runtime" config TEST_OVERFLOW - tristate "Test check_*_overflow() functions at runtime" + tristate "Test check_*_overflow() functions at runtime" if !KUNIT_ALL_TESTS + depends on KUNIT + default KUNIT_ALL_TESTS config TEST_RHASHTABLE tristate "Perform selftest on resizable hash table" diff --git a/lib/test_overflow.c b/lib/test_overflow.c index 7a4b6f6c5473..1719b924f04f 100644 --- a/lib/test_overflow.c +++ b/lib/test_overflow.c @@ -2,10 +2,9 @@ /* * Test cases for arithmetic overflow checks. */ -#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt +#include <kunit/test.h> #include <linux/device.h> -#include <linux/init.h> #include <linux/kernel.h> #include <linux/mm.h> #include <linux/module.h> @@ -19,7 +18,7 @@ t a, b; \ t sum, diff, prod; \ bool s_of, d_of, p_of; \ - } t ## _tests[] __initconst + } t ## _tests[] DEFINE_TEST_ARRAY(u8) = { {0, 0, 0, 0, 0, false, false, false}, @@ -215,48 +214,38 @@ DEFINE_TEST_ARRAY(s64) = { {0, -S64_MAX, -S64_MAX, S64_MAX, 0, false, false, false}, }; -#define check_one_op(t, fmt, op, sym, a, b, r, of) do { \ - t _r; \ - bool _of; \ - \ - _of = check_ ## op ## _overflow(a, b, &_r); \ - if (_of != of) { \ - pr_warn("expected "fmt" "sym" "fmt \ - " to%s overflow (type %s)\n", \ - a, b, of ? "" : " not", #t); \ - err = 1; \ - } \ - if (_r != r) { \ - pr_warn("expected "fmt" "sym" "fmt" == " \ - fmt", got "fmt" (type %s)\n", \ - a, b, r, _r, #t); \ - err = 1; \ - } \ +#define check_one_op(t, fmt, op, sym, a, b, want, of) do { \ + t _got; \ + bool _of; \ + \ + _of = check_ ## op ## _overflow(a, b, &_got); \ + KUNIT_EXPECT_EQ_MSG(test, _of, of, \ + "expected " fmt " " sym " " fmt " to%s overflow " \ + "(type %s)", a, b, of ? "" : " not", #t); \ + \ + KUNIT_EXPECT_EQ_MSG(test, _got, want, \ + "wrong result for " fmt " " sym " " fmt " (type %s)", \ + a, b, #t); \ } while (0) #define DEFINE_TEST_FUNC(t, fmt) \ -static int __init do_test_ ## t(const struct test_ ## t *p) \ -{ \ - int err = 0; \ - \ +static void do_test_ ## t(struct kunit *test, const struct test_ ## t *p)\ +{ \ check_one_op(t, fmt, add, "+", p->a, p->b, p->sum, p->s_of); \ check_one_op(t, fmt, add, "+", p->b, p->a, p->sum, p->s_of); \ check_one_op(t, fmt, sub, "-", p->a, p->b, p->diff, p->d_of); \ check_one_op(t, fmt, mul, "*", p->a, p->b, p->prod, p->p_of); \ check_one_op(t, fmt, mul, "*", p->b, p->a, p->prod, p->p_of); \ - \ - return err; \ } \ \ -static int __init test_ ## t ## _overflow(void) { \ - int err = 0; \ +static void test_ ## t ## _overflow(struct kunit *test) \ +{ \ unsigned i; \ \ pr_info("%-3s: %zu arithmetic tests\n", #t, \ ARRAY_SIZE(t ## _tests)); \ for (i = 0; i < ARRAY_SIZE(t ## _tests); ++i) \ - err |= do_test_ ## t(&t ## _tests[i]); \ - return err; \ + do_test_ ## t(test, &t ## _tests[i]); \ } DEFINE_TEST_FUNC(u8, "%d"); @@ -270,30 +259,10 @@ DEFINE_TEST_FUNC(u64, "%llu"); DEFINE_TEST_FUNC(s64, "%lld"); #endif -static int __init test_overflow_calculation(void) +static void test_overflow_shift(struct kunit *test) { - int err = 0; - - err |= test_u8_overflow(); - err |= test_s8_overflow(); - err |= test_u16_overflow(); - err |= test_s16_overflow(); - err |= test_u32_overflow(); - err |= test_s32_overflow(); -#if BITS_PER_LONG == 64 - err |= test_u64_overflow(); - err |= test_s64_overflow(); -#endif - - return err; -} - -static int __init test_overflow_shift(void) -{ - int err = 0; - /* Args are: value, shift, type, expected result, overflow expected */ -#define TEST_ONE_SHIFT(a, s, t, expect, of) ({ \ +#define TEST_ONE_SHIFT(a, s, t, expect, of) do { \ int __failed = 0; \ typeof(a) __a = (a); \ typeof(s) __s = (s); \ @@ -316,153 +285,153 @@ static int __init test_overflow_shift(void) if (!__failed) \ pr_info("ok: (%s)(%s << %s) == %s\n", #t, #a, #s, \ of ? "overflow" : #expect); \ - __failed; \ -}) + KUNIT_EXPECT_EQ_MSG(test, __failed, 0, \ + "(%s)(%s << %s) failed", #t, #a, #s); \ +} while (0) /* Sane shifts. */ - err |= TEST_ONE_SHIFT(1, 0, u8, 1 << 0, false); - err |= TEST_ONE_SHIFT(1, 4, u8, 1 << 4, false); - err |= TEST_ONE_SHIFT(1, 7, u8, 1 << 7, false); - err |= TEST_ONE_SHIFT(0xF, 4, u8, 0xF << 4, false); - err |= TEST_ONE_SHIFT(1, 0, u16, 1 << 0, false); - err |= TEST_ONE_SHIFT(1, 10, u16, 1 << 10, false); - err |= TEST_ONE_SHIFT(1, 15, u16, 1 << 15, false); - err |= TEST_ONE_SHIFT(0xFF, 8, u16, 0xFF << 8, false); - err |= TEST_ONE_SHIFT(1, 0, int, 1 << 0, false); - err |= TEST_ONE_SHIFT(1, 16, int, 1 << 16, false); - err |= TEST_ONE_SHIFT(1, 30, int, 1 << 30, false); - err |= TEST_ONE_SHIFT(1, 0, s32, 1 << 0, false); - err |= TEST_ONE_SHIFT(1, 16, s32, 1 << 16, false); - err |= TEST_ONE_SHIFT(1, 30, s32, 1 << 30, false); - err |= TEST_ONE_SHIFT(1, 0, unsigned int, 1U << 0, false); - err |= TEST_ONE_SHIFT(1, 20, unsigned int, 1U << 20, false); - err |= TEST_ONE_SHIFT(1, 31, unsigned int, 1U << 31, false); - err |= TEST_ONE_SHIFT(0xFFFFU, 16, unsigned int, 0xFFFFU << 16, false); - err |= TEST_ONE_SHIFT(1, 0, u32, 1U << 0, false); - err |= TEST_ONE_SHIFT(1, 20, u32, 1U << 20, false); - err |= TEST_ONE_SHIFT(1, 31, u32, 1U << 31, false); - err |= TEST_ONE_SHIFT(0xFFFFU, 16, u32, 0xFFFFU << 16, false); - err |= TEST_ONE_SHIFT(1, 0, u64, 1ULL << 0, false); - err |= TEST_ONE_SHIFT(1, 40, u64, 1ULL << 40, false); - err |= TEST_ONE_SHIFT(1, 63, u64, 1ULL << 63, false); - err |= TEST_ONE_SHIFT(0xFFFFFFFFULL, 32, u64, - 0xFFFFFFFFULL << 32, false); + TEST_ONE_SHIFT(1, 0, u8, 1 << 0, false); + TEST_ONE_SHIFT(1, 4, u8, 1 << 4, false); + TEST_ONE_SHIFT(1, 7, u8, 1 << 7, false); + TEST_ONE_SHIFT(0xF, 4, u8, 0xF << 4, false); + TEST_ONE_SHIFT(1, 0, u16, 1 << 0, false); + TEST_ONE_SHIFT(1, 10, u16, 1 << 10, false); + TEST_ONE_SHIFT(1, 15, u16, 1 << 15, false); + TEST_ONE_SHIFT(0xFF, 8, u16, 0xFF << 8, false); + TEST_ONE_SHIFT(1, 0, int, 1 << 0, false); + TEST_ONE_SHIFT(1, 16, int, 1 << 16, false); + TEST_ONE_SHIFT(1, 30, int, 1 << 30, false); + TEST_ONE_SHIFT(1, 0, s32, 1 << 0, false); + TEST_ONE_SHIFT(1, 16, s32, 1 << 16, false); + TEST_ONE_SHIFT(1, 30, s32, 1 << 30, false); + TEST_ONE_SHIFT(1, 0, unsigned int, 1U << 0, false); + TEST_ONE_SHIFT(1, 20, unsigned int, 1U << 20, false); + TEST_ONE_SHIFT(1, 31, unsigned int, 1U << 31, false); + TEST_ONE_SHIFT(0xFFFFU, 16, unsigned int, 0xFFFFU << 16, false); + TEST_ONE_SHIFT(1, 0, u32, 1U << 0, false); + TEST_ONE_SHIFT(1, 20, u32, 1U << 20, false); + TEST_ONE_SHIFT(1, 31, u32, 1U << 31, false); + TEST_ONE_SHIFT(0xFFFFU, 16, u32, 0xFFFFU << 16, false); + TEST_ONE_SHIFT(1, 0, u64, 1ULL << 0, false); + TEST_ONE_SHIFT(1, 40, u64, 1ULL << 40, false); + TEST_ONE_SHIFT(1, 63, u64, 1ULL << 63, false); + TEST_ONE_SHIFT(0xFFFFFFFFULL, 32, u64, 0xFFFFFFFFULL << 32, false); /* Sane shift: start and end with 0, without a too-wide shift. */ - err |= TEST_ONE_SHIFT(0, 7, u8, 0, false); - err |= TEST_ONE_SHIFT(0, 15, u16, 0, false); - err |= TEST_ONE_SHIFT(0, 31, unsigned int, 0, false); - err |= TEST_ONE_SHIFT(0, 31, u32, 0, false); - err |= TEST_ONE_SHIFT(0, 63, u64, 0, false); + TEST_ONE_SHIFT(0, 7, u8, 0, false); + TEST_ONE_SHIFT(0, 15, u16, 0, false); + TEST_ONE_SHIFT(0, 31, unsigned int, 0, false); + TEST_ONE_SHIFT(0, 31, u32, 0, false); + TEST_ONE_SHIFT(0, 63, u64, 0, false); /* Sane shift: start and end with 0, without reaching signed bit. */ - err |= TEST_ONE_SHIFT(0, 6, s8, 0, false); - err |= TEST_ONE_SHIFT(0, 14, s16, 0, false); - err |= TEST_ONE_SHIFT(0, 30, int, 0, false); - err |= TEST_ONE_SHIFT(0, 30, s32, 0, false); - err |= TEST_ONE_SHIFT(0, 62, s64, 0, false); + TEST_ONE_SHIFT(0, 6, s8, 0, false); + TEST_ONE_SHIFT(0, 14, s16, 0, false); + TEST_ONE_SHIFT(0, 30, int, 0, false); + TEST_ONE_SHIFT(0, 30, s32, 0, false); + TEST_ONE_SHIFT(0, 62, s64, 0, false); /* Overflow: shifted the bit off the end. */ - err |= TEST_ONE_SHIFT(1, 8, u8, 0, true); - err |= TEST_ONE_SHIFT(1, 16, u16, 0, true); - err |= TEST_ONE_SHIFT(1, 32, unsigned int, 0, true); - err |= TEST_ONE_SHIFT(1, 32, u32, 0, true); - err |= TEST_ONE_SHIFT(1, 64, u64, 0, true); + TEST_ONE_SHIFT(1, 8, u8, 0, true); + TEST_ONE_SHIFT(1, 16, u16, 0, true); + TEST_ONE_SHIFT(1, 32, unsigned int, 0, true); + TEST_ONE_SHIFT(1, 32, u32, 0, true); + TEST_ONE_SHIFT(1, 64, u64, 0, true); /* Overflow: shifted into the signed bit. */ - err |= TEST_ONE_SHIFT(1, 7, s8, 0, true); - err |= TEST_ONE_SHIFT(1, 15, s16, 0, true); - err |= TEST_ONE_SHIFT(1, 31, int, 0, true); - err |= TEST_ONE_SHIFT(1, 31, s32, 0, true); - err |= TEST_ONE_SHIFT(1, 63, s64, 0, true); + TEST_ONE_SHIFT(1, 7, s8, 0, true); + TEST_ONE_SHIFT(1, 15, s16, 0, true); + TEST_ONE_SHIFT(1, 31, int, 0, true); + TEST_ONE_SHIFT(1, 31, s32, 0, true); + TEST_ONE_SHIFT(1, 63, s64, 0, true); /* Overflow: high bit falls off unsigned types. */ /* 10010110 */ - err |= TEST_ONE_SHIFT(150, 1, u8, 0, true); + TEST_ONE_SHIFT(150, 1, u8, 0, true); /* 1000100010010110 */ - err |= TEST_ONE_SHIFT(34966, 1, u16, 0, true); + TEST_ONE_SHIFT(34966, 1, u16, 0, true); /* 10000100000010001000100010010110 */ - err |= TEST_ONE_SHIFT(2215151766U, 1, u32, 0, true); - err |= TEST_ONE_SHIFT(2215151766U, 1, unsigned int, 0, true); + TEST_ONE_SHIFT(2215151766U, 1, u32, 0, true); + TEST_ONE_SHIFT(2215151766U, 1, unsigned int, 0, true); /* 1000001000010000010000000100000010000100000010001000100010010110 */ - err |= TEST_ONE_SHIFT(9372061470395238550ULL, 1, u64, 0, true); + TEST_ONE_SHIFT(9372061470395238550ULL, 1, u64, 0, true); /* Overflow: bit shifted into signed bit on signed types. */ /* 01001011 */ - err |= TEST_ONE_SHIFT(75, 1, s8, 0, true); + TEST_ONE_SHIFT(75, 1, s8, 0, true); /* 0100010001001011 */ - err |= TEST_ONE_SHIFT(17483, 1, s16, 0, true); + TEST_ONE_SHIFT(17483, 1, s16, 0, true); /* 01000010000001000100010001001011 */ - err |= TEST_ONE_SHIFT(1107575883, 1, s32, 0, true); - err |= TEST_ONE_SHIFT(1107575883, 1, int, 0, true); + TEST_ONE_SHIFT(1107575883, 1, s32, 0, true); + TEST_ONE_SHIFT(1107575883, 1, int, 0, true); /* 0100000100001000001000000010000001000010000001000100010001001011 */ - err |= TEST_ONE_SHIFT(4686030735197619275LL, 1, s64, 0, true); + TEST_ONE_SHIFT(4686030735197619275LL, 1, s64, 0, true); /* Overflow: bit shifted past signed bit on signed types. */ /* 01001011 */ - err |= TEST_ONE_SHIFT(75, 2, s8, 0, true); + TEST_ONE_SHIFT(75, 2, s8, 0, true); /* 0100010001001011 */ - err |= TEST_ONE_SHIFT(17483, 2, s16, 0, true); + TEST_ONE_SHIFT(17483, 2, s16, 0, true); /* 01000010000001000100010001001011 */ - err |= TEST_ONE_SHIFT(1107575883, 2, s32, 0, true); - err |= TEST_ONE_SHIFT(1107575883, 2, int, 0, true); + TEST_ONE_SHIFT(1107575883, 2, s32, 0, true); + TEST_ONE_SHIFT(1107575883, 2, int, 0, true); /* 0100000100001000001000000010000001000010000001000100010001001011 */ - err |= TEST_ONE_SHIFT(4686030735197619275LL, 2, s64, 0, true); + TEST_ONE_SHIFT(4686030735197619275LL, 2, s64, 0, true); /* Overflow: values larger than destination type. */ - err |= TEST_ONE_SHIFT(0x100, 0, u8, 0, true); - err |= TEST_ONE_SHIFT(0xFF, 0, s8, 0, true); - err |= TEST_ONE_SHIFT(0x10000U, 0, u16, 0, true); - err |= TEST_ONE_SHIFT(0xFFFFU, 0, s16, 0, true); - err |= TEST_ONE_SHIFT(0x100000000ULL, 0, u32, 0, true); - err |= TEST_ONE_SHIFT(0x100000000ULL, 0, unsigned int, 0, true); - err |= TEST_ONE_SHIFT(0xFFFFFFFFUL, 0, s32, 0, true); - err |= TEST_ONE_SHIFT(0xFFFFFFFFUL, 0, int, 0, true); - err |= TEST_ONE_SHIFT(0xFFFFFFFFFFFFFFFFULL, 0, s64, 0, true); + TEST_ONE_SHIFT(0x100, 0, u8, 0, true); + TEST_ONE_SHIFT(0xFF, 0, s8, 0, true); + TEST_ONE_SHIFT(0x10000U, 0, u16, 0, true); + TEST_ONE_SHIFT(0xFFFFU, 0, s16, 0, true); + TEST_ONE_SHIFT(0x100000000ULL, 0, u32, 0, true); + TEST_ONE_SHIFT(0x100000000ULL, 0, unsigned int, 0, true); + TEST_ONE_SHIFT(0xFFFFFFFFUL, 0, s32, 0, true); + TEST_ONE_SHIFT(0xFFFFFFFFUL, 0, int, 0, true); + TEST_ONE_SHIFT(0xFFFFFFFFFFFFFFFFULL, 0, s64, 0, true); /* Nonsense: negative initial value. */ - err |= TEST_ONE_SHIFT(-1, 0, s8, 0, true); - err |= TEST_ONE_SHIFT(-1, 0, u8, 0, true); - err |= TEST_ONE_SHIFT(-5, 0, s16, 0, true); - err |= TEST_ONE_SHIFT(-5, 0, u16, 0, true); - err |= TEST_ONE_SHIFT(-10, 0, int, 0, true); - err |= TEST_ONE_SHIFT(-10, 0, unsigned int, 0, true); - err |= TEST_ONE_SHIFT(-100, 0, s32, 0, true); - err |= TEST_ONE_SHIFT(-100, 0, u32, 0, true); - err |= TEST_ONE_SHIFT(-10000, 0, s64, 0, true); - err |= TEST_ONE_SHIFT(-10000, 0, u64, 0, true); + TEST_ONE_SHIFT(-1, 0, s8, 0, true); + TEST_ONE_SHIFT(-1, 0, u8, 0, true); + TEST_ONE_SHIFT(-5, 0, s16, 0, true); + TEST_ONE_SHIFT(-5, 0, u16, 0, true); + TEST_ONE_SHIFT(-10, 0, int, 0, true); + TEST_ONE_SHIFT(-10, 0, unsigned int, 0, true); + TEST_ONE_SHIFT(-100, 0, s32, 0, true); + TEST_ONE_SHIFT(-100, 0, u32, 0, true); + TEST_ONE_SHIFT(-10000, 0, s64, 0, true); + TEST_ONE_SHIFT(-10000, 0, u64, 0, true); /* Nonsense: negative shift values. */ - err |= TEST_ONE_SHIFT(0, -5, s8, 0, true); - err |= TEST_ONE_SHIFT(0, -5, u8, 0, true); - err |= TEST_ONE_SHIFT(0, -10, s16, 0, true); - err |= TEST_ONE_SHIFT(0, -10, u16, 0, true); - err |= TEST_ONE_SHIFT(0, -15, int, 0, true); - err |= TEST_ONE_SHIFT(0, -15, unsigned int, 0, true); - err |= TEST_ONE_SHIFT(0, -20, s32, 0, true); - err |= TEST_ONE_SHIFT(0, -20, u32, 0, true); - err |= TEST_ONE_SHIFT(0, -30, s64, 0, true); - err |= TEST_ONE_SHIFT(0, -30, u64, 0, true); + TEST_ONE_SHIFT(0, -5, s8, 0, true); + TEST_ONE_SHIFT(0, -5, u8, 0, true); + TEST_ONE_SHIFT(0, -10, s16, 0, true); + TEST_ONE_SHIFT(0, -10, u16, 0, true); + TEST_ONE_SHIFT(0, -15, int, 0, true); + TEST_ONE_SHIFT(0, -15, unsigned int, 0, true); + TEST_ONE_SHIFT(0, -20, s32, 0, true); + TEST_ONE_SHIFT(0, -20, u32, 0, true); + TEST_ONE_SHIFT(0, -30, s64, 0, true); + TEST_ONE_SHIFT(0, -30, u64, 0, true); /* Overflow: shifted at or beyond entire type's bit width. */ - err |= TEST_ONE_SHIFT(0, 8, u8, 0, true); - err |= TEST_ONE_SHIFT(0, 9, u8, 0, true); - err |= TEST_ONE_SHIFT(0, 8, s8, 0, true); - err |= TEST_ONE_SHIFT(0, 9, s8, 0, true); - err |= TEST_ONE_SHIFT(0, 16, u16, 0, true); - err |= TEST_ONE_SHIFT(0, 17, u16, 0, true); - err |= TEST_ONE_SHIFT(0, 16, s16, 0, true); - err |= TEST_ONE_SHIFT(0, 17, s16, 0, true); - err |= TEST_ONE_SHIFT(0, 32, u32, 0, true); - err |= TEST_ONE_SHIFT(0, 33, u32, 0, true); - err |= TEST_ONE_SHIFT(0, 32, int, 0, true); - err |= TEST_ONE_SHIFT(0, 33, int, 0, true); - err |= TEST_ONE_SHIFT(0, 32, s32, 0, true); - err |= TEST_ONE_SHIFT(0, 33, s32, 0, true); - err |= TEST_ONE_SHIFT(0, 64, u64, 0, true); - err |= TEST_ONE_SHIFT(0, 65, u64, 0, true); - err |= TEST_ONE_SHIFT(0, 64, s64, 0, true); - err |= TEST_ONE_SHIFT(0, 65, s64, 0, true); + TEST_ONE_SHIFT(0, 8, u8, 0, true); + TEST_ONE_SHIFT(0, 9, u8, 0, true); + TEST_ONE_SHIFT(0, 8, s8, 0, true); + TEST_ONE_SHIFT(0, 9, s8, 0, true); + TEST_ONE_SHIFT(0, 16, u16, 0, true); + TEST_ONE_SHIFT(0, 17, u16, 0, true); + TEST_ONE_SHIFT(0, 16, s16, 0, true); + TEST_ONE_SHIFT(0, 17, s16, 0, true); + TEST_ONE_SHIFT(0, 32, u32, 0, true); + TEST_ONE_SHIFT(0, 33, u32, 0, true); + TEST_ONE_SHIFT(0, 32, int, 0, true); + TEST_ONE_SHIFT(0, 33, int, 0, true); + TEST_ONE_SHIFT(0, 32, s32, 0, true); + TEST_ONE_SHIFT(0, 33, s32, 0, true); + TEST_ONE_SHIFT(0, 64, u64, 0, true); + TEST_ONE_SHIFT(0, 65, u64, 0, true); + TEST_ONE_SHIFT(0, 64, s64, 0, true); + TEST_ONE_SHIFT(0, 65, s64, 0, true); /* * Corner case: for unsigned types, we fail when we've shifted @@ -473,13 +442,11 @@ static int __init test_overflow_shift(void) * signed bit). So, for now, we will test this condition but * mark it as not expected to overflow. */ - err |= TEST_ONE_SHIFT(0, 7, s8, 0, false); - err |= TEST_ONE_SHIFT(0, 15, s16, 0, false); - err |= TEST_ONE_SHIFT(0, 31, int, 0, false); - err |= TEST_ONE_SHIFT(0, 31, s32, 0, false); - err |= TEST_ONE_SHIFT(0, 63, s64, 0, false); - - return err; + TEST_ONE_SHIFT(0, 7, s8, 0, false); + TEST_ONE_SHIFT(0, 15, s16, 0, false); + TEST_ONE_SHIFT(0, 31, int, 0, false); + TEST_ONE_SHIFT(0, 31, s32, 0, false); + TEST_ONE_SHIFT(0, 63, s64, 0, false); } /* @@ -499,7 +466,7 @@ static int __init test_overflow_shift(void) #define TEST_SIZE (5 * 4096) #define DEFINE_TEST_ALLOC(func, free_func, want_arg, want_gfp, want_node)\ -static int __init test_ ## func (void *arg) \ +static void test_ ## func (struct kunit *test, void *arg) \ { \ volatile size_t a = TEST_SIZE; \ volatile size_t b = (SIZE_MAX / TEST_SIZE) + 1; \ @@ -507,31 +474,26 @@ static int __init test_ ## func (void *arg) \ \ /* Tiny allocation test. */ \ ptr = alloc ## want_arg ## want_gfp ## want_node (func, arg, 1);\ - if (!ptr) { \ - pr_warn(#func " failed regular allocation?!\n"); \ - return 1; \ - } \ + KUNIT_ASSERT_NOT_ERR_OR_NULL_MSG(test, ptr, \ + #func " failed regular allocation?!"); \ free ## want_arg (free_func, arg, ptr); \ \ /* Wrapped allocation test. */ \ ptr = alloc ## want_arg ## want_gfp ## want_node (func, arg, \ a * b); \ - if (!ptr) { \ - pr_warn(#func " unexpectedly failed bad wrapping?!\n"); \ - return 1; \ - } \ + KUNIT_ASSERT_NOT_ERR_OR_NULL_MSG(test, ptr, \ + #func " unexpectedly failed bad wrapping?!"); \ free ## want_arg (free_func, arg, ptr); \ \ /* Saturated allocation test. */ \ ptr = alloc ## want_arg ## want_gfp ## want_node (func, arg, \ array_size(a, b)); \ if (ptr) { \ - pr_warn(#func " missed saturation!\n"); \ + KUNIT_FAIL(test, #func " missed saturation!\n"); \ free ## want_arg (free_func, arg, ptr); \ - return 1; \ + return; \ } \ pr_info(#func " detected saturation\n"); \ - return 0; \ } /* @@ -555,60 +517,56 @@ DEFINE_TEST_ALLOC(kvzalloc_node, kvfree, 0, 1, 1); DEFINE_TEST_ALLOC(devm_kmalloc, devm_kfree, 1, 1, 0); DEFINE_TEST_ALLOC(devm_kzalloc, devm_kfree, 1, 1, 0); -static int __init test_overflow_allocation(void) +static void test_overflow_allocation(struct kunit *test) { const char device_name[] = "overflow-test"; struct device *dev; - int err = 0; /* Create dummy device for devm_kmalloc()-family tests. */ dev = root_device_register(device_name); - if (IS_ERR(dev)) { - pr_warn("Cannot register test device\n"); - return 1; - } - - err |= test_kmalloc(NULL); - err |= test_kmalloc_node(NULL); - err |= test_kzalloc(NULL); - err |= test_kzalloc_node(NULL); - err |= test_kvmalloc(NULL); - err |= test_kvmalloc_node(NULL); - err |= test_kvzalloc(NULL); - err |= test_kvzalloc_node(NULL); - err |= test_vmalloc(NULL); - err |= test_vmalloc_node(NULL); - err |= test_vzalloc(NULL); - err |= test_vzalloc_node(NULL); - err |= test_devm_kmalloc(dev); - err |= test_devm_kzalloc(dev); + KUNIT_ASSERT_NOT_ERR_OR_NULL_MSG(test, dev, + "Cannot register test device"); + + test_kmalloc(test, NULL); + test_kmalloc_node(test, NULL); + test_kzalloc(test, NULL); + test_kzalloc_node(test, NULL); + test_kvmalloc(test, NULL); + test_kvmalloc_node(test, NULL); + test_kvzalloc(test, NULL); + test_kvzalloc_node(test, NULL); + test_vmalloc(test, NULL); + test_vmalloc_node(test, NULL); + test_vzalloc(test, NULL); + test_vzalloc_node(test, NULL); + test_devm_kmalloc(test, dev); + test_devm_kzalloc(test, dev); device_unregister(dev); - - return err; } -static int __init test_module_init(void) -{ - int err = 0; - - err |= test_overflow_calculation(); - err |= test_overflow_shift(); - err |= test_overflow_allocation(); +static struct kunit_case overflow_test_cases[] = { + KUNIT_CASE(test_u8_overflow), + KUNIT_CASE(test_s8_overflow), + KUNIT_CASE(test_u16_overflow), + KUNIT_CASE(test_s16_overflow), + KUNIT_CASE(test_u32_overflow), + KUNIT_CASE(test_s32_overflow), +#if BITS_PER_LONG == 64 + KUNIT_CASE(test_u64_overflow), + KUNIT_CASE(test_s64_overflow), +#endif - if (err) { - pr_warn("FAIL!\n"); - err = -EINVAL; - } else { - pr_info("all tests passed\n"); - } + KUNIT_CASE(test_overflow_shift), + KUNIT_CASE(test_overflow_allocation), + {} +}; - return err; -} +static struct kunit_suite overflow_test_suite = { + .name = "lib_overflow", + .test_cases = overflow_test_cases, +}; -static void __exit test_module_exit(void) -{ } +kunit_test_suites(&overflow_test_suite); -module_init(test_module_init); -module_exit(test_module_exit); MODULE_LICENSE("Dual MIT/GPL"); base-commit: cda689f8708b6bef0b921c3a17fcdecbe959a079 -- 2.31.1.527.g47e6f16901-goog

4 years, 2 months

2
2
0 0

[PATCH v4] KVM: x86: Fix KVM_GET_CPUID2 ioctl to return cpuid entries count

by Valeriy Vdovin

KVM_GET_CPUID2 kvm ioctl is not very well documented, but the way it is implemented in function kvm_vcpu_ioctl_get_cpuid2 suggests that even at error path it will try to return number of entries to the caller. But The dispatcher kvm vcpu ioctl dispatcher code in kvm_arch_vcpu_ioctl ignores any output from this function if it sees the error return code. It's very explicit by the code that it was designed to receive some small number of entries to return E2BIG along with the corrected number. This lost logic in the dispatcher code has been restored by removing the lines that check for function return code and skip if error is found. Without it, the ioctl caller will see both the number of entries and the correct error. In selftests relevant function vcpu_get_cpuid has also been modified to utilize the number of cpuid entries returned along with errno E2BIG. Signed-off-by: Valeriy Vdovin <valeriy.vdovin(a)virtuozzo.com> --- v4: - Added description to documentation of KVM_GET_CPUID2. - Copy back nent only if E2BIG is returned. - Fixed error code sign. - Corrected version message Documentation/virt/kvm/api.rst | 81 ++++++++++++------- arch/x86/kvm/x86.c | 11 ++- .../selftests/kvm/lib/x86_64/processor.c | 20 +++-- 3 files changed, 73 insertions(+), 39 deletions(-) diff --git a/Documentation/virt/kvm/api.rst b/Documentation/virt/kvm/api.rst index 245d80581f15..c7cfe4b9614e 100644 --- a/Documentation/virt/kvm/api.rst +++ b/Documentation/virt/kvm/api.rst @@ -711,7 +711,34 @@ resulting CPUID configuration through KVM_GET_CPUID2 in case. }; -4.21 KVM_SET_SIGNAL_MASK +4.21 KVM_GET_CPUID2 +------------------ + +:Capability: basic +:Architectures: x86 +:Type: vcpu ioctl +:Parameters: struct kvm_cpuid (in/out) +:Returns: 0 on success, -1 on error + +Returns a full list of cpuid entries that are supported by this vcpu and were +previously set by KVM_SET_CPUID/KVM_SET_CPUID2. + +The userspace must specify the number of cpuid entries it is ready to accept +from the kernel in the 'nent' field of 'struct kmv_cpuid'. + +The kernel will try to return all the cpuid entries it has in the response. +If the userspace nent value is too small for the full response, the kernel will +set the error code to -E2BIG, set the same 'nent' field to the actual number of +cpuid_entries and return without writing back any entries to the userspace. +The userspace can thus implement a two-call sequence, where the first call is +made with nent set to 0 to read the number of entries from the kernel and +use this response to allocate enough memory for a full response for the second +call. + +The call cal also return with error code -EFAULT in case of other errors. + + +4.22 KVM_SET_SIGNAL_MASK ------------------------ :Capability: basic @@ -737,7 +764,7 @@ signal mask. }; -4.22 KVM_GET_FPU +4.23 KVM_GET_FPU ---------------- :Capability: basic @@ -766,7 +793,7 @@ Reads the floating point state from the vcpu. }; -4.23 KVM_SET_FPU +4.24 KVM_SET_FPU ---------------- :Capability: basic @@ -795,7 +822,7 @@ Writes the floating point state to the vcpu. }; -4.24 KVM_CREATE_IRQCHIP +4.25 KVM_CREATE_IRQCHIP ----------------------- :Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390) @@ -817,7 +844,7 @@ Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled before KVM_CREATE_IRQCHIP can be used. -4.25 KVM_IRQ_LINE +4.26 KVM_IRQ_LINE ----------------- :Capability: KVM_CAP_IRQCHIP @@ -886,7 +913,7 @@ be used for a userspace interrupt controller. }; -4.26 KVM_GET_IRQCHIP +4.27 KVM_GET_IRQCHIP -------------------- :Capability: KVM_CAP_IRQCHIP @@ -911,7 +938,7 @@ KVM_CREATE_IRQCHIP into a buffer provided by the caller. }; -4.27 KVM_SET_IRQCHIP +4.28 KVM_SET_IRQCHIP -------------------- :Capability: KVM_CAP_IRQCHIP @@ -936,7 +963,7 @@ KVM_CREATE_IRQCHIP from a buffer provided by the caller. }; -4.28 KVM_XEN_HVM_CONFIG +4.29 KVM_XEN_HVM_CONFIG ----------------------- :Capability: KVM_CAP_XEN_HVM @@ -972,7 +999,7 @@ fields must be zero. No other flags are currently valid in the struct kvm_xen_hvm_config. -4.29 KVM_GET_CLOCK +4.30 KVM_GET_CLOCK ------------------ :Capability: KVM_CAP_ADJUST_CLOCK @@ -1005,7 +1032,7 @@ TSC is not stable. }; -4.30 KVM_SET_CLOCK +4.31 KVM_SET_CLOCK ------------------ :Capability: KVM_CAP_ADJUST_CLOCK @@ -1027,7 +1054,7 @@ such as migration. }; -4.31 KVM_GET_VCPU_EVENTS +4.32 KVM_GET_VCPU_EVENTS ------------------------ :Capability: KVM_CAP_VCPU_EVENTS @@ -1146,7 +1173,7 @@ directly to the virtual CPU). __u32 reserved[12]; }; -4.32 KVM_SET_VCPU_EVENTS +4.33 KVM_SET_VCPU_EVENTS ------------------------ :Capability: KVM_CAP_VCPU_EVENTS @@ -1209,7 +1236,7 @@ exceptions by manipulating individual registers using the KVM_SET_ONE_REG API. See KVM_GET_VCPU_EVENTS for the data structure. -4.33 KVM_GET_DEBUGREGS +4.34 KVM_GET_DEBUGREGS ---------------------- :Capability: KVM_CAP_DEBUGREGS @@ -1231,7 +1258,7 @@ Reads debug registers from the vcpu. }; -4.34 KVM_SET_DEBUGREGS +4.35 KVM_SET_DEBUGREGS ---------------------- :Capability: KVM_CAP_DEBUGREGS @@ -1246,7 +1273,7 @@ See KVM_GET_DEBUGREGS for the data structure. The flags field is unused yet and must be cleared on entry. -4.35 KVM_SET_USER_MEMORY_REGION +4.36 KVM_SET_USER_MEMORY_REGION ------------------------------- :Capability: KVM_CAP_USER_MEMORY @@ -1315,7 +1342,7 @@ The KVM_SET_MEMORY_REGION does not allow fine grained control over memory allocation and is deprecated. -4.36 KVM_SET_TSS_ADDR +4.37 KVM_SET_TSS_ADDR --------------------- :Capability: KVM_CAP_SET_TSS_ADDR @@ -1335,7 +1362,7 @@ because of a quirk in the virtualization implementation (see the internals documentation when it pops into existence). -4.37 KVM_ENABLE_CAP +4.38 KVM_ENABLE_CAP ------------------- :Capability: KVM_CAP_ENABLE_CAP @@ -1390,7 +1417,7 @@ function properly, this is the place to put them. The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl for vm-wide capabilities. -4.38 KVM_GET_MP_STATE +4.39 KVM_GET_MP_STATE --------------------- :Capability: KVM_CAP_MP_STATE @@ -1438,7 +1465,7 @@ For arm/arm64: The only states that are valid are KVM_MP_STATE_STOPPED and KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not. -4.39 KVM_SET_MP_STATE +4.40 KVM_SET_MP_STATE --------------------- :Capability: KVM_CAP_MP_STATE @@ -1460,7 +1487,7 @@ For arm/arm64: The only states that are valid are KVM_MP_STATE_STOPPED and KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not. -4.40 KVM_SET_IDENTITY_MAP_ADDR +4.41 KVM_SET_IDENTITY_MAP_ADDR ------------------------------ :Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR @@ -1484,7 +1511,7 @@ documentation when it pops into existence). Fails if any VCPU has already been created. -4.41 KVM_SET_BOOT_CPU_ID +4.42 KVM_SET_BOOT_CPU_ID ------------------------ :Capability: KVM_CAP_SET_BOOT_CPU_ID @@ -1499,7 +1526,7 @@ is vcpu 0. This ioctl has to be called before vcpu creation, otherwise it will return EBUSY error. -4.42 KVM_GET_XSAVE +4.43 KVM_GET_XSAVE ------------------ :Capability: KVM_CAP_XSAVE @@ -1518,7 +1545,7 @@ otherwise it will return EBUSY error. This ioctl would copy current vcpu's xsave struct to the userspace. -4.43 KVM_SET_XSAVE +4.44 KVM_SET_XSAVE ------------------ :Capability: KVM_CAP_XSAVE @@ -1537,7 +1564,7 @@ This ioctl would copy current vcpu's xsave struct to the userspace. This ioctl would copy userspace's xsave struct to the kernel. -4.44 KVM_GET_XCRS +4.45 KVM_GET_XCRS ----------------- :Capability: KVM_CAP_XCRS @@ -1564,7 +1591,7 @@ This ioctl would copy userspace's xsave struct to the kernel. This ioctl would copy current vcpu's xcrs to the userspace. -4.45 KVM_SET_XCRS +4.46 KVM_SET_XCRS ----------------- :Capability: KVM_CAP_XCRS @@ -1591,7 +1618,7 @@ This ioctl would copy current vcpu's xcrs to the userspace. This ioctl would set vcpu's xcr to the value userspace specified. -4.46 KVM_GET_SUPPORTED_CPUID +4.47 KVM_GET_SUPPORTED_CPUID ---------------------------- :Capability: KVM_CAP_EXT_CPUID @@ -1676,7 +1703,7 @@ if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the feature in userspace, then you can enable the feature for KVM_SET_CPUID2. -4.47 KVM_PPC_GET_PVINFO +4.48 KVM_PPC_GET_PVINFO ----------------------- :Capability: KVM_CAP_PPC_GET_PVINFO diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index efc7a82ab140..3f941b1f4e78 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -4773,14 +4773,17 @@ long kvm_arch_vcpu_ioctl(struct file *filp, r = -EFAULT; if (copy_from_user(&cpuid, cpuid_arg, sizeof(cpuid))) goto out; + r = kvm_vcpu_ioctl_get_cpuid2(vcpu, &cpuid, cpuid_arg->entries); - if (r) + + if (r && r != -E2BIG) goto out; - r = -EFAULT; - if (copy_to_user(cpuid_arg, &cpuid, sizeof(cpuid))) + + if (copy_to_user(cpuid_arg, &cpuid, sizeof(cpuid))) { + r = -EFAULT; goto out; - r = 0; + } break; } case KVM_GET_MSRS: { diff --git a/tools/testing/selftests/kvm/lib/x86_64/processor.c b/tools/testing/selftests/kvm/lib/x86_64/processor.c index a8906e60a108..a412b39ad791 100644 --- a/tools/testing/selftests/kvm/lib/x86_64/processor.c +++ b/tools/testing/selftests/kvm/lib/x86_64/processor.c @@ -727,17 +727,21 @@ struct kvm_cpuid2 *vcpu_get_cpuid(struct kvm_vm *vm, uint32_t vcpuid) cpuid = allocate_kvm_cpuid2(); max_ent = cpuid->nent; + cpuid->nent = 0; - for (cpuid->nent = 1; cpuid->nent <= max_ent; cpuid->nent++) { - rc = ioctl(vcpu->fd, KVM_GET_CPUID2, cpuid); - if (!rc) - break; + rc = ioctl(vcpu->fd, KVM_GET_CPUID2, cpuid); + TEST_ASSERT(rc == -1 && errno == E2BIG, + "KVM_GET_CPUID2 should return E2BIG: %d %d", + rc, errno); - TEST_ASSERT(rc == -1 && errno == E2BIG, - "KVM_GET_CPUID2 should either succeed or give E2BIG: %d %d", - rc, errno); - } + TEST_ASSERT(cpuid->nent, + "KVM_GET_CPUID2 failed to set cpuid->nent with E2BIG"); + + TEST_ASSERT(cpuid->nent < max_ent, + "KVM_GET_CPUID2 has %d entries, expected maximum: %d", + cpuid->nent, max_ent); + rc = ioctl(vcpu->fd, KVM_GET_CPUID2, cpuid); TEST_ASSERT(rc == 0, "KVM_GET_CPUID2 failed, rc: %i errno: %i", rc, errno); -- 2.17.1

4 years, 2 months

6
9
0 0

[PATCH] selftests: kvm: remove reassignment of non-absolute variables

by Bill Wendling

Clang's integrated assembler does not allow symbols with non-absolute values to be reassigned. Modify the interrupt entry loop macro to be compatible with IAS by using a label and an offset. Cc: Jian Cai <caij2003(a)gmail.com> Signed-off-by: Bill Wendling <morbo(a)google.com> References: https://lore.kernel.org/lkml/20200714233024.1789985-1-caij2003@gmail.com/ --- tools/testing/selftests/kvm/lib/x86_64/handlers.S | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/kvm/lib/x86_64/handlers.S b/tools/testing/selftests/kvm/lib/x86_64/handlers.S index aaf7bc7d2ce1..3f9181e9a0a7 100644 --- a/tools/testing/selftests/kvm/lib/x86_64/handlers.S +++ b/tools/testing/selftests/kvm/lib/x86_64/handlers.S @@ -54,9 +54,9 @@ idt_handlers: .align 8 /* Fetch current address and append it to idt_handlers. */ - current_handler = . +0 : .pushsection .rodata -.quad current_handler + .quad 0b .popsection .if ! \has_error -- 2.29.2.576.ga3fc446d84-goog

4 years, 2 months

4
9
0 0

Subject: [PATCH 0/1] SGX self test fails

by Tim Gardner

I'm just starting my learning curve on SGX, so I don't know if I've missed some setup for the SGX device entries. After looking at arch/x86/kernel/cpu/sgx/driver.c I see that there is no mode value for either sgx_dev_enclave or sgx_dev_provision. With this patch I can get the SGX self test to complete: sudo ./test_sgx Warning: no execute permissions on device file /dev/sgx_enclave 0x0000000000000000 0x0000000000002000 0x03 0x0000000000002000 0x0000000000001000 0x05 0x0000000000003000 0x0000000000003000 0x03 SUCCESS Is the warning even necessary ? Tim

4 years, 2 months

4
7
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-kselftest-mirror