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

[PATCH v4 0/9] riscv: vector: misc ptrace fixes for debug use-cases

by Sergey Matyukevich

This patch series suggests fixes for several corner cases in the RISC-V vector ptrace implementation: - init vector context with proper vlenb, to avoid reading zero vlenb by an early attached debugger - follow gdbserver expectations and return ENODATA instead of EINVAL if vector extension is supported but not yet activated for the traced process - validate input vector csr registers in ptrace, to maintain an accurate view of the tracee's vector context across multiple halt/resume debug cycles For detailed description see the appropriate commit messages. A new test suite v_ptrace is added into the tools/testing/selftests/riscv/vector to verify some of the vector ptrace functionality and corner cases. Previous versions: - v3: https://lore.kernel.org/linux-riscv/20251025210655.43099-1-geomatsi@gmail.c… - v2: https://lore.kernel.org/linux-riscv/20250821173957.563472-1-geomatsi@gmail.… - v1: https://lore.kernel.org/linux-riscv/20251007115840.2320557-1-geomatsi@gmail… Changes in v4: The form 'vsetvli x0, x0, ...' can only be used if VLMAX remains unchanged, see spec 6.2. This condition was not met by the initial values in the selftests w.r.t. the initial zeroed context. QEMU accepted such values, but actual hardware (c908, BananaPi CanMV Zero board) did not, setting vill. So fix the selftests after testing on hardware: - replace 'vsetvli x0, x0, ...' by 'vsetvli rd, x0, ...' - fixed instruction returns VLMAX, so use it in checks as well - replace fixed vlenb == 16 in the syscall test Changes in v3: Address the review comments by Andy Chiu and rework the approach: - drop forced vector context save entirely - perform strict validation of vector csr regs in ptrace Changes in v2: - add thread_info flag to allow to force vector context save - force vector context save after vector ptrace to ensure valid vector context in the next ptrace operations - force vector context save on the first context switch after vector context init to get proper vlenb --- Ilya Mamay (1): riscv: ptrace: return ENODATA for inactive vector extension Sergey Matyukevich (8): selftests: riscv: test ptrace vector interface selftests: riscv: verify initial vector state with ptrace riscv: vector: init vector context with proper vlenb riscv: csr: define vtype registers elements riscv: ptrace: validate input vector csr registers selftests: riscv: verify ptrace rejects invalid vector csr inputs selftests: riscv: verify ptrace accepts valid vector csr values selftests: riscv: verify syscalls discard vector context arch/riscv/include/asm/csr.h | 11 + arch/riscv/kernel/ptrace.c | 72 +- arch/riscv/kernel/vector.c | 12 +- .../testing/selftests/riscv/vector/.gitignore | 1 + tools/testing/selftests/riscv/vector/Makefile | 5 +- .../testing/selftests/riscv/vector/v_ptrace.c | 754 ++++++++++++++++++ 6 files changed, 847 insertions(+), 8 deletions(-) create mode 100644 tools/testing/selftests/riscv/vector/v_ptrace.c base-commit: e811c33b1f137be26a20444b79db8cbc1fca1c89 -- 2.51.0

4 days, 3 hours

2
11
0 0

[PATCH net-next 12/12] selftests: net: selftest for ipvlan-macnat mode

by Dmitry Skorodumov

Implemented a self-test for ipvlan in l2macnat mode. The test verifies: 1) It's not possible to configure an ip in l2macnat mode on ipvtap 2) It creates several net namespaces - Default namespace emulates host, - ipvlan-tst-phy emulates some host in remote network - ipvlan-tst-0/1 emulate VMs on host. Test verifies, that MAC addresses are as expected in ARP/NEIGH tables: all MACs in 'tst-phy' points to "host" mac-address all MACs in Default and tst are real ones 3) The l2macnat mode has limited number of addresses remembered on port. Test verifies, that this limit really works. Signed-off-by: Dmitry Skorodumov <skorodumov.dmitry(a)huawei.com> --- tools/testing/selftests/net/Makefile | 2 + .../selftests/net/ipvtap_macnat_bridge.py | 168 +++++++++ .../selftests/net/ipvtap_macnat_test.sh | 333 ++++++++++++++++++ 3 files changed, 503 insertions(+) create mode 100755 tools/testing/selftests/net/ipvtap_macnat_bridge.py create mode 100755 tools/testing/selftests/net/ipvtap_macnat_test.sh diff --git a/tools/testing/selftests/net/Makefile b/tools/testing/selftests/net/Makefile index b5127e968108..050d864f0bd9 100644 --- a/tools/testing/selftests/net/Makefile +++ b/tools/testing/selftests/net/Makefile @@ -49,6 +49,7 @@ TEST_PROGS := \ ipv6_flowlabel.sh \ ipv6_force_forwarding.sh \ ipv6_route_update_soft_lockup.sh \ + ipvtap_macnat_test.sh \ l2_tos_ttl_inherit.sh \ l2tp.sh \ link_netns.py \ @@ -191,6 +192,7 @@ TEST_GEN_PROGS := \ TEST_FILES := \ fcnal-test.sh \ in_netns.sh \ + ipvtap_macnat_bridge.py \ lib.sh \ settings \ setup_loopback.sh \ diff --git a/tools/testing/selftests/net/ipvtap_macnat_bridge.py b/tools/testing/selftests/net/ipvtap_macnat_bridge.py new file mode 100755 index 000000000000..7dc4a626e5bb --- /dev/null +++ b/tools/testing/selftests/net/ipvtap_macnat_bridge.py @@ -0,0 +1,168 @@ +#!/usr/bin/env python3 +# SPDX-License-Identifier: GPL-2.0 + +""" +Script to bridge ipvtap and tap, +needed to simulate behaviour of virtual machine using ipvtap. + +ipvtap in macnat mode cannot have IP address. +Due to limitations of ipvtap, it also cannot be plugged +into bridge. +Use this script to connect ipvtap and tap and assing IP to tap. +""" + +import socket +import os +import select +import sys +import signal +import fcntl +import struct +import subprocess + +# Linux TUN/TAP constants +TUNSETIFF = 0x400454ca +IFF_TUN = 0x0001 +IFF_TAP = 0x0002 +IFF_NO_PI = 0x1000 + +ns_name = "non-initialized" + +class TapBridge: + """Simple class to bridge ipvtap and tap interfaces""" + def __init__(self, tap, ipvtap, buffer_size=65536): + self.tap_name = tap + self.ipvtap_name = ipvtap + self.buffer_size = buffer_size + self.running = False + + def open_ipvtap_sock(self, tap_name): + """Open a IPVTAP interface using raw socket""" + try: + sock = socket.socket(socket.AF_PACKET, + socket.SOCK_RAW, + socket.ntohs(0x0003)) + sock.bind((tap_name, 0)) + sock.setblocking(False) + print(f"Connected to IPVTAP interface: {tap_name}") + return sock + + except (OSError, IOError) as e: + print(f"Error opening IPVTAP interface {tap_name}: {e}") + return None + + def create_tap_interface(self, tap_name): + """Create and configure a TAP interface using /dev/net/tun""" + try: + # Open the tun device + tun_fd = os.open('/dev/net/tun', os.O_RDWR) + if tun_fd < 0: + raise OSError("Failed to open /dev/net/tun (err: {os.errno})") + + # Prepare the ifr structure + tap_name_bytes = tap_name.encode('utf-8') + ifr = struct.pack('16sH', tap_name_bytes, IFF_TAP | IFF_NO_PI) + + # Set the interface name and flags + result = fcntl.ioctl(tun_fd, TUNSETIFF, ifr) + + # Get the actual interface name that was set + unpacked = struct.unpack('16sH', result) + actual_name = unpacked[0].split(b'\x00')[0].decode() + print(f"Created TAP interface: {actual_name}") + + return tun_fd + + except (OSError, IOError) as e: + print(f"Error creating TAP interface {tap_name}: {e}") + return None + + def forward_data(self, from_fd, to_fd, description): + """Forward data from one file descriptor to another""" + try: + data = os.read(from_fd, self.buffer_size) + if data: + os.write(to_fd, data) + return True + return False + + except BlockingIOError: + return True + except (OSError, IOError) as e: + print(f"Error forwarding data {description}: {e}") + return False + + def run(self): + """Main bridge loop""" + # Create TAP interfaces + tap1_fd = self.create_tap_interface(self.tap_name) + + sock = self.open_ipvtap_sock(self.ipvtap_name) + tap2_fd = sock.fileno() + + if tap1_fd is None or tap2_fd is None: + print("Failed to create TAP interfaces") + return + + print("Press Ctrl+C to stop\n") + + self.running = True + stats = {'tap1_to_tap2': 0, 'tap2_to_tap1': 0} + while self.running: + try: + # Use select to monitor both file descriptors + readable, _, _ = select.select([tap1_fd, tap2_fd], [], [], 1.0) + + for fd in readable: + if fd == tap1_fd: + descr = f"from {self.tap_name} to {self.ipvtap_name}" + if self.forward_data(tap1_fd, tap2_fd, descr): + stats['tap1_to_tap2'] += 1 + else: + self.running = False + elif fd == tap2_fd: + descr = f"from {self.ipvtap_name} to {self.tap_name}" + if self.forward_data(tap2_fd, tap1_fd, descr): + stats['tap2_to_tap1'] += 1 + else: + self.running = False + + except KeyboardInterrupt: + print("\nShutting down...") + self.running = False + except (OSError, IOError) as e: + print(f"Error in main loop: {e}") + self.running = False + + # Cleanup + os.close(tap1_fd) + os.close(tap2_fd) + print(f"Bridge stopped in {ns_name}. Stats: {stats}") + + +def signal_handler(_sig, _frame): + """SIGINT handler for macnat bridge""" + print(f'\nReceived interrupt signal, shutting down bridge in {ns_name}') + sys.exit(0) + + +if __name__ == "__main__": + ns_name = subprocess.getoutput("ip netns identify") or "default" + + signal.signal(signal.SIGINT, signal_handler) + + # Check if running as root + if os.geteuid() != 0: + print("ERROR: This script must be run as root!") + sys.exit(1) + + if len(sys.argv) != 3: + print("Usage: tap_bridge.py tap_name ipvtap_name") + sys.exit(1) + + TAP = sys.argv[1] + IPVTAP = sys.argv[2] + + print(f"Starting TAP bridge between {TAP} and {IPVTAP} in {ns_name}") + bridge = TapBridge(TAP, IPVTAP) + bridge.run() diff --git a/tools/testing/selftests/net/ipvtap_macnat_test.sh b/tools/testing/selftests/net/ipvtap_macnat_test.sh new file mode 100755 index 000000000000..927d75af776b --- /dev/null +++ b/tools/testing/selftests/net/ipvtap_macnat_test.sh @@ -0,0 +1,333 @@ +#!/bin/bash +# SPDX-License-Identifier: GPL-2.0 +# +# Tests for ipvtap in macnat mode + +NS_TST0=ipvlan-tst-0 +NS_TST1=ipvlan-tst-1 +NS_PHY=ipvlan-tst-phy + +IP_HOST=172.25.0.1 +IP_PHY=172.25.0.2 +IP_TST0=172.25.0.10 +IP_TST1=172.25.0.30 + +IP_OK0=("172.25.0.10" "172.25.0.11" "172.25.0.12" "172.25.0.13") +IP6_OK0=("fc00::10" "fc00::11" "fc00::12" "fc00::13" ) + +IP_OVFL0="172.25.0.14" +IP6_OVFL0="fc00::14" + +IP6_HOST=fc00::1 +IP6_PHY=fc00::2 +IP6_TST0=fc00::10 +IP6_TST1=fc00::30 + +MAC_HOST="92:3a:00:00:00:01" +MAC_PHY="92:3a:00:00:00:02" +MAC_TST0="92:3a:00:00:00:10" +MAC_TST1="92:3a:00:00:00:30" + +VETH_HOST=vethtst +VETH_PHY=vethtst.p + +# +# The testing environment looks this way: +# +# |------HOST------| |------PHY-------| +# | veth<----------------->veth | +# |------|--|------| |----------------| +# | | +# | | |-----TST0-------| +# | |------------|----ipvtap | +# | |----------------| +# | +# | |-----TST1-------| +# |---------------|----ivtap | +# |----------------| +# +# The macnat mode is for virtual machines, so ipvtap-interface is supposed +# to be used only for traffic monitoring and doesn't have ip-address. +# +# To simulate a virtual machine on ipvtap, we create TAP-interfaces +# in TST environments and assing IP-addresses to them. +# TAP and IPVTAP are connected with simple python script. +# + +ns_run() { + ns=$1 + shift + if [[ "$ns" == "default" ]]; then + "$@" >/dev/null + else + ip netns exec "$ns" "$@" >/dev/null + fi +} + +configure_ns() { + local ns=$1 + local n=$2 + local ip=$3 + local ip6=$4 + local mac=$5 + + ns_run "$ns" ip link set lo up + + if ! ip link add netns "$ns" name "ipvtap0.$n" link $VETH_HOST \ + type ipvtap mode l2macnat bridge; then + exit_error "FAIL: Failed to configure ipvtap link." + fi + ns_run "$ns" ip link set "ipvtap0.$n" up + + ns_run "$ns" ip tuntap add mode tap "tap0.$n" + ns_run "$ns" ip link set dev "tap0.$n" address "$mac" + # disable dad + ns_run "$ns" sysctl -w "net/ipv6/conf/tap0.$n/accept_dad"=0 + ns_run "$ns" ip link set "tap0.$n" up + ns_run "$ns" ip a a "$ip/24" dev "tap0.$n" + ns_run "$ns" ip a a "$ip6/64" dev "tap0.$n" +} + +start_macnat_bridge() { + local ns=$1 + local n=$2 + ip netns exec "$ns" python3 ipvtap_macnat_bridge.py \ + "tap0.$n" "ipvtap0.$n" & +} + +configure_veth() { + local ns=$1 + local veth=$2 + local ip=$3 + local ip6=$4 + local mac=$5 + + ns_run "$ns" ip link set lo up + ns_run "$ns" ethtool -K "$veth" tx off rx off + ns_run "$ns" ip link set dev "$veth" address "$mac" + ns_run "$ns" ip link set "$veth" up + ns_run "$ns" ip a a "$ip/24" dev "$veth" + ns_run "$ns" ip a a "$ip6/64" dev "$veth" +} + +setup_env() { + ip netns add $NS_TST0 + ip netns add $NS_TST1 + ip netns add $NS_PHY + + # setup simulated other-host (phy) and host itself + ip link add $VETH_HOST type veth peer name $VETH_PHY \ + netns $NS_PHY >/dev/null + + # host config + configure_veth default $VETH_HOST $IP_HOST $IP6_HOST $MAC_HOST + configure_veth $NS_PHY $VETH_PHY $IP_PHY $IP6_PHY $MAC_PHY + + # TST namespaces config + configure_ns $NS_TST0 0 $IP_TST0 $IP6_TST0 $MAC_TST0 + configure_ns $NS_TST1 1 $IP_TST1 $IP6_TST1 $MAC_TST1 +} + +ping_all() { + # This will learn MAC/IP addresses on ipvtap + local ns=$1 + + ns_run "$ns" ping -c 1 $IP_TST0 + ns_run "$ns" ping -c 1 $IP6_TST0 + + ns_run "$ns" ping -c 1 $IP_TST1 + ns_run "$ns" ping -c 1 $IP6_TST1 + + ns_run "$ns" ping -c 1 $IP_HOST + ns_run "$ns" ping -c 1 $IP6_HOST + + ns_run "$ns" ping -c 1 $IP_PHY + ns_run "$ns" ping -c 1 $IP6_PHY +} + +check_mac_eq() { + # Ensure IP corresponds to MAC. + local ns=$1 + local ip=$2 + local mac=$3 + local dev=$4 + + if [[ "$ns" == "default" ]]; then + out=$( + ip neigh show "$ip" dev "$dev" \ + | grep "$ip" \ + | grep "$mac" + ) + else + out=$( + ip netns exec "$ns" \ + ip neigh show "$ip" dev "$dev" \ + | grep "$ip" \ + | grep "$mac" + ) + fi + + if [[ $out'X' == "X" ]]; then + exit_error "FAIL: '$ip' is not '$mac'" + fi +} + +cleanup_env() { + ip link del $VETH_HOST + ip netns del $NS_TST0 + ip netns del $NS_TST1 + ip netns del $NS_PHY +} + +exit_error() { + echo "$1" + exit 1 +} + +test_check_mac() { + # All IPs in NS_PHY should have MAC of the host + check_mac_eq $NS_PHY $IP_TST0 $MAC_HOST $VETH_PHY + check_mac_eq $NS_PHY $IP6_TST0 $MAC_HOST $VETH_PHY + check_mac_eq $NS_PHY $IP_TST1 $MAC_HOST $VETH_PHY + check_mac_eq $NS_PHY $IP6_TST1 $MAC_HOST $VETH_PHY + check_mac_eq $NS_PHY $IP_HOST $MAC_HOST $VETH_PHY + check_mac_eq $NS_PHY $IP6_HOST $MAC_HOST $VETH_PHY + + # All IPs in TST0 should have corresponding MAC + check_mac_eq $NS_TST0 $IP_HOST $MAC_HOST tap0.0 + check_mac_eq $NS_TST0 $IP6_HOST $MAC_HOST tap0.0 + check_mac_eq $NS_TST0 $IP_TST1 $MAC_TST1 tap0.0 + check_mac_eq $NS_TST0 $IP6_TST1 $MAC_TST1 tap0.0 + check_mac_eq $NS_TST0 $IP_PHY $MAC_PHY tap0.0 + check_mac_eq $NS_TST0 $IP6_PHY $MAC_PHY tap0.0 + + # All IPs in host should have corresponding MAC + check_mac_eq default $IP_TST0 $MAC_TST0 $VETH_HOST + check_mac_eq default $IP6_TST0 $MAC_TST0 $VETH_HOST + check_mac_eq default $IP_TST1 $MAC_TST1 $VETH_HOST + check_mac_eq default $IP6_TST1 $MAC_TST1 $VETH_HOST + check_mac_eq default $IP_PHY $MAC_PHY $VETH_HOST + check_mac_eq default $IP6_PHY $MAC_PHY $VETH_HOST +} + +test_ip_add() { + # adding IPs to ipvtap should be forbidden and should fail + if ns_run $NS_TST0 ip a a 172.26.0.1/24 dev ipvtap0.0; then + exit_error "FAIL: Module allowed to add ip to ipvtap." + fi + + if ns_run $NS_TST0 ip a a fc01::1/64 dev ipvtap0.0; then + exit_error "FAIL: Module allowed to add ip6 to ipvtap." + fi +} + +test_ip_overflow() { + # The ipvtap remembers limited number of addresses on interface. + # Let's overflow it and check that oldest one doesn't work. + + ns_run $NS_TST0 ip addr flush dev tap0.0 + + # Add exactly 4 ip addresses + for ip in "${IP_OK0[@]}"; do + ns_run $NS_TST0 ip a a "$ip/24" dev tap0.0 + ns_run $NS_TST0 ping -c 1 $IP_HOST -I "$ip" + done + + # Initial check that ping works + if ! ping -c 2 $IP_TST0; then + exit_error "FAIL: Failed to ping tst0" + fi + + # Add 1 more ip addresses + ns_run "$NS_TST0" ip a a $IP_OVFL0/24 dev tap0.0 + ns_run $NS_TST0 ping -c 1 $IP_HOST -I $IP_OVFL0 + # check that ping to oldest one from host fails. + echo "the next ping should fail:" + if ping -c 2 $IP_TST0; then + exit_error "FAIL: IP-0 still exists on interface" + fi + + # ping host using address-0 and force relearn of IP0. + # Host should be able ping after that + ns_run $NS_TST0 ping -c 1 $IP_HOST -I $IP_TST0 + + if ! ping -c 2 $IP_TST0; then + exit_error "FAIL: Failed to ping tst0 at stage 3" + fi +} + +test_ip6_overflow() { + # The ipvtap stores limited number of addresses on interface. + # Let's overflow it and check that oldest one doesn't work. + + ns_run $NS_TST0 ip addr flush dev tap0.0 + + # Add exactly 4 ip addresses + for ip6 in "${IP6_OK0[@]}"; do + ns_run $NS_TST0 ip a a "$ip6/64" dev tap0.0 + ns_run $NS_TST0 ping -c 1 $IP6_HOST -I "$ip6" + done + + # Initial check that ping6 works + if ! ping -c 2 $IP6_TST0; then + exit_error "FAIL: Failed to ping6 tst0" + fi + + # Add 1 more ip6 addresses + ns_run $NS_TST0 ip a a $IP6_OVFL0/64 dev tap0.0 + ns_run $NS_TST0 ping -c 1 $IP6_HOST -I $IP6_OVFL0 + # check that ping to oldest one from host fails. + echo "the next ping should fail:" + if ping -c 2 $IP6_TST0; then + exit_error "FAIL: IP6-0 still exists on interface" + fi + + # ping host using address-0 and force relearn of IP0. + # Host should be able ping after that + ns_run $NS_TST0 ping -c 1 $IP6_HOST -I $IP6_TST0 + if ! ping -c 2 $IP6_TST0; then + exit_error "FAIL: Failed to ping6 tst0 at stage 3" + fi +} + +exec_test() { + echo "TEST: $2" + $1 + echo "PASSED: $2" +} + +trap cleanup_env EXIT + +echo "ipvlan macnat tests" +echo "===================" + +modprobe -q tap +modprobe -q ipvlan +modprobe -q ipvtap + +setup_env + +exec_test test_ip_add "ip add not allowed" + +start_macnat_bridge $NS_TST0 0 +mb_pid1=$! +start_macnat_bridge $NS_TST1 1 +mb_pid2=$! + +echo "<<< Preparation: pinging all...." +ping_all default +ping_all $NS_TST0 +ping_all $NS_TST1 +ping_all $NS_PHY +echo "Finished preparational pinging all. >>>" + +exec_test test_check_mac "mac correctness" +exec_test test_ip_overflow "ip learn capacity overflow" +exec_test test_ip6_overflow "ip6 learn capacity overflow" + +kill -INT $mb_pid1 +kill -INT $mb_pid2 +wait $mb_pid1 +wait $mb_pid2 + +echo "All tests passed" -- 2.25.1

4 days, 4 hours

1
0
0 0

[RFC PATCH 00/13] Introduce per-vCPU vLPI injection control API

by Maximilian Dittgen

At the moment, the ability to direct-inject vLPIs is only enableable on an all-or-nothing per-VM basis, causing unnecessary I/O performance loss in cases where a VM's vCPU count exceeds available vPEs. This RFC introduces per-vCPU control over vLPI injection to realize potential I/O performance gain in such situations. Background ---------- The value of dynamically enabling the direct injection of vLPIs on a per-vCPU basis is the ability to run guest VMs with simultaneous hardware-forwarded and software-forwarded message-signaled interrupts. Currently, hardware-forwarded vLPI direct injection on a KVM guest requires GICv4 and is enabled on a per-VM, all-or-nothing basis. vLPI injection enablment happens in two stages: 1) At vGIC initialization, allocate direct injection structures for each vCPU (doorbell IRQ, vPE table entry, virtual pending table, vPEID). 2) When a PCI device is configured for passthrough, map its MSIs to vLPIs using the structures allocated in step 1. Step 1 is all-or-nothing; if any vCPU cannot be configured with the vPE structures necessary for direct injection, the vPEs of all vCPUs are torn down and direct injection is disabled VM-wide. This universality of direct vLPI injection enablement sparks several issues, with the most pressing being performance degradation on overcommitted hosts. VM-wide vLPI enablement creates resource inefficiency when guest VMs have more vCPUs than the host has available vPEIDs. The amount of vPEIDs (and consequently, vPEs) a host can allocate is constrained by hardware and defined by GICD_TYPER2.VID + 1 (ITS_MAX_VPEID). Since direct injection requires a vCPU to be assigned a vPEID, at most ITS_MAX_VPEID vCPUs can be configured for direct injection at a time. Because vLPI direct injection is all-or-nothing on a VM, if a new guest VM would exhaust remaining vPEIDs, all vCPUs on that VM would fall back to hypervisor-forwarded LPIs, causing considerable I/O performance degradation. Such performance degradation is exemplified on hosts with CPU overcommitment. Overcommitting an arbitrarily high number of vCPUs enables a VM's vCPU count to easily exceed the host's available vPEIDs. Even with marginally more vCPUs than vPEIDs, the current all-or-nothing vLPI paradigm disables direct injection entirely. This creates two problems: first, a single many-vCPU overcommitted VM loses all direct injection despite having vPEIDs available; second, on multi-tenant hosts, VMs booted first consume all vPEIDs, leaving later VMs without direct injection regardless of their I/O intensity. Per-vCPU control would allow userspace to allocate available vPEIDs across VMs based on I/O workload rather than boot order or per-VM vCPU count. This per-vCPU granularity recovers most of the direct injection performance benefit instead of losing it completely. To allow this per-vCPU granularity, this RFC introduces three new ioctls to the KVM API that enables userspace the ability to activate/deactivate direct vLPI injection capability and resources to vCPUs ad-hoc during VM runtime. This RFC proposes userspace control, rather than kernel control, over vPEID allocation for simplicity of implementation, ease of testability, and autonomy over resource usage. In the future, the vLPI enable/disable building blocks from this RFC may be used to implement a full vPE allocation policy in the kernel. The solution comes in several parts ----------------------------------- 1) [P 1] General declarations (ioctl definitions/stubs, kconfig option) 2) [P 2] Conditionally disable auto vLPI injection init routines To prevent vCPUs from exceeding vPEID allocation limits upon VM boot, disable automatic vPEID allocation in the GICv4 initialization routine when the per-vCPU kconfig is active. Likewise, disable automatic hardware forwarding for PCI device-backed MSIs upon device registration. 3) [P 3-6] Implement per-vCPU vLPI enablement routine, which: a) Creates per-vCPU doorbell IRQ on new vCPU-scoped, rather than VM-scoped, interrupt domain hierarchies. b) Allocates per-vCPU vPE table entries and virtual pending table, linking them to the vCPU's doorbell IRQ. c) Iterates through interrupt translation table to set hardware forwarding for all PCI device–backed interrupts targeting the specific vCPU. 3) [P 7-8] Implement per-vCPU vLPI disablement routine, which a) Iterates through interrupt translation table to unset hardware forwarding for all interrupts targeting the specific vCPU. b) Frees per-vCPU vPE table entries, virtual pending table, and doorbell IRQ, then removes vgic_dist's pointer to the vCPU's freed vPE. 4) [P 9] Couple vSGI enablement with per-vCPU vPE allocation Since vSGIs cannot be direct-injected without an allocated vPE on the receiving vCPU, couple vSGI enablement with vLPI enablement on GICv4.1. 5) [P 10-13] Write selftests for vLPI direct injection PCI devices cannot be passed through to selftest guests, so define an ioctl that mocks a hardware source for software-defined MSI interrupts and sets vLPI "hardware" forwarding for the MSIs. Use these vLPIs to selftest per-vCPU vLPI enablement/disablement ioctls. Testing ------- Testing has been carried out via selftests and QEMU-emulated guests. Selftests have covered diverse vLPI configurations and race conditions. These include: 1) Stress testing LPI injection across multiple vCPUs while concurrently and repeatedly toggling the vCPUs' vLPI injection capability. 2) Enabling/disabling vLPI direct injection while scheduling or unscheduling a vCPU. 3) Allocating and freeing a single vPEID to multiple vCPUs, ensuring reusability. 4) Attempting to allocate a vPEID when all are already allocated, validating an error is thrown. 5) Calling enable/disable vLPI ioctls when GIC is not initialized. 6) Idempotent ioctl calls. PCI device passthrough and interrupt injection to QEMU guest demonstrated: 1) Complete hypervisor circumvention when vLPI injection is enabled on a vCPU, hypervisor forwarding when vLPI injection is disabled. 2) Interrupts are not lost when received during per-vCPU vLPI state transitions. Caveats ------- 1) Pending interrupts are flushed when vLPI injection is disabled for a vCPU; hardware pending state is not transfered to software. This may cause pending interrupts to be lost upon vPE disablement. Unlike vSGIs, vLPIs do not expose their pending state through a GICD_ISPENDR register. Thus, we would need to read the pending state of the vLPI from the vPT. To read the pending status of the vLPI from vPT, we would need to invalidate any vPT cache associated with the vCPU's vPE. This requires unmapping the vPE and halting the vCPU, which would be incredibly expensive and unecessary given that MSIs are usually recoverable by the driver. 2) Direct-injected vSGIs (GICv4.1) require vCPUs to have associated vPEs. Since disabling vLPI injection on a vCPU frees its vPE, vSGI direct injection must simultaenously be disabled as well. At the moment, we use the per-vCPU vSGI toggle mechanism introduced in commit bacf2c6 to enable/disable vSGI injection alongside vLPI injection. Maximilian Dittgen (13): KVM: Introduce config option for per-vCPU vLPI enablement KVM: arm64: Disable auto vCPU vPE assignment with per-vCPU vLPI config KVM: arm64: Refactor out locked section of kvm_vgic_v4_set_forwarding() KVM: arm64: Implement vLPI QUERY ioctl for per-vCPU vLPI injection API KVM: arm64: Implement vLPI ENABLE ioctl for per-vCPU vLPI injection API KVM: arm64: Resolve race between vCPU scheduling and vLPI enablement KVM: arm64: Implement vLPI DISABLE ioctl for per-vCPU vLPI Injection API KVM: arm64: Make per-vCPU vLPI control ioctls atomic KVM: arm64: Couple vSGI enablement with per-vCPU vPE allocation KVM: selftests: fix MAPC RDbase target formatting in vgic_lpi_stress KVM: Ioctl to set up userspace-injected MSIs as software-bypassing vLPIs KVM: arm64: selftests: Add support for stress testing direct-injected vLPIs KVM: arm64: selftests: Add test for per-vCPU vLPI control API Documentation/virt/kvm/api.rst | 56 +++ arch/arm64/kvm/arm.c | 89 +++++ arch/arm64/kvm/vgic/vgic-its.c | 142 ++++++- arch/arm64/kvm/vgic/vgic-v3.c | 14 +- arch/arm64/kvm/vgic/vgic-v4.c | 370 +++++++++++++++++- arch/arm64/kvm/vgic/vgic.h | 10 + drivers/irqchip/Kconfig | 13 + drivers/irqchip/irq-gic-v3-its.c | 58 ++- drivers/irqchip/irq-gic-v4.c | 75 +++- include/kvm/arm_vgic.h | 8 + include/linux/irqchip/arm-gic-v3.h | 5 + include/linux/irqchip/arm-gic-v4.h | 10 +- include/linux/kvm_host.h | 11 + include/uapi/linux/kvm.h | 22 ++ tools/testing/selftests/kvm/Makefile.kvm | 1 + .../selftests/kvm/arm64/per_vcpu_vlpi.c | 274 +++++++++++++ .../selftests/kvm/arm64/vgic_lpi_stress.c | 181 ++++++++- .../selftests/kvm/lib/arm64/gic_v3_its.c | 9 +- 18 files changed, 1307 insertions(+), 41 deletions(-) create mode 100644 tools/testing/selftests/kvm/arm64/per_vcpu_vlpi.c -- 2.50.1 (Apple Git-155) Amazon Web Services Development Center Germany GmbH Tamara-Danz-Str. 13 10243 Berlin Geschaeftsfuehrung: Christian Schlaeger, Christof Hellmis Eingetragen am Amtsgericht Charlottenburg unter HRB 257764 B Sitz: Berlin Ust-ID: DE 365 538 597

4 days, 5 hours

2
15
0 0

[PATCH net-next v5 0/5] netconsole: support automatic target recovery

by Andre Carvalho

This patchset introduces target resume capability to netconsole allowing it to recover targets when underlying low-level interface comes back online. The patchset starts by refactoring netconsole state representation in order to allow representing deactivated targets (targets that are disabled due to interfaces going down). It then modifies netconsole to handle NETDEV_UP events for such targets and setups netpoll. Targets are matched with incoming interfaces depending on how they were initially bound in netconsole (by mac or interface name). The patchset includes a selftest that validates netconsole target state transitions and that target is functional after resumed. Signed-off-by: Andre Carvalho <asantostc(a)gmail.com> --- Changes in v5: - patch 3: Set (de)enslaved target as DISABLED instead of DEACTIVATED to prevent resuming it. - selftest: Fix test cleanup by moving trap line to outside of loop and remove unneeded 'local' keyword - Rename maybe_resume_target to resume_target, add netconsole_ prefix to process_resumable_targets. - Hold device reference before calling __netpoll_setup. - Link to v4: https://lore.kernel.org/r/20251116-netcons-retrigger-v4-0-5290b5f140c2@gmai… Changes in v4: - Simplify selftest cleanup, removing trap setup in loop. - Drop netpoll helper (__setup_netpoll_hold) and manage reference inside netconsole. - Move resume_list processing logic to separate function. - Link to v3: https://lore.kernel.org/r/20251109-netcons-retrigger-v3-0-1654c280bbe6@gmai… Changes in v3: - Resume by mac or interface name depending on how target was created. - Attempt to resume target without holding target list lock, by moving the target to a temporary list. This is required as netpoll may attempt to allocate memory. - Link to v2: https://lore.kernel.org/r/20250921-netcons-retrigger-v2-0-a0e84006237f@gmai… Changes in v2: - Attempt to resume target in the same thread, instead of using workqueue . - Add wrapper around __netpoll_setup (patch 4). - Renamed resume_target to maybe_resume_target and moved conditionals to inside its implementation, keeping code more clear. - Verify that device addr matches target mac address when target was setup using mac. - Update selftest to cover targets bound by mac and interface name. - Fix typo in selftest comment and sort tests alphabetically in Makefile. - Link to v1: https://lore.kernel.org/r/20250909-netcons-retrigger-v1-0-3aea904926cf@gmai… --- Andre Carvalho (3): netconsole: convert 'enabled' flag to enum for clearer state management netconsole: resume previously deactivated target selftests: netconsole: validate target resume Breno Leitao (2): netconsole: add target_state enum netconsole: add STATE_DEACTIVATED to track targets disabled by low level drivers/net/netconsole.c | 155 +++++++++++++++++---- tools/testing/selftests/drivers/net/Makefile | 1 + .../selftests/drivers/net/lib/sh/lib_netcons.sh | 35 ++++- .../selftests/drivers/net/netcons_resume.sh | 97 +++++++++++++ 4 files changed, 254 insertions(+), 34 deletions(-) --- base-commit: a057e8e4ac5b1ddd12be590e2e039fa08d0c8aa4 change-id: 20250816-netcons-retrigger-a4f547bfc867 Best regards, -- Andre Carvalho <asantostc(a)gmail.com>

4 days, 6 hours

2
7
0 0

[PATCH RFC 0/4] landlock: add LANDLOCK_SCOPE_MEMFD_EXEC execution

by Abhinav Saxena

This patch series introduces LANDLOCK_SCOPE_MEMFD_EXEC, a new Landlock scoping mechanism that restricts execution of anonymous memory file descriptors (memfd) created via memfd_create(2). This addresses security gaps where processes can bypass W^X policies and execute arbitrary code through anonymous memory objects. Fixes: https://github.com/landlock-lsm/linux/issues/37 SECURITY PROBLEM ================ Current Landlock filesystem restrictions do not cover memfd objects, allowing processes to: 1. Read-to-execute bypass: Create writable memfd, inject code, then execute via mmap(PROT_EXEC) or direct execve() 2. Anonymous execution: Execute code without touching the filesystem via execve("/proc/self/fd/N") where N is a memfd descriptor 3. Cross-domain access violations: Pass memfd between processes to bypass domain restrictions These scenarios can occur in sandboxed environments where filesystem access is restricted but memfd creation remains possible. IMPLEMENTATION ============== The implementation adds hierarchical execution control through domain scoping: Core Components: - is_memfd_file(): Reliable memfd detection via "memfd:" dentry prefix - domain_is_scoped(): Cross-domain hierarchy checking (moved to domain.c) - LSM hooks: mmap_file, file_mprotect, bprm_creds_for_exec - Creation-time restrictions: hook_file_alloc_security Security Matrix: Execution decisions follow domain hierarchy rules preventing both same-domain bypass attempts and cross-domain access violations while preserving legitimate hierarchical access patterns. Domain Hierarchy with LANDLOCK_SCOPE_MEMFD_EXEC: =============================================== Root (no domain) - No restrictions | +-- Domain A [SCOPE_MEMFD_EXEC] Layer 1 | +-- memfd_A (tagged with Domain A as creator) | | | +-- Domain A1 (child) [NO SCOPE] Layer 2 | | +-- Inherits Layer 1 restrictions from parent | | +-- memfd_A1 (can create, inherits restrictions) | | +-- Domain A1a [SCOPE_MEMFD_EXEC] Layer 3 | | +-- memfd_A1a (tagged with Domain A1a) | | | +-- Domain A2 (child) [SCOPE_MEMFD_EXEC] Layer 2 | +-- memfd_A2 (tagged with Domain A2 as creator) | +-- CANNOT access memfd_A1 (different subtree) | +-- Domain B [SCOPE_MEMFD_EXEC] Layer 1 +-- memfd_B (tagged with Domain B as creator) +-- CANNOT access ANY memfd from Domain A subtree Execution Decision Matrix: ======================== Executor-> | A | A1 | A1a | A2 | B | Root Creator | | | | | | ------------|-----|----|-----|----|----|----- Domain A | X | X | X | X | X | Y Domain A1 | Y | X | X | X | X | Y Domain A1a | Y | Y | X | X | X | Y Domain A2 | Y | X | X | X | X | Y Domain B | X | X | X | X | X | Y Root | Y | Y | Y | Y | Y | Y Legend: Y = Execution allowed, X = Execution denied Scenarios Covered: - Direct mmap(PROT_EXEC) on memfd files - Two-stage mmap(PROT_READ) + mprotect(PROT_EXEC) bypass attempts - execve("/proc/self/fd/N") anonymous execution - execveat() and fexecve() file descriptor execution - Cross-process memfd inheritance and IPC passing TESTING ======= All patches have been validated with: - scripts/checkpatch.pl --strict (clean) - Selftests covering same-domain restrictions, cross-domain hierarchy enforcement, and regular file isolation - KUnit tests for memfd detection edge cases DISCLAIMER ========== My understanding of Landlock scoping semantics may be limited, but this implementation reflects my current understanding based on available documentation and code analysis. I welcome feedback and corrections regarding the scoping logic and domain hierarchy enforcement. Signed-off-by: Abhinav Saxena <xandfury(a)gmail.com> --- Abhinav Saxena (4): landlock: add LANDLOCK_SCOPE_MEMFD_EXEC scope landlock: implement memfd detection landlock: add memfd exec LSM hooks and scoping selftests/landlock: add memfd execution tests include/uapi/linux/landlock.h | 5 + security/landlock/.kunitconfig | 1 + security/landlock/audit.c | 4 + security/landlock/audit.h | 1 + security/landlock/cred.c | 14 - security/landlock/domain.c | 67 ++++ security/landlock/domain.h | 4 + security/landlock/fs.c | 405 ++++++++++++++++++++- security/landlock/limits.h | 2 +- security/landlock/task.c | 67 ---- .../selftests/landlock/scoped_memfd_exec_test.c | 325 +++++++++++++++++ 11 files changed, 812 insertions(+), 83 deletions(-) --- base-commit: 5b74b2eff1eeefe43584e5b7b348c8cd3b723d38 change-id: 20250716-memfd-exec-ac0d582018c3 Best regards, -- Abhinav Saxena <xandfury(a)gmail.com>

4 days, 6 hours

3
11
0 0

[PATCH v2 00/14] selftests: vDSO: Stop using libc types for vDSO calls

by Thomas Weißschuh

Currently the vDSO selftests use the time-related types from libc. This works on glibc by chance today but will break with other libc implementations or on distributions which switch to 64-bit times everywhere. The kernel's UAPI headers provide the proper types to use with the vDSO (and raw syscalls) but are not necessarily compatible with libc types. Introduce a new header which makes the UAPI headers compatible with the libc. Also contains some related cleanups. Signed-off-by: Thomas Weißschuh <thomas.weissschuh(a)linutronix.de> --- Changes in v2: - Use __kernel_old_time_t in vdso_time_t. - Add vdso_syscalls.h. - Add a test for the time() function. - Validate return value of syscall(clock_getres) in vdso_test_abi - Link to v1: https://lore.kernel.org/r/20251111-vdso-test-types-v1-0-03b31f88c659@linutr… --- Thomas Weißschuh (14): Revert "selftests: vDSO: parse_vdso: Use UAPI headers instead of libc headers" selftests: vDSO: Introduce vdso_types.h selftests: vDSO: Introduce vdso_syscalls.h selftests: vDSO: vdso_test_gettimeofday: Remove nolibc checks selftests: vDSO: vdso_test_gettimeofday: Use types from vdso_types.h selftests: vDSO: vdso_test_abi: Use types from vdso_types.h selftests: vDSO: vdso_test_abi: Validate return value of syscall(clock_getres) selftests: vDSO: vdso_test_abi: Use system call wrappers from vdso_syscalls.h selftests: vDSO: vdso_test_correctness: Drop SYS_getcpu fallbacks selftests: vDSO: vdso_test_correctness: Make ts_leq() and tv_leq() more generic selftests: vDSO: vdso_test_correctness: Use types from vdso_types.h selftests: vDSO: vdso_test_correctness: Use system call wrappers from vdso_syscalls.h selftests: vDSO: vdso_test_correctness: Use facilities from parse_vdso.c selftests: vDSO: vdso_test_correctness: Add a test for time() tools/testing/selftests/vDSO/Makefile | 6 +- tools/testing/selftests/vDSO/parse_vdso.c | 3 +- tools/testing/selftests/vDSO/vdso_syscalls.h | 93 ++++++++++ tools/testing/selftests/vDSO/vdso_test_abi.c | 46 +++-- .../testing/selftests/vDSO/vdso_test_correctness.c | 190 +++++++++++---------- .../selftests/vDSO/vdso_test_gettimeofday.c | 9 +- tools/testing/selftests/vDSO/vdso_types.h | 70 ++++++++ 7 files changed, 285 insertions(+), 132 deletions(-) --- base-commit: 1b2eb8c1324859864f4aa79dc3cfbb2f7ef5c524 change-id: 20251110-vdso-test-types-68ce0c712b79 Best regards, -- Thomas Weißschuh <thomas.weissschuh(a)linutronix.de>

4 days, 8 hours

2
23
0 0

[PATCH] kselftest: futex: Fix memset with zero size warning

by Wake Liu

The `FIXTURE(args)` macro defines an empty `struct _test_data_args`, leading to `sizeof(struct _test_data_args)` evaluating to 0. This caused a build error due to a compiler warning on a `memset` call with a zero size argument. Adding a dummy member to the struct ensures its size is non-zero, resolving the build issue. Signed-off-by: Wake Liu <wakel(a)google.com> --- tools/testing/selftests/futex/functional/futex_requeue_pi.c | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/testing/selftests/futex/functional/futex_requeue_pi.c b/tools/testing/selftests/futex/functional/futex_requeue_pi.c index f299d75848cd..000fec468835 100644 --- a/tools/testing/selftests/futex/functional/futex_requeue_pi.c +++ b/tools/testing/selftests/futex/functional/futex_requeue_pi.c @@ -52,6 +52,7 @@ struct thread_arg { FIXTURE(args) { + char dummy; }; FIXTURE_SETUP(args) -- 2.52.0.rc1.455.g30608eb744-goog

4 days, 10 hours

3
2
0 0

[PATCH v5 2/2] selftests/riscv: Add Zicbop prefetch test

by Yao Zihong

Add selftests to cbo.c to verify Zicbop extension behavior, and split the previous `--sigill` mode into two options so they can be tested independently. The test checks: - That hwprobe correctly reports Zicbop presence and block size. - That prefetch instructions execute without exception on valid and NULL addresses when Zicbop is present. Signed-off-by: Yao Zihong <zihong.plct(a)isrc.iscas.ac.cn> --- tools/testing/selftests/riscv/hwprobe/cbo.c | 165 ++++++++++++++++---- 1 file changed, 136 insertions(+), 29 deletions(-) diff --git a/tools/testing/selftests/riscv/hwprobe/cbo.c b/tools/testing/selftests/riscv/hwprobe/cbo.c index 5e96ef785d0d..6d99726aceac 100644 --- a/tools/testing/selftests/riscv/hwprobe/cbo.c +++ b/tools/testing/selftests/riscv/hwprobe/cbo.c @@ -15,24 +15,31 @@ #include <linux/compiler.h> #include <linux/kernel.h> #include <asm/ucontext.h> +#include <getopt.h> #include "hwprobe.h" #include "../../kselftest.h" #define MK_CBO(fn) le32_bswap((uint32_t)(fn) << 20 | 10 << 15 | 2 << 12 | 0 << 7 | 15) +#define MK_PREFETCH(fn) \ + le32_bswap(0 << 25 | (uint32_t)(fn) << 20 | 10 << 15 | 6 << 12 | 0 << 7 | 19) static char mem[4096] __aligned(4096) = { [0 ... 4095] = 0xa5 }; -static bool illegal_insn; +static bool got_fault; -static void sigill_handler(int sig, siginfo_t *info, void *context) +static void fault_handler(int sig, siginfo_t *info, void *context) { unsigned long *regs = (unsigned long *)&((ucontext_t *)context)->uc_mcontext; uint32_t insn = *(uint32_t *)regs[0]; - assert(insn == MK_CBO(regs[11])); + if (sig == SIGILL) + assert(insn == MK_CBO(regs[11])); - illegal_insn = true; + if (sig == SIGSEGV || sig == SIGBUS) + assert(insn == MK_PREFETCH(regs[11])); + + got_fault = true; regs[0] += 4; } @@ -45,39 +52,51 @@ static void sigill_handler(int sig, siginfo_t *info, void *context) : : "r" (base), "i" (fn), "i" (MK_CBO(fn)) : "a0", "a1", "memory"); \ }) +#define prefetch_insn(base, fn) \ +({ \ + asm volatile( \ + "mv a0, %0\n" \ + "li a1, %1\n" \ + ".4byte %2\n" \ + : : "r" (base), "i" (fn), "i" (MK_PREFETCH(fn)) : "a0", "a1"); \ +}) + static void cbo_inval(char *base) { cbo_insn(base, 0); } static void cbo_clean(char *base) { cbo_insn(base, 1); } static void cbo_flush(char *base) { cbo_insn(base, 2); } static void cbo_zero(char *base) { cbo_insn(base, 4); } +static void prefetch_i(char *base) { prefetch_insn(base, 0); } +static void prefetch_r(char *base) { prefetch_insn(base, 1); } +static void prefetch_w(char *base) { prefetch_insn(base, 3); } static void test_no_cbo_inval(void *arg) { ksft_print_msg("Testing cbo.inval instruction remain privileged\n"); - illegal_insn = false; + got_fault = false; cbo_inval(&mem[0]); - ksft_test_result(illegal_insn, "No cbo.inval\n"); + ksft_test_result(got_fault, "No cbo.inval\n"); } static void test_no_zicbom(void *arg) { ksft_print_msg("Testing Zicbom instructions remain privileged\n"); - illegal_insn = false; + got_fault = false; cbo_clean(&mem[0]); - ksft_test_result(illegal_insn, "No cbo.clean\n"); + ksft_test_result(got_fault, "No cbo.clean\n"); - illegal_insn = false; + got_fault = false; cbo_flush(&mem[0]); - ksft_test_result(illegal_insn, "No cbo.flush\n"); + ksft_test_result(got_fault, "No cbo.flush\n"); } static void test_no_zicboz(void *arg) { ksft_print_msg("No Zicboz, testing cbo.zero remains privileged\n"); - illegal_insn = false; + got_fault = false; cbo_zero(&mem[0]); - ksft_test_result(illegal_insn, "No cbo.zero\n"); + ksft_test_result(got_fault, "No cbo.zero\n"); } static bool is_power_of_2(__u64 n) @@ -85,6 +104,51 @@ static bool is_power_of_2(__u64 n) return n != 0 && (n & (n - 1)) == 0; } +static void test_zicbop(void *arg) +{ + struct riscv_hwprobe pair = { + .key = RISCV_HWPROBE_KEY_ZICBOP_BLOCK_SIZE, + }; + struct sigaction act = { + .sa_sigaction = &fault_handler, + .sa_flags = SA_SIGINFO + }; + struct sigaction dfl = { + .sa_handler = SIG_DFL + }; + cpu_set_t *cpus = (cpu_set_t *)arg; + __u64 block_size; + long rc; + + rc = sigaction(SIGSEGV, &act, NULL); + assert(rc == 0); + rc = sigaction(SIGBUS, &act, NULL); + assert(rc == 0); + + rc = riscv_hwprobe(&pair, 1, sizeof(cpu_set_t), (unsigned long *)cpus, 0); + block_size = pair.value; + ksft_test_result(rc == 0 && pair.key == RISCV_HWPROBE_KEY_ZICBOP_BLOCK_SIZE && + is_power_of_2(block_size), "Zicbop block size\n"); + ksft_print_msg("Zicbop block size: %llu\n", block_size); + + got_fault = false; + prefetch_i(&mem[0]); + prefetch_r(&mem[0]); + prefetch_w(&mem[0]); + ksft_test_result(!got_fault, "Zicbop prefetch.* on valid address\n"); + + got_fault = false; + prefetch_i(NULL); + prefetch_r(NULL); + prefetch_w(NULL); + ksft_test_result(!got_fault, "Zicbop prefetch.* on NULL\n"); + + rc = sigaction(SIGBUS, &dfl, NULL); + assert(rc == 0); + rc = sigaction(SIGSEGV, &dfl, NULL); + assert(rc == 0); +} + static void test_zicbom(void *arg) { struct riscv_hwprobe pair = { @@ -100,13 +164,13 @@ static void test_zicbom(void *arg) is_power_of_2(block_size), "Zicbom block size\n"); ksft_print_msg("Zicbom block size: %llu\n", block_size); - illegal_insn = false; + got_fault = false; cbo_clean(&mem[block_size]); - ksft_test_result(!illegal_insn, "cbo.clean\n"); + ksft_test_result(!got_fault, "cbo.clean\n"); - illegal_insn = false; + got_fault = false; cbo_flush(&mem[block_size]); - ksft_test_result(!illegal_insn, "cbo.flush\n"); + ksft_test_result(!got_fault, "cbo.flush\n"); } static void test_zicboz(void *arg) @@ -125,11 +189,11 @@ static void test_zicboz(void *arg) is_power_of_2(block_size), "Zicboz block size\n"); ksft_print_msg("Zicboz block size: %llu\n", block_size); - illegal_insn = false; + got_fault = false; cbo_zero(&mem[block_size]); - ksft_test_result(!illegal_insn, "cbo.zero\n"); + ksft_test_result(!got_fault, "cbo.zero\n"); - if (illegal_insn || !is_power_of_2(block_size)) { + if (got_fault || !is_power_of_2(block_size)) { ksft_test_result_skip("cbo.zero check\n"); return; } @@ -177,7 +241,19 @@ static void check_no_zicbo_cpus(cpu_set_t *cpus, __u64 cbo) rc = riscv_hwprobe(&pair, 1, sizeof(cpu_set_t), (unsigned long *)&one_cpu, 0); assert(rc == 0 && pair.key == RISCV_HWPROBE_KEY_IMA_EXT_0); - cbostr = cbo == RISCV_HWPROBE_EXT_ZICBOZ ? "Zicboz" : "Zicbom"; + switch (cbo) { + case RISCV_HWPROBE_EXT_ZICBOZ: + cbostr = "Zicboz"; + break; + case RISCV_HWPROBE_EXT_ZICBOM: + cbostr = "Zicbom"; + break; + case RISCV_HWPROBE_EXT_ZICBOP: + cbostr = "Zicbop"; + break; + default: + ksft_exit_fail_msg("Internal error: invalid cbo %llu\n", cbo); + } if (pair.value & cbo) ksft_exit_fail_msg("%s is only present on a subset of harts.\n" @@ -194,6 +270,7 @@ enum { TEST_ZICBOM, TEST_NO_ZICBOM, TEST_NO_CBO_INVAL, + TEST_ZICBOP, }; static struct test_info { @@ -206,26 +283,51 @@ static struct test_info { [TEST_ZICBOM] = { .nr_tests = 3, test_zicbom }, [TEST_NO_ZICBOM] = { .nr_tests = 2, test_no_zicbom }, [TEST_NO_CBO_INVAL] = { .nr_tests = 1, test_no_cbo_inval }, + [TEST_ZICBOP] = { .nr_tests = 3, test_zicbop }, +}; + +static const struct option long_opts[] = { + {"zicbom-raises-sigill", no_argument, 0, 'm'}, + {"zicboz-raises-sigill", no_argument, 0, 'z'}, + {0, 0, 0, 0} }; int main(int argc, char **argv) { struct sigaction act = { - .sa_sigaction = &sigill_handler, + .sa_sigaction = &fault_handler, .sa_flags = SA_SIGINFO, }; struct riscv_hwprobe pair; unsigned int plan = 0; cpu_set_t cpus; long rc; - int i; - - if (argc > 1 && !strcmp(argv[1], "--sigill")) { - rc = sigaction(SIGILL, &act, NULL); - assert(rc == 0); - tests[TEST_NO_ZICBOZ].enabled = true; - tests[TEST_NO_ZICBOM].enabled = true; - tests[TEST_NO_CBO_INVAL].enabled = true; + int i, opt, long_index; + + long_index = 0; + + while ((opt = getopt_long(argc, argv, "mz", long_opts, &long_index)) != -1) { + switch (opt) { + case 'm': + tests[TEST_NO_ZICBOM].enabled = true; + tests[TEST_NO_CBO_INVAL].enabled = true; + rc = sigaction(SIGILL, &act, NULL); + assert(rc == 0); + break; + case 'z': + tests[TEST_NO_ZICBOZ].enabled = true; + tests[TEST_NO_CBO_INVAL].enabled = true; + rc = sigaction(SIGILL, &act, NULL); + assert(rc == 0); + break; + case '?': + fprintf(stderr, + "Usage: %s [--zicbom-raises-sigill|-m] [--zicboz-raises-sigill|-z]\n", + argv[0]); + exit(1); + default: + break; + } } rc = sched_getaffinity(0, sizeof(cpu_set_t), &cpus); @@ -253,6 +355,11 @@ int main(int argc, char **argv) check_no_zicbo_cpus(&cpus, RISCV_HWPROBE_EXT_ZICBOM); } + if (pair.value & RISCV_HWPROBE_EXT_ZICBOP) + tests[TEST_ZICBOP].enabled = true; + else + check_no_zicbo_cpus(&cpus, RISCV_HWPROBE_EXT_ZICBOP); + for (i = 0; i < ARRAY_SIZE(tests); ++i) plan += tests[i].enabled ? tests[i].nr_tests : 0; -- 2.47.2

4 days, 10 hours

2
2
0 0

[PATCH v3 6/6] KVM: LoongArch: selftests: Add time counter test

by Bibo Mao

With time counter test, it is to verify that time count starts from 0 and always grows up then. Signed-off-by: Bibo Mao <maobibo(a)loongson.cn> --- .../selftests/kvm/lib/loongarch/processor.c | 9 ++++++ .../selftests/kvm/loongarch/arch_timer.c | 29 +++++++++++++++++++ 2 files changed, 38 insertions(+) diff --git a/tools/testing/selftests/kvm/lib/loongarch/processor.c b/tools/testing/selftests/kvm/lib/loongarch/processor.c index 436990258068..ac2ffd076bff 100644 --- a/tools/testing/selftests/kvm/lib/loongarch/processor.c +++ b/tools/testing/selftests/kvm/lib/loongarch/processor.c @@ -3,6 +3,7 @@ #include <assert.h> #include <linux/compiler.h> +#include <asm/kvm.h> #include "kvm_util.h" #include "processor.h" #include "ucall_common.h" @@ -256,6 +257,11 @@ static void loongarch_set_csr(struct kvm_vcpu *vcpu, uint64_t id, uint64_t val) __vcpu_set_reg(vcpu, csrid, val); } +static void loongarch_set_reg(struct kvm_vcpu *vcpu, uint64_t id, uint64_t val) +{ + __vcpu_set_reg(vcpu, id, val); +} + static void loongarch_vcpu_setup(struct kvm_vcpu *vcpu) { int width; @@ -279,6 +285,9 @@ static void loongarch_vcpu_setup(struct kvm_vcpu *vcpu) loongarch_set_csr(vcpu, LOONGARCH_CSR_ECFG, 0); loongarch_set_csr(vcpu, LOONGARCH_CSR_TCFG, 0); loongarch_set_csr(vcpu, LOONGARCH_CSR_ASID, 1); + /* time count start from 0 */ + val = 0; + loongarch_set_reg(vcpu, KVM_REG_LOONGARCH_COUNTER, val); val = 0; width = vm->page_shift - 3; diff --git a/tools/testing/selftests/kvm/loongarch/arch_timer.c b/tools/testing/selftests/kvm/loongarch/arch_timer.c index ff6c141f0e2b..d1cdf7f4ae55 100644 --- a/tools/testing/selftests/kvm/loongarch/arch_timer.c +++ b/tools/testing/selftests/kvm/loongarch/arch_timer.c @@ -132,10 +132,39 @@ static void guest_test_emulate_timer(uint32_t cpu) local_irq_enable(); } +static void guest_time_count_test(uint32_t cpu) +{ + uint32_t config_iter; + unsigned long start, end, prev, us; + + /* Assuming that test case starts to run in 1 second */ + start = timer_get_cycles(); + us = msec_to_cycles(1000); + __GUEST_ASSERT(start <= us, + "start = 0x%lx, us = 0x%lx.\n", + start, us); + + us = msec_to_cycles(test_args.timer_period_ms); + for (config_iter = 0; config_iter < test_args.nr_iter; config_iter++) { + start = timer_get_cycles(); + end = start + us; + /* test time count growing up always */ + while (start < end) { + prev = start; + start = timer_get_cycles(); + __GUEST_ASSERT(prev <= start, + "prev = 0x%lx, start = 0x%lx.\n", + prev, start); + } + } +} + static void guest_code(void) { uint32_t cpu = guest_get_vcpuid(); + /* must run at first */ + guest_time_count_test(cpu); timer_irq_enable(); local_irq_enable(); guest_test_oneshot_timer(cpu); -- 2.39.3

4 days, 15 hours

1
0
0 0

[PATCH v3 5/6] KVM: LoongArch: selftests: Add SW emulated timer test

by Bibo Mao

This test case setup one-shot timer and execute idle instruction immediately to indicate giving up CPU, hypervisor will emulate SW hrtimer and wakeup vCPU when SW hrtimer is fired. Signed-off-by: Bibo Mao <maobibo(a)loongson.cn> --- .../selftests/kvm/loongarch/arch_timer.c | 40 +++++++++++++++++++ 1 file changed, 40 insertions(+) diff --git a/tools/testing/selftests/kvm/loongarch/arch_timer.c b/tools/testing/selftests/kvm/loongarch/arch_timer.c index a8b7ff05faf6..ff6c141f0e2b 100644 --- a/tools/testing/selftests/kvm/loongarch/arch_timer.c +++ b/tools/testing/selftests/kvm/loongarch/arch_timer.c @@ -93,6 +93,45 @@ static void guest_test_period_timer(uint32_t cpu) irq_iter); } +static void do_idle(void) +{ + unsigned int intid; + unsigned long estat; + + __asm__ __volatile__("idle 0" : : : "memory"); + + estat = csr_read(LOONGARCH_CSR_ESTAT); + intid = !!(estat & BIT(INT_TI)); + + /* Make sure pending timer IRQ arrived */ + GUEST_ASSERT_EQ(intid, 1); + csr_write(CSR_TINTCLR_TI, LOONGARCH_CSR_TINTCLR); +} + +static void guest_test_emulate_timer(uint32_t cpu) +{ + uint32_t config_iter; + uint64_t xcnt_diff_us, us; + struct test_vcpu_shared_data *shared_data = &vcpu_shared_data[cpu]; + + local_irq_disable(); + shared_data->nr_iter = 0; + us = msecs_to_usecs(test_args.timer_period_ms); + for (config_iter = 0; config_iter < test_args.nr_iter; config_iter++) { + shared_data->xcnt = timer_get_cycles(); + + /* Setup the next interrupt */ + timer_set_next_cmp_ms(test_args.timer_period_ms, false); + do_idle(); + + xcnt_diff_us = cycles_to_usec(timer_get_cycles() - shared_data->xcnt); + __GUEST_ASSERT(xcnt_diff_us >= us, + "xcnt_diff_us = 0x%lx, us = 0x%lx.\n", + xcnt_diff_us, us); + } + local_irq_enable(); +} + static void guest_code(void) { uint32_t cpu = guest_get_vcpuid(); @@ -101,6 +140,7 @@ static void guest_code(void) local_irq_enable(); guest_test_oneshot_timer(cpu); guest_test_period_timer(cpu); + guest_test_emulate_timer(cpu); GUEST_DONE(); } -- 2.39.3

4 days, 15 hours

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-kselftest-mirror