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

[bpf-next v1 3/5] selftests/bpf: move common TCP helpers into bpf_tracing_net.h

by Hoyeon Lee

Some BPF selftests contain identical copies of the min(), max(), before(), and after() helpers. These repeated snippets are the same across the tests and do not need to be defined separately. Move these helpers into bpf_tracing_net.h so they can be shared by TCP related BPF programs. This removes repeated code and keeps the helpers in a single place. Signed-off-by: Hoyeon Lee <hoyeon.lee(a)suse.com> --- tools/testing/selftests/bpf/progs/bpf_cc_cubic.c | 9 --------- tools/testing/selftests/bpf/progs/bpf_cubic.c | 7 ------- tools/testing/selftests/bpf/progs/bpf_dctcp.c | 6 ------ tools/testing/selftests/bpf/progs/bpf_tracing_net.h | 11 +++++++++++ .../selftests/bpf/progs/tcp_ca_write_sk_pacing.c | 2 -- 5 files changed, 11 insertions(+), 24 deletions(-) diff --git a/tools/testing/selftests/bpf/progs/bpf_cc_cubic.c b/tools/testing/selftests/bpf/progs/bpf_cc_cubic.c index 4e51785e7606..9af19dfe4e80 100644 --- a/tools/testing/selftests/bpf/progs/bpf_cc_cubic.c +++ b/tools/testing/selftests/bpf/progs/bpf_cc_cubic.c @@ -22,10 +22,6 @@ #define TCP_PACING_CA_RATIO (120) #define TCP_REORDERING (12) -#define min(a, b) ((a) < (b) ? (a) : (b)) -#define max(a, b) ((a) > (b) ? (a) : (b)) -#define after(seq2, seq1) before(seq1, seq2) - extern void cubictcp_init(struct sock *sk) __ksym; extern void cubictcp_cwnd_event(struct sock *sk, enum tcp_ca_event event) __ksym; extern __u32 cubictcp_recalc_ssthresh(struct sock *sk) __ksym; @@ -34,11 +30,6 @@ extern __u32 tcp_reno_undo_cwnd(struct sock *sk) __ksym; extern void cubictcp_acked(struct sock *sk, const struct ack_sample *sample) __ksym; extern void cubictcp_cong_avoid(struct sock *sk, __u32 ack, __u32 acked) __ksym; -static bool before(__u32 seq1, __u32 seq2) -{ - return (__s32)(seq1-seq2) < 0; -} - static __u64 div64_u64(__u64 dividend, __u64 divisor) { return dividend / divisor; diff --git a/tools/testing/selftests/bpf/progs/bpf_cubic.c b/tools/testing/selftests/bpf/progs/bpf_cubic.c index f089faa97ae6..46fb2b37d3a7 100644 --- a/tools/testing/selftests/bpf/progs/bpf_cubic.c +++ b/tools/testing/selftests/bpf/progs/bpf_cubic.c @@ -20,13 +20,6 @@ char _license[] SEC("license") = "GPL"; #define clamp(val, lo, hi) min((typeof(val))max(val, lo), hi) -#define min(a, b) ((a) < (b) ? (a) : (b)) -#define max(a, b) ((a) > (b) ? (a) : (b)) -static bool before(__u32 seq1, __u32 seq2) -{ - return (__s32)(seq1-seq2) < 0; -} -#define after(seq2, seq1) before(seq1, seq2) extern __u32 tcp_slow_start(struct tcp_sock *tp, __u32 acked) __ksym; extern void tcp_cong_avoid_ai(struct tcp_sock *tp, __u32 w, __u32 acked) __ksym; diff --git a/tools/testing/selftests/bpf/progs/bpf_dctcp.c b/tools/testing/selftests/bpf/progs/bpf_dctcp.c index 32c511bcd60b..1cc83140849f 100644 --- a/tools/testing/selftests/bpf/progs/bpf_dctcp.c +++ b/tools/testing/selftests/bpf/progs/bpf_dctcp.c @@ -13,16 +13,10 @@ #ifndef EBUSY #define EBUSY 16 #endif -#define min(a, b) ((a) < (b) ? (a) : (b)) -#define max(a, b) ((a) > (b) ? (a) : (b)) #define min_not_zero(x, y) ({ \ typeof(x) __x = (x); \ typeof(y) __y = (y); \ __x == 0 ? __y : ((__y == 0) ? __x : min(__x, __y)); }) -static bool before(__u32 seq1, __u32 seq2) -{ - return (__s32)(seq1-seq2) < 0; -} char _license[] SEC("license") = "GPL"; diff --git a/tools/testing/selftests/bpf/progs/bpf_tracing_net.h b/tools/testing/selftests/bpf/progs/bpf_tracing_net.h index 17db400f0e0d..39e98e16c113 100644 --- a/tools/testing/selftests/bpf/progs/bpf_tracing_net.h +++ b/tools/testing/selftests/bpf/progs/bpf_tracing_net.h @@ -5,6 +5,17 @@ #include <vmlinux.h> #include <bpf/bpf_core_read.h> +#define min(a, b) ((a) < (b) ? (a) : (b)) +#define max(a, b) ((a) > (b) ? (a) : (b)) + +static inline bool before(__u32 seq1, __u32 seq2) +{ + return (__s32)(seq1 - seq2) < 0; +} + +#define after(seq2, seq1) before(seq1, seq2) + + #define AF_INET 2 #define AF_INET6 10 diff --git a/tools/testing/selftests/bpf/progs/tcp_ca_write_sk_pacing.c b/tools/testing/selftests/bpf/progs/tcp_ca_write_sk_pacing.c index a58b5194fc89..022291f21dfb 100644 --- a/tools/testing/selftests/bpf/progs/tcp_ca_write_sk_pacing.c +++ b/tools/testing/selftests/bpf/progs/tcp_ca_write_sk_pacing.c @@ -8,8 +8,6 @@ char _license[] SEC("license") = "GPL"; #define USEC_PER_SEC 1000000UL -#define min(a, b) ((a) < (b) ? (a) : (b)) - static unsigned int tcp_left_out(const struct tcp_sock *tp) { return tp->sacked_out + tp->lost_out; -- 2.51.1

3 days, 5 hours

1
0
0 0

[bpf-next v1 2/5] selftests/bpf: use sockaddr_storage instead of sa46 in select_reuseport test

by Hoyeon Lee

The select_reuseport selftest uses a custom sa46 union to represent IPv4 and IPv6 addresses. This custom wrapper requires extra manual handling for address family and field extraction. Replace sa46 with sockaddr_storage and update the helper functions to operate on native socket structures. This simplifies the code and removes unnecessary custom address-handling logic. No functional changes intended. Signed-off-by: Hoyeon Lee <hoyeon.lee(a)suse.com> --- .../bpf/prog_tests/select_reuseport.c | 67 ++++++++++--------- 1 file changed, 34 insertions(+), 33 deletions(-) diff --git a/tools/testing/selftests/bpf/prog_tests/select_reuseport.c b/tools/testing/selftests/bpf/prog_tests/select_reuseport.c index 036d4760d2c1..3dbcc091f16c 100644 --- a/tools/testing/selftests/bpf/prog_tests/select_reuseport.c +++ b/tools/testing/selftests/bpf/prog_tests/select_reuseport.c @@ -41,11 +41,7 @@ static struct bpf_object *obj; static __u32 index_zero; static int epfd; -static union sa46 { - struct sockaddr_in6 v6; - struct sockaddr_in v4; - sa_family_t family; -} srv_sa; +static struct sockaddr_storage srv_sa; #define RET_IF(condition, tag, format...) ({ \ if (CHECK_FAIL(condition)) { \ @@ -135,24 +131,24 @@ static int prepare_bpf_obj(void) return 0; } -static void sa46_init_loopback(union sa46 *sa, sa_family_t family) +static void ss_init_loopback(struct sockaddr_storage *sa, sa_family_t family) { memset(sa, 0, sizeof(*sa)); - sa->family = family; - if (sa->family == AF_INET6) - sa->v6.sin6_addr = in6addr_loopback; + sa->ss_family = family; + if (sa->ss_family == AF_INET6) + ((struct sockaddr_in6 *)sa)->sin6_addr = in6addr_loopback; else - sa->v4.sin_addr.s_addr = htonl(INADDR_LOOPBACK); + ((struct sockaddr_in *)sa)->sin_addr.s_addr = htonl(INADDR_LOOPBACK); } -static void sa46_init_inany(union sa46 *sa, sa_family_t family) +static void ss_init_inany(struct sockaddr_storage *sa, sa_family_t family) { memset(sa, 0, sizeof(*sa)); - sa->family = family; - if (sa->family == AF_INET6) - sa->v6.sin6_addr = in6addr_any; + sa->ss_family = family; + if (sa->ss_family == AF_INET6) + ((struct sockaddr_in6 *)sa)->sin6_addr = in6addr_any; else - sa->v4.sin_addr.s_addr = INADDR_ANY; + ((struct sockaddr_in *)sa)->sin_addr.s_addr = INADDR_ANY; } static int read_int_sysctl(const char *sysctl) @@ -228,7 +224,7 @@ static void check_data(int type, sa_family_t family, const struct cmd *cmd, int cli_fd) { struct data_check expected = {}, result; - union sa46 cli_sa; + struct sockaddr_storage cli_sa; socklen_t addrlen; int err; @@ -251,26 +247,32 @@ static void check_data(int type, sa_family_t family, const struct cmd *cmd, } if (family == AF_INET6) { + struct sockaddr_in6 *srv_v6 = (struct sockaddr_in6 *)&srv_sa; + struct sockaddr_in6 *cli_v6 = (struct sockaddr_in6 *)&cli_sa; + expected.eth_protocol = htons(ETH_P_IPV6); - expected.bind_inany = !srv_sa.v6.sin6_addr.s6_addr32[3] && - !srv_sa.v6.sin6_addr.s6_addr32[2] && - !srv_sa.v6.sin6_addr.s6_addr32[1] && - !srv_sa.v6.sin6_addr.s6_addr32[0]; + expected.bind_inany = !srv_v6->sin6_addr.s6_addr32[3] && + !srv_v6->sin6_addr.s6_addr32[2] && + !srv_v6->sin6_addr.s6_addr32[1] && + !srv_v6->sin6_addr.s6_addr32[0]; - memcpy(&expected.skb_addrs[0], cli_sa.v6.sin6_addr.s6_addr32, - sizeof(cli_sa.v6.sin6_addr)); + memcpy(&expected.skb_addrs[0], cli_v6->sin6_addr.s6_addr32, + sizeof(cli_v6->sin6_addr)); memcpy(&expected.skb_addrs[4], &in6addr_loopback, sizeof(in6addr_loopback)); - expected.skb_ports[0] = cli_sa.v6.sin6_port; - expected.skb_ports[1] = srv_sa.v6.sin6_port; + expected.skb_ports[0] = cli_v6->sin6_port; + expected.skb_ports[1] = srv_v6->sin6_port; } else { + struct sockaddr_in *srv_v4 = (struct sockaddr_in *)&srv_sa; + struct sockaddr_in *cli_v4 = (struct sockaddr_in *)&cli_sa; + expected.eth_protocol = htons(ETH_P_IP); - expected.bind_inany = !srv_sa.v4.sin_addr.s_addr; + expected.bind_inany = !srv_v4->sin_addr.s_addr; - expected.skb_addrs[0] = cli_sa.v4.sin_addr.s_addr; + expected.skb_addrs[0] = cli_v4->sin_addr.s_addr; expected.skb_addrs[1] = htonl(INADDR_LOOPBACK); - expected.skb_ports[0] = cli_sa.v4.sin_port; - expected.skb_ports[1] = srv_sa.v4.sin_port; + expected.skb_ports[0] = cli_v4->sin_port; + expected.skb_ports[1] = srv_v4->sin_port; } if (memcmp(&result, &expected, offsetof(struct data_check, @@ -364,16 +366,15 @@ static void check_results(void) static int send_data(int type, sa_family_t family, void *data, size_t len, enum result expected) { - union sa46 cli_sa; + struct sockaddr_storage cli_sa; int fd, err; fd = socket(family, type, 0); RET_ERR(fd == -1, "socket()", "fd:%d errno:%d\n", fd, errno); - sa46_init_loopback(&cli_sa, family); + ss_init_loopback(&cli_sa, family); err = bind(fd, (struct sockaddr *)&cli_sa, sizeof(cli_sa)); RET_ERR(fd == -1, "bind(cli_sa)", "err:%d errno:%d\n", err, errno); - err = sendto(fd, data, len, MSG_FASTOPEN, (struct sockaddr *)&srv_sa, sizeof(srv_sa)); RET_ERR(err != len && expected >= PASS, @@ -589,9 +590,9 @@ static void prepare_sk_fds(int type, sa_family_t family, bool inany) socklen_t addrlen; if (inany) - sa46_init_inany(&srv_sa, family); + ss_init_inany(&srv_sa, family); else - sa46_init_loopback(&srv_sa, family); + ss_init_loopback(&srv_sa, family); addrlen = sizeof(srv_sa); /* -- 2.51.1

3 days, 5 hours

1
0
0 0

[PATCH v2 3/7] KVM: LoongArch: selftests: Add basic interfaces

by Bibo Mao

Add some basic function interfaces such as CSR register access, local irq enable or disable APIs. Signed-off-by: Bibo Mao <maobibo(a)loongson.cn> --- .../kvm/include/loongarch/processor.h | 52 +++++++++++++++++++ .../selftests/kvm/lib/loongarch/processor.c | 5 ++ 2 files changed, 57 insertions(+) diff --git a/tools/testing/selftests/kvm/include/loongarch/processor.h b/tools/testing/selftests/kvm/include/loongarch/processor.h index a18ac7bff303..b027f8f4dac7 100644 --- a/tools/testing/selftests/kvm/include/loongarch/processor.h +++ b/tools/testing/selftests/kvm/include/loongarch/processor.h @@ -118,6 +118,28 @@ #define CSR_TLBREHI_PS_SHIFT 0 #define CSR_TLBREHI_PS (0x3fUL << CSR_TLBREHI_PS_SHIFT) +#define csr_read(csr) \ +({ \ + register unsigned long __v; \ + __asm__ __volatile__( \ + "csrrd %[val], %[reg]\n\t" \ + : [val] "=r" (__v) \ + : [reg] "i" (csr) \ + : "memory"); \ + __v; \ +}) + +#define csr_write(v, csr) \ +({ \ + register unsigned long __v = v; \ + __asm__ __volatile__ ( \ + "csrwr %[val], %[reg]\n\t" \ + : [val] "+r" (__v) \ + : [reg] "i" (csr) \ + : "memory"); \ + __v; \ +}) + #define EXREGS_GPRS (32) #ifndef __ASSEMBLER__ @@ -147,6 +169,36 @@ struct handlers { void vm_init_descriptor_tables(struct kvm_vm *vm); void vm_install_exception_handler(struct kvm_vm *vm, int vector, handler_fn handler); +static inline void local_irq_enable(void) +{ + unsigned int flags = CSR_CRMD_IE; + + register unsigned int mask asm("$t0") = CSR_CRMD_IE; + + __asm__ __volatile__( + "csrxchg %[val], %[mask], %[reg]\n\t" + : [val] "+r" (flags) + : [mask] "r" (mask), [reg] "i" (LOONGARCH_CSR_CRMD) + : "memory"); +} + +static inline void local_irq_disable(void) +{ + unsigned int flags = 0; + + register unsigned int mask asm("$t0") = CSR_CRMD_IE; + + __asm__ __volatile__( + "csrxchg %[val], %[mask], %[reg]\n\t" + : [val] "+r" (flags) + : [mask] "r" (mask), [reg] "i" (LOONGARCH_CSR_CRMD) + : "memory"); +} + +static inline void cpu_relax(void) +{ + asm volatile("nop" ::: "memory"); +} #else #define PC_OFFSET_EXREGS ((EXREGS_GPRS + 0) * 8) #define ESTAT_OFFSET_EXREGS ((EXREGS_GPRS + 1) * 8) diff --git a/tools/testing/selftests/kvm/lib/loongarch/processor.c b/tools/testing/selftests/kvm/lib/loongarch/processor.c index be537c5ff74e..20ba476ccb72 100644 --- a/tools/testing/selftests/kvm/lib/loongarch/processor.c +++ b/tools/testing/selftests/kvm/lib/loongarch/processor.c @@ -373,3 +373,8 @@ void vcpu_arch_set_entry_point(struct kvm_vcpu *vcpu, void *guest_code) regs.pc = (uint64_t)guest_code; vcpu_regs_set(vcpu, &regs); } + +uint32_t guest_get_vcpuid(void) +{ + return csr_read(LOONGARCH_CSR_CPUID); +} -- 2.39.3

4 days

2
1
0 0

Re: [PATCH] selftests/cgroup: conform test to TAP format output

by Michal Koutný

On Fri, Nov 14, 2025 at 11:55:48AM +0800, Guopeng Zhang <zhangguopeng(a)kylinos.cn> wrote: > Actually, selftests are no longer just something for developers to view locally; they are now extensively > run in CI and stable branch regression testing. Using a standardized layout means that general test runners > and CI systems can parse the cgroup test results without any special handling. Nice. I appreciate you took this up. > This patch is not part of a formal, tree-wide conversion series I am running; it is an incremental step to align the > cgroup C tests with the existing TAP usage. I started here because these tests already use ksft_test_result_*() and > only require minor changes to generate proper TAP output. The tests are in various state of usage, correctness and usefulness, hence... > > > I'm asking to better asses whether also the scripts listed in > > Makefile:TEST_PROGS should be converted too. > > I agree that having them produce TAP output would benefit tooling and CI. I did not want to mix > that into this change, but if you and other maintainers think this direction is reasonable, > I would be happy to follow up and convert the cgroup shell tests to TAP as well. ...I'd suggest next focus on test_cpuset_prs.sh (as discussed, it may need more changes to adapt its output too). Michal

4 days

2
1
0 0

[PATCH bpf-next] selftests/bpf: simplify the kernel_count bench trigger

by Menglong Dong

Remove the "trigger_count" in trigger_bench.c and reuse trigger_driver() instead for trigger_kernel_count_setup(). With the calling to bpf_get_numa_node_id(), the result for "kernel_count" will become a little more accurate. It will also easier if we want to test the performance of livepatch, just hook the bpf_get_numa_node_id() and run the "kernel_count" bench trigger. Signed-off-by: Menglong Dong <dongml2(a)chinatelecom.cn> --- .../selftests/bpf/benchs/bench_trigger.c | 5 +---- .../testing/selftests/bpf/progs/trigger_bench.c | 17 +++++------------ 2 files changed, 6 insertions(+), 16 deletions(-) diff --git a/tools/testing/selftests/bpf/benchs/bench_trigger.c b/tools/testing/selftests/bpf/benchs/bench_trigger.c index 1e2aff007c2a..34fd8fa3b803 100644 --- a/tools/testing/selftests/bpf/benchs/bench_trigger.c +++ b/tools/testing/selftests/bpf/benchs/bench_trigger.c @@ -179,11 +179,8 @@ static void trigger_syscall_count_setup(void) static void trigger_kernel_count_setup(void) { setup_ctx(); - bpf_program__set_autoload(ctx.skel->progs.trigger_driver, false); - bpf_program__set_autoload(ctx.skel->progs.trigger_count, true); + ctx.skel->rodata->kernel_count = 1; load_ctx(); - /* override driver program */ - ctx.driver_prog_fd = bpf_program__fd(ctx.skel->progs.trigger_count); } static void trigger_kprobe_setup(void) diff --git a/tools/testing/selftests/bpf/progs/trigger_bench.c b/tools/testing/selftests/bpf/progs/trigger_bench.c index 3d5f30c29ae3..6564d1909c7b 100644 --- a/tools/testing/selftests/bpf/progs/trigger_bench.c +++ b/tools/testing/selftests/bpf/progs/trigger_bench.c @@ -39,26 +39,19 @@ int bench_trigger_uprobe_multi(void *ctx) return 0; } +const volatile int kernel_count = 0; const volatile int batch_iters = 0; -SEC("?raw_tp") -int trigger_count(void *ctx) -{ - int i; - - for (i = 0; i < batch_iters; i++) - inc_counter(); - - return 0; -} - SEC("?raw_tp") int trigger_driver(void *ctx) { int i; - for (i = 0; i < batch_iters; i++) + for (i = 0; i < batch_iters; i++) { (void)bpf_get_numa_node_id(); /* attach point for benchmarking */ + if (kernel_count) + inc_counter(); + } return 0; } -- 2.51.2

4 days, 1 hour

2
2
0 0

[PATCH net-next] selftests: drv-net: xdp: make the XDP qstats tests less flaky

by Jakub Kicinski

The XDP qstats tests send 2k packets over a single socket. Looks like when netdev CI is busy running those tests in QEMU occasionally flakes. The target doesn't get to run at all before all 2000 packets are sent. Lower the number of packets to 1000 and reopen the socket every 50 packets, to give RSS a chance to spread the packets to multiple queues. For the netdev CI testing either lowering the count or using multiple sockets is enough, but let's do both for extra resiliency. Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> --- CC: shuah(a)kernel.org CC: ast(a)kernel.org CC: hawk(a)kernel.org CC: john.fastabend(a)gmail.com CC: sdf(a)fomichev.me CC: linux-kselftest(a)vger.kernel.org --- tools/testing/selftests/drivers/net/xdp.py | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) diff --git a/tools/testing/selftests/drivers/net/xdp.py b/tools/testing/selftests/drivers/net/xdp.py index a148004e1c36..834a37ae7d0d 100755 --- a/tools/testing/selftests/drivers/net/xdp.py +++ b/tools/testing/selftests/drivers/net/xdp.py @@ -687,9 +687,12 @@ from lib.py import ip, bpftool, defer "/dev/null" # Listener runs on "remote" in case of XDP_TX rx_host = cfg.remote if act == XDPAction.TX else None - # We want to spew 2000 packets quickly, bash seems to do a good enough job - tx_udp = f"exec 5<>/dev/udp/{cfg.addr}/{port}; " \ - "for i in `seq 2000`; do echo a >&5; done; exec 5>&-" + # We want to spew 1000 packets quickly, bash seems to do a good enough job + # Each reopening of the socket gives us a differenot local port (for RSS) + tx_udp = "for _ in `seq 20`; do " \ + f"exec 5<>/dev/udp/{cfg.addr}/{port}; " \ + "for i in `seq 50`; do echo a >&5; done; " \ + "exec 5>&-; done" cfg.wait_hw_stats_settle() # Qstats have more clearly defined semantics than rtnetlink. @@ -704,11 +707,11 @@ from lib.py import ip, bpftool, defer cfg.wait_hw_stats_settle() after = cfg.netnl.qstats_get({"ifindex": cfg.ifindex}, dump=True)[0] - ksft_ge(after['rx-packets'] - before['rx-packets'], 2000) + expected_pkts = 1000 + ksft_ge(after['rx-packets'] - before['rx-packets'], expected_pkts) if act == XDPAction.TX: - ksft_ge(after['tx-packets'] - before['tx-packets'], 2000) + ksft_ge(after['tx-packets'] - before['tx-packets'], expected_pkts) - expected_pkts = 2000 stats = _get_stats(prog_info["maps"]["map_xdp_stats"]) ksft_eq(stats[XDPStats.RX.value], expected_pkts, "XDP RX stats mismatch") if act == XDPAction.TX: -- 2.51.1

4 days, 2 hours

3
2
0 0

[PATCH net-next] selftests: drv-net: xdp: Fix register spill error with clang 20

by Dimitri Daskalakis

On clang 20.1.8 the XDP program fails to load with a register spill error. Since hdr_len is a __u32, the compiler decided it only needed the lower 32-bits of ctx->data, which later triggers the register spill verifier error. Suggested-by: Martin KaFai Lau <martin.lau(a)kernel.org> Signed-off-by: Dimitri Daskalakis <dimitri.daskalakis1(a)gmail.com> Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> --- libbpf: prog 'xdp_prog': BPF program load failed: Permission denied libbpf: prog 'xdp_prog': -- BEGIN PROG LOAD LOG -- 0: R1=ctx() R10=fp0 ; return xdp_prog_common(ctx); @ xdp_native.bpf.c:670 0: (85) call pc+1 caller: R10=fp0 callee: frame1: R1=ctx() R10=fp0 2: frame1: R1=ctx() R10=fp0 ; static int xdp_prog_common(struct xdp_md *ctx) @ xdp_native.bpf.c:635 2: (bf) r7 = r1 ; frame1: R1=ctx() R7_w=ctx() 3: (b4) w1 = 0 ; frame1: R1_w=0 ; key = XDP_MODE; @ xdp_native.bpf.c:640 4: (63) *(u32 *)(r10 -336) = r1 ; frame1: R1_w=0 R10=fp0 fp-336=????0 5: (bf) r2 = r10 ; frame1: R2_w=fp0 R10=fp0 6: (07) r2 += -336 ; frame1: R2_w=fp-336 ; mode = bpf_map_lookup_elem(&map_xdp_setup, &key); @ xdp_native.bpf.c:641 7: (18) r1 = 0xff110001099ada00 ; frame1: R1_w=map_ptr(map=map_xdp_setup,ks=4,vs=4) 9: (85) call bpf_map_lookup_elem#1 ; frame1: R0=map_value(map=map_xdp_setup,ks=4,vs=4) 10: (bf) r8 = r0 ; frame1: R0=map_value(map=map_xdp_setup,ks=4,vs=4) R8_w=map_value(map=map_xdp_setup,ks=4,vs=4) 11: (b4) w6 = 2 ; frame1: R6_w=2 ; if (!mode) @ xdp_native.bpf.c:642 12: (15) if r8 == 0x0 goto pc+669 ; frame1: R8_w=map_value(map=map_xdp_setup,ks=4,vs=4) 13: (b4) w1 = 1 ; frame1: R1_w=1 ; key = XDP_PORT; @ xdp_native.bpf.c:645 14: (63) *(u32 *)(r10 -336) = r1 ; frame1: R1_w=1 R10=fp0 fp-336=????1 15: (bf) r2 = r10 ; frame1: R2_w=fp0 R10=fp0 16: (07) r2 += -336 ; frame1: R2_w=fp-336 ; port = bpf_map_lookup_elem(&map_xdp_setup, &key); @ xdp_native.bpf.c:646 17: (18) r1 = 0xff110001099ada00 ; frame1: R1_w=map_ptr(map=map_xdp_setup,ks=4,vs=4) 19: (85) call bpf_map_lookup_elem#1 ; frame1: R0=map_value(map=map_xdp_setup,ks=4,vs=4) ; if (!port) @ xdp_native.bpf.c:647 20: (15) if r0 == 0x0 goto pc+661 ; frame1: R0=map_value(map=map_xdp_setup,ks=4,vs=4) ; switch (*mode) { @ xdp_native.bpf.c:650 21: (61) r1 = *(u32 *)(r8 +0) ; frame1: R1_w=scalar(smin=0,smax=umax=0xffffffff,var_off=(0x0; 0xffffffff)) R8=map_value(map=map_xdp_setup,ks=4,vs=4) 22: (66) if w1 s> 0x1 goto pc+20 43: frame1: R0=map_value(map=map_xdp_setup,ks=4,vs=4) R1=scalar(smin=umin=smin32=umin32=2,smax=umax=umax32=0x7fffffff,var_off=(0x0; 0x7fffffff)) R6=2 R7=ctx() R8=map_value(map=map_xdp_setup,ks=4,vs=4) R10=fp0 fp-336=????1 ; switch (*mode) { @ xdp_native.bpf.c:650 43: (16) if w1 == 0x2 goto pc+26 ; frame1: R1=scalar(smin=umin=smin32=umin32=3,smax=umax=umax32=0x7fffffff,var_off=(0x0; 0x7fffffff)) 44: (16) if w1 == 0x3 goto pc+123 168: frame1: R0=map_value(map=map_xdp_setup,ks=4,vs=4) R1=3 R6=2 R7=ctx() R8=map_value(map=map_xdp_setup,ks=4,vs=4) R10=fp0 fp-336=????1 ; return xdp_adjst_tail(ctx, (__u16)(*port)); @ xdp_native.bpf.c:658 168: (61) r2 = *(u32 *)(r0 +0) ; frame1: R0=map_value(map=map_xdp_setup,ks=4,vs=4) R2_w=scalar(smin=0,smax=umax=0xffffffff,var_off=(0x0; 0xffffffff)) ; udph = filter_udphdr(ctx, port); @ xdp_native.bpf.c:430 169: (54) w2 &= 65535 ; frame1: R2_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=0xffff,var_off=(0x0; 0xffff)) 170: (bf) r1 = r7 ; frame1: R1_w=ctx() R7=ctx() 171: (85) call pc+512 caller: frame1: R6=2 R7=ctx() R8=map_value(map=map_xdp_setup,ks=4,vs=4) R10=fp0 fp-336=????1 callee: frame2: R1_w=ctx() R2_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=0xffff,var_off=(0x0; 0xffff)) R10=fp0 684: frame2: R1=ctx() R2=scalar(smin=smin32=0,smax=umax=smax32=umax32=0xffff,var_off=(0x0; 0xffff)) R10=fp0 ; static struct udphdr *filter_udphdr(struct xdp_md *ctx, __u16 port) @ xdp_native.bpf.c:71 684: (bc) w6 = w2 ; frame2: R2=scalar(id=54,smin=smin32=0,smax=umax=smax32=umax32=0xffff,var_off=(0x0; 0xffff)) R6_w=scalar(id=54,smin=smin32=0,smax=umax=smax32=umax32=0xffff,var_off=(0x0; 0xffff)) 685: (bf) r7 = r1 ; frame2: R1=ctx() R7_w=ctx() ; err = bpf_xdp_pull_data(ctx, sizeof(*eth)); @ xdp_native.bpf.c:78 686: (b4) w2 = 14 ; frame2: R2_w=14 687: (85) call bpf_xdp_pull_data#85514 ; frame2: R0_w=scalar() 688: (bc) w1 = w0 ; frame2: R0_w=scalar() R1_w=scalar(smin=0,smax=umax=0xffffffff,var_off=(0x0; 0xffffffff)) 689: (b7) r0 = 0 ; frame2: R0_w=0 ; if (err) @ xdp_native.bpf.c:79 690: (56) if w1 != 0x0 goto pc+55 ; frame2: R1_w=0 ; data_end = (void *)(long)ctx->data_end; @ xdp_native.bpf.c:82 691: (61) r2 = *(u32 *)(r7 +4) ; frame2: R2_w=pkt_end() R7_w=ctx() ; data = eth = (void *)(long)ctx->data; @ xdp_native.bpf.c:83 692: (61) r1 = *(u32 *)(r7 +0) ; frame2: R1_w=pkt(r=0) R7_w=ctx() ; if (data + sizeof(*eth) > data_end) @ xdp_native.bpf.c:85 693: (bf) r3 = r1 ; frame2: R1_w=pkt(r=0) R3_w=pkt(r=0) 694: (07) r3 += 14 ; frame2: R3=pkt(off=14,r=0) 695: (2d) if r3 > r2 goto pc+50 ; frame2: R2=pkt_end() R3=pkt(off=14,r=14) ; if (eth->h_proto == bpf_htons(ETH_P_IP)) { @ xdp_native.bpf.c:88 696: (71) r2 = *(u8 *)(r1 +12) ; frame2: R1=pkt(r=14) R2_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=255,var_off=(0x0; 0xff)) 697: (71) r1 = *(u8 *)(r1 +13) ; frame2: R1_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=255,var_off=(0x0; 0xff)) 698: (64) w1 <<= 8 ; frame2: R1_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=0xff00,var_off=(0x0; 0xff00)) 699: (4c) w1 |= w2 ; frame2: R1_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=0xffff,var_off=(0x0; 0xffff)) R2_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=255,var_off=(0x0; 0xff)) 700: (16) if w1 == 0xdd86 goto pc+15 ; frame2: R1_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=0xffff,var_off=(0x0; 0xffff)) 701: (56) if w1 != 0x8 goto pc+44 ; frame2: R1_w=8 ; err = bpf_xdp_pull_data(ctx, sizeof(*eth) + sizeof(*iph) + @ xdp_native.bpf.c:91 702: (bf) r1 = r7 ; frame2: R1_w=ctx() R7=ctx() 703: (b4) w2 = 42 ; frame2: R2_w=42 704: (85) call bpf_xdp_pull_data#85514 ; frame2: R0=scalar() 705: (bc) w1 = w0 ; frame2: R0=scalar() R1_w=scalar(smin=0,smax=umax=0xffffffff,var_off=(0x0; 0xffffffff)) 706: (b7) r0 = 0 ; frame2: R0_w=0 ; if (err) @ xdp_native.bpf.c:93 707: (56) if w1 != 0x0 goto pc+38 ; frame2: R1_w=0 ; data_end = (void *)(long)ctx->data_end; @ xdp_native.bpf.c:96 708: (61) r1 = *(u32 *)(r7 +4) ; frame2: R1_w=pkt_end() R7=ctx() ; data = (void *)(long)ctx->data; @ xdp_native.bpf.c:97 709: (61) r2 = *(u32 *)(r7 +0) ; frame2: R2_w=pkt(r=0) R7=ctx() ; if (iph + 1 > (struct iphdr *)data_end || @ xdp_native.bpf.c:101 710: (bf) r7 = r2 ; frame2: R2_w=pkt(r=0) R7_w=pkt(r=0) 711: (07) r7 += 34 ; frame2: R7_w=pkt(off=34,r=0) 712: (2d) if r7 > r1 goto pc+33 ; frame2: R1_w=pkt_end() R7_w=pkt(off=34,r=34) ; iph->protocol != IPPROTO_UDP) @ xdp_native.bpf.c:102 713: (71) r2 = *(u8 *)(r2 +23) ; frame2: R2=scalar(smin=smin32=0,smax=umax=smax32=umax32=255,var_off=(0x0; 0xff)) ; if (iph + 1 > (struct iphdr *)data_end || @ xdp_native.bpf.c:101 714: (16) if w2 == 0x11 goto pc+14 729: frame2: R0=0 R1=pkt_end() R2=17 R6=scalar(id=54,smin=smin32=0,smax=umax=smax32=umax32=0xffff,var_off=(0x0; 0xffff)) R7=pkt(off=34,r=34) R10=fp0 ; if (udph + 1 > (struct udphdr *)data_end) @ xdp_native.bpf.c:128 729: (bf) r2 = r7 ; frame2: R2_w=pkt(off=34,r=34) R7=pkt(off=34,r=34) 730: (07) r2 += 8 ; frame2: R2=pkt(off=42,r=34) 731: (2d) if r2 > r1 goto pc+14 ; frame2: R1=pkt_end() R2=pkt(off=42,r=42) ; if (udph->dest != bpf_htons(port)) @ xdp_native.bpf.c:131 732: (dc) r6 = be16 r6 ; frame2: R6_w=scalar() 733: (69) r1 = *(u16 *)(r7 +2) ; frame2: R1_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=0xffff,var_off=(0x0; 0xffff)) R7=pkt(off=34,r=42) 734: (5e) if w1 != w6 goto pc+11 ; frame2: R1_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=0xffff,var_off=(0x0; 0xffff)) R6_w=scalar(smax=0x7fffffff0000ffff,umax=0xffffffff0000ffff,smin32=0,smax32=umax32=0xffff,var_off=(0x0; 0xffffffff0000ffff)) 735: (b4) w1 = 0 ; frame2: R1_w=0 736: (63) *(u32 *)(r10 -4) = r1 ; frame2: R1_w=0 R10=fp0 fp-8=0000???? 737: (bf) r2 = r10 ; frame2: R2_w=fp0 R10=fp0 738: (07) r2 += -4 ; frame2: R2_w=fp-4 ; count = bpf_map_lookup_elem(&map_xdp_stats, &stat_type); @ xdp_native.bpf.c:65 739: (18) r1 = 0xff110001099ad200 ; frame2: R1_w=map_ptr(map=map_xdp_stats,ks=4,vs=8) 741: (85) call bpf_map_lookup_elem#1 ; frame2: R0=map_value(map=map_xdp_stats,ks=4,vs=8) ; if (count) @ xdp_native.bpf.c:67 742: (15) if r0 == 0x0 goto pc+2 ; frame2: R0=map_value(map=map_xdp_stats,ks=4,vs=8) 743: (b7) r1 = 1 ; frame2: R1_w=1 ; __sync_fetch_and_add(count, 1); @ xdp_native.bpf.c:68 744: (db) r1 = atomic64_fetch_add((u64 *)(r0 +0), r1) ; frame2: R0=map_value(map=map_xdp_stats,ks=4,vs=8) R1_w=scalar() 745: (bf) r0 = r7 ; frame2: R0_w=pkt(off=34,r=42) R7=pkt(off=34,r=42) ; } @ xdp_native.bpf.c:137 746: (95) exit returning from callee: frame2: R0_w=pkt(off=34,r=42) R1_w=scalar() R6=scalar(smax=0x7fffffff0000ffff,umax=0xffffffff0000ffff,smin32=0,smax32=umax32=0xffff,var_off=(0x0; 0xffffffff0000ffff)) R7=pkt(off=34,r=42) R10=fp0 fp-8=0000???? to caller at 172: frame1: R0_w=pkt(off=34,r=42) R6=2 R7=ctx() R8=map_value(map=map_xdp_setup,ks=4,vs=4) R10=fp0 fp-336=????1 ; udph = filter_udphdr(ctx, port); @ xdp_native.bpf.c:430 172: (bf) r8 = r0 ; frame1: R0_w=pkt(off=34,r=42) R8_w=pkt(off=34,r=42) ; if (!udph) @ xdp_native.bpf.c:431 173: (15) if r8 == 0x0 goto pc+508 ; frame1: R8_w=pkt(off=34,r=42) ; hdr_len = (void *)udph - (void *)(long)ctx->data + @ xdp_native.bpf.c:434 174: (61) r9 = *(u32 *)(r7 +0) ; frame1: R7=ctx() R9_w=pkt(r=0) ; key = XDP_ADJST_OFFSET; @ xdp_native.bpf.c:436 175: (63) *(u32 *)(r10 -328) = r6 ; frame1: R6=2 R10=fp0 fp-328=????2 176: (bf) r2 = r10 ; frame1: R2_w=fp0 R10=fp0 177: (07) r2 += -328 ; frame1: R2_w=fp-328 ; adjust_offset = bpf_map_lookup_elem(&map_xdp_setup, &key); @ xdp_native.bpf.c:437 178: (18) r1 = 0xff110001099ada00 ; frame1: R1_w=map_ptr(map=map_xdp_setup,ks=4,vs=4) 180: (85) call bpf_map_lookup_elem#1 ; frame1: R0=map_value(map=map_xdp_setup,ks=4,vs=4) ; if (!adjust_offset) @ xdp_native.bpf.c:438 181: (15) if r0 == 0x0 goto pc+500 ; frame1: R0=map_value(map=map_xdp_setup,ks=4,vs=4) 182: (63) *(u32 *)(r10 -344) = r9 invalid size of register spill processed 2736 insns (limit 1000000) max_states_per_insn 5 total_states 190 peak_states 156 mark_read 19 -- END PROG LOAD LOG -- libbpf: prog 'xdp_prog': failed to load: -13 libbpf: failed to load object '/root/ksft-net-drv/net/lib/xdp_native.bpf.o' tools/testing/selftests/net/lib/xdp_native.bpf.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/net/lib/xdp_native.bpf.c b/tools/testing/selftests/net/lib/xdp_native.bpf.c index c368fc045f4b..64f05229ab24 100644 --- a/tools/testing/selftests/net/lib/xdp_native.bpf.c +++ b/tools/testing/selftests/net/lib/xdp_native.bpf.c @@ -332,7 +332,7 @@ static __u16 csum_fold_helper(__u32 csum) } static int xdp_adjst_tail_shrnk_data(struct xdp_md *ctx, __u16 offset, - __u32 hdr_len) + unsigned long hdr_len) { char tmp_buff[MAX_ADJST_OFFSET]; __u32 buff_pos, udp_csum = 0; @@ -422,8 +422,9 @@ static int xdp_adjst_tail(struct xdp_md *ctx, __u16 port) { struct udphdr *udph = NULL; __s32 *adjust_offset, *val; - __u32 key, hdr_len; + unsigned long hdr_len; void *offset_ptr; + __u32 key; __u8 tag; int ret; -- 2.47.3

4 days, 2 hours

2
1
0 0

[PATCH v4 0/2] libbpf: fix BTF dedup to support recursive typedef

by Paul Houssel

Pahole fails to encode BTF for some Go projects (e.g. Kubernetes and Podman) due to recursive type definitions that create reference loops not representable in C. These recursive typedefs trigger a failure in the BTF deduplication algorithm. This patch extends btf_dedup_struct_types() to properly handle potential recursion for BTF_KIND_TYPEDEF, similar to how recursion is already handled for BTF_KIND_STRUCT. This allows pahole to successfully generate BTF for Go binaries using recursive types without impacting existing C-based workflows. Changes in v4: fix typo found by Claude-based CI Changes in v3: 1. Patch 1: Adjusted the comment of btf_dedup_ref_type() to refer to typedef as well. 2. Patch 2: Update of the "dedup: recursive typedef" test to include a duplicated version of the types to make sure deduplication still happens in this case. Changes in v2: 1. Patch 1: Refactored code to prevent copying existing logic. Instead of adding a new function we modify the existing btf_dedup_struct_type() function to handle the BTF_KIND_TYPEDEF case. Calls to btf_hash_struct() and btf_shallow_equal_struct() are replaced with calls to functions that select btf_hash_struct() / btf_hash_typedef() based on the type. 2. Patch 2: Added tests v3: https://lore.kernel.org/lkml/cover.1763024337.git.paul.houssel@orange.com/ v2: https://lore.kernel.org/lkml/cover.1762956564.git.paul.houssel@orange.com/ v1: https://lore.kernel.org/lkml/20251107153408.159342-1-paulhoussel2@gmail.com/ Paul Houssel (2): libbpf: fix BTF dedup to support recursive typedef definitions selftests/bpf: add BTF dedup tests for recursive typedef definitions tools/lib/bpf/btf.c | 71 +++++++++++++++----- tools/testing/selftests/bpf/prog_tests/btf.c | 65 ++++++++++++++++++ 2 files changed, 120 insertions(+), 16 deletions(-) -- 2.51.0

4 days, 3 hours

3
5
0 0

[RFC PATCH v1 00/37] guest_memfd: In-place conversion support

by Ackerley Tng

Hello, IIUC this is the first independent patch series for guest_memfd's in-place conversion series! Happy to finally bring this out on its own. Previous versions of this feature, part of other series, are available at [1][2][3]. Many prior discussions have led up to these main features of this series, and these are the main points I'd like feedback on. 1. Having private/shared status stored in a maple tree (Thanks Michael for your support of using maple trees over xarrays for performance! [4]). 2. Having a new guest_memfd ioctl (not a vm ioctl) that performs conversions. 3. Using ioctls/structs/input attribute similar to the existing vm ioctl KVM_SET_MEMORY_ATTRIBUTES to perform conversions. 4. Storing requested attributes directly in the maple tree. 5. Using a KVM module-wide param to toggle between setting memory attributes via vm and guest_memfd ioctls (making them mututally exclusive - a single loaded KVM module can only do one of the two.) 6. Skipping LRU in guest_memfd folios - make guest_memfd folios not participate in LRU to avoid LRU refcounts from interfering with conversions. This series is based on kvm/next, followed by + v12 of NUMA mempolicy support patches [5] + 3 cleanup patches from Sean [6][7][8] Everything is stitched together here for your convenience https://github.com/googleprodkernel/linux-cc/commits/guest_memfd-inplace-co… Thank you all for helping with this series! If I missed out your comment from a previous series, it's not intentional! Please do raise it again. TODOs: + There might be an issue with memory failure handling because when guest_memfd folios stop participating in LRU. From a preliminary analysis, HWPoisonHandlable() is only true if PageLRU() is true. This needs further investigation. [1] https://lore.kernel.org/all/bd163de3118b626d1005aa88e71ef2fb72f0be0f.172600… [2] https://lore.kernel.org/all/20250117163001.2326672-6-tabba@google.com/ [3] https://lore.kernel.org/all/b784326e9ccae6a08388f1bf39db70a2204bdc51.174726… [4] https://lore.kernel.org/all/20250529054227.hh2f4jmyqf6igd3i@amd.com/ [5] https://lore.kernel.org/all/20251007221420.344669-1-seanjc@google.com/T/ [6] https://lore.kernel.org/all/20250924174255.2141847-1-seanjc@google.com/ [7] https://lore.kernel.org/all/20251007224515.374516-1-seanjc@google.com/ [8] https://lore.kernel.org/all/20251007223625.369939-1-seanjc@google.com/ Ackerley Tng (19): KVM: guest_memfd: Update kvm_gmem_populate() to use gmem attributes KVM: Introduce KVM_SET_MEMORY_ATTRIBUTES2 KVM: guest_memfd: Don't set FGP_ACCESSED when getting folios KVM: guest_memfd: Skip LRU for guest_memfd folios KVM: guest_memfd: Add support for KVM_SET_MEMORY_ATTRIBUTES KVM: selftests: Update framework to use KVM_SET_MEMORY_ATTRIBUTES2 KVM: selftests: guest_memfd: Test basic single-page conversion flow KVM: selftests: guest_memfd: Test conversion flow when INIT_SHARED KVM: selftests: guest_memfd: Test indexing in guest_memfd KVM: selftests: guest_memfd: Test conversion before allocation KVM: selftests: guest_memfd: Convert with allocated folios in different layouts KVM: selftests: guest_memfd: Test precision of conversion KVM: selftests: guest_memfd: Test that truncation does not change shared/private status KVM: selftests: guest_memfd: Test conversion with elevated page refcount KVM: selftests: Reset shared memory after hole-punching KVM: selftests: Provide function to look up guest_memfd details from gpa KVM: selftests: Make TEST_EXPECT_SIGBUS thread-safe KVM: selftests: Update private_mem_conversions_test to mmap() guest_memfd KVM: selftests: Add script to exercise private_mem_conversions_test Sean Christopherson (18): KVM: guest_memfd: Introduce per-gmem attributes, use to guard user mappings KVM: Rename KVM_GENERIC_MEMORY_ATTRIBUTES to KVM_VM_MEMORY_ATTRIBUTES KVM: Enumerate support for PRIVATE memory iff kvm_arch_has_private_mem is defined KVM: Stub in ability to disable per-VM memory attribute tracking KVM: guest_memfd: Wire up kvm_get_memory_attributes() to per-gmem attributes KVM: guest_memfd: Enable INIT_SHARED on guest_memfd for x86 Coco VMs KVM: Move KVM_VM_MEMORY_ATTRIBUTES config definition to x86 KVM: Let userspace disable per-VM mem attributes, enable per-gmem attributes KVM: selftests: Create gmem fd before "regular" fd when adding memslot KVM: selftests: Rename guest_memfd{,_offset} to gmem_{fd,offset} KVM: selftests: Add support for mmap() on guest_memfd in core library KVM: selftests: Add helpers for calling ioctls on guest_memfd KVM: selftests: guest_memfd: Test that shared/private status is consistent across processes KVM: selftests: Add selftests global for guest memory attributes capability KVM: selftests: Provide common function to set memory attributes KVM: selftests: Check fd/flags provided to mmap() when setting up memslot KVM: selftests: Update pre-fault test to work with per-guest_memfd attributes KVM: selftests: Update private memory exits test work with per-gmem attributes Documentation/virt/kvm/api.rst | 72 ++- arch/x86/include/asm/kvm_host.h | 2 +- arch/x86/kvm/Kconfig | 15 +- arch/x86/kvm/mmu/mmu.c | 4 +- arch/x86/kvm/x86.c | 13 +- include/linux/kvm_host.h | 44 +- include/trace/events/kvm.h | 4 +- include/uapi/linux/kvm.h | 17 + mm/filemap.c | 1 + mm/memcontrol.c | 2 + tools/testing/selftests/kvm/.gitignore | 1 + tools/testing/selftests/kvm/Makefile.kvm | 1 + .../kvm/guest_memfd_conversions_test.c | 498 ++++++++++++++++++ .../testing/selftests/kvm/include/kvm_util.h | 127 ++++- .../testing/selftests/kvm/include/test_util.h | 29 +- tools/testing/selftests/kvm/lib/kvm_util.c | 128 +++-- tools/testing/selftests/kvm/lib/test_util.c | 7 - .../selftests/kvm/pre_fault_memory_test.c | 2 +- .../kvm/x86/private_mem_conversions_test.c | 55 +- .../kvm/x86/private_mem_conversions_test.py | 159 ++++++ .../kvm/x86/private_mem_kvm_exits_test.c | 36 +- virt/kvm/Kconfig | 4 +- virt/kvm/guest_memfd.c | 414 +++++++++++++-- virt/kvm/kvm_main.c | 104 +++- 24 files changed, 1554 insertions(+), 185 deletions(-) create mode 100644 tools/testing/selftests/kvm/guest_memfd_conversions_test.c create mode 100755 tools/testing/selftests/kvm/x86/private_mem_conversions_test.py -- 2.51.0.858.gf9c4a03a3a-goog

4 days, 3 hours

7
62
0 0

[PATCH] KVM: selftests: Add SYNC after guest ITS setup in vgic_lpi_stress

by Maximilian Dittgen

vgic_lpi_stress sends MAPTI and MAPC commands during guest GIC setup to map interrupt events to ITT entries and collection IDs to redistributors, respectively. Theoretically, we have no guarantee that the ITS will finish handling these mapping commands before the selftest calls KVM_SIGNAL_MSI to inject LPIs to the guest. If LPIs are injected before ITS mapping completes, the ITS cannot properly pass the interrupt on to the redistributor. In practice, KVM processes ITS commands synchronously, so SYNC calls are functionally unnecessary and ignored in vgic_its_handle_command(). However, selftests should test based on ARM specification and be blind to KVM-specific implementation optimizations. Thus, we must update the test to be architecturally compliant and logically correct. Fix by adding a SYNC command to the selftests ITS library, then calling SYNC after ITS mapping to ensure mapping completes before signal_lpi() writes to GITS_TRANSLATER. This patch depends on commit a24f7afce048 ("KVM: selftests: fix MAPC RDbase target formatting in vgic_lpi_stress"), which is queued in kvmarm/fixes. Signed-off-by: Maximilian Dittgen <mdittgen(a)amazon.de> --- Validated by the following debug logging to the GITS_CMD_SYNC handler in vgic_its_handle_command(): kvm_info("ITS SYNC command: %016llx %016llx %016llx %016llx\n", its_cmd[0], its_cmd[1], its_cmd[2], its_cmd[3]); Initialized a selftest guest with 4 vCPUs by: ./vgic_lpi_stress -v 4 Confirmed that an ITS SYNC was successfully called for all 4 vCPUs: kvm [5094]: ITS SYNC command: 0000000000000005 0000000000000000 0000000000000000 0000000000000000 kvm [5094]: ITS SYNC command: 0000000000000005 0000000000000000 0000000000010000 0000000000000000 kvm [5094]: ITS SYNC command: 0000000000000005 0000000000000000 0000000000020000 0000000000000000 kvm [5094]: ITS SYNC command: 0000000000000005 0000000000000000 0000000000030000 0000000000000000 --- tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c | 4 ++++ .../testing/selftests/kvm/include/arm64/gic_v3_its.h | 1 + tools/testing/selftests/kvm/lib/arm64/gic_v3_its.c | 11 +++++++++++ 3 files changed, 16 insertions(+) diff --git a/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c b/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c index 687d04463983..e857a605f577 100644 --- a/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c +++ b/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c @@ -118,6 +118,10 @@ static void guest_setup_gic(void) guest_setup_its_mappings(); guest_invalidate_all_rdists(); + + /* SYNC to ensure ITS setup is complete */ + for (cpuid = 0; cpuid < test_data.nr_cpus; cpuid++) + its_send_sync_cmd(test_data.cmdq_base_va, cpuid); } static void guest_code(size_t nr_lpis) diff --git a/tools/testing/selftests/kvm/include/arm64/gic_v3_its.h b/tools/testing/selftests/kvm/include/arm64/gic_v3_its.h index 3722ed9c8f96..58feef3eb386 100644 --- a/tools/testing/selftests/kvm/include/arm64/gic_v3_its.h +++ b/tools/testing/selftests/kvm/include/arm64/gic_v3_its.h @@ -15,5 +15,6 @@ void its_send_mapc_cmd(void *cmdq_base, u32 vcpu_id, u32 collection_id, bool val void its_send_mapti_cmd(void *cmdq_base, u32 device_id, u32 event_id, u32 collection_id, u32 intid); void its_send_invall_cmd(void *cmdq_base, u32 collection_id); +void its_send_sync_cmd(void *cmdq_base, u32 vcpu_id); #endif // __SELFTESTS_GIC_V3_ITS_H__ diff --git a/tools/testing/selftests/kvm/lib/arm64/gic_v3_its.c b/tools/testing/selftests/kvm/lib/arm64/gic_v3_its.c index 0e2f8ed90f30..d9ee331074ea 100644 --- a/tools/testing/selftests/kvm/lib/arm64/gic_v3_its.c +++ b/tools/testing/selftests/kvm/lib/arm64/gic_v3_its.c @@ -253,3 +253,14 @@ void its_send_invall_cmd(void *cmdq_base, u32 collection_id) its_send_cmd(cmdq_base, &cmd); } + +void its_send_sync_cmd(void *cmdq_base, u32 vcpu_id) +{ + struct its_cmd_block cmd = {}; + + its_encode_cmd(&cmd, GITS_CMD_SYNC); + its_encode_target(&cmd, procnum_to_rdbase(vcpu_id)); + + its_send_cmd(cmdq_base, &cmd); +} + -- 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, 8 hours

3
2
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-kselftest-mirror