| Message ID | 20221216183122.2040142-2-davemarchevsky@fb.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | BPF |
| Headers | show |
| Series | [v2,bpf-next,1/2] bpf, x86: Improve PROBE_MEM runtime load check | expand |
On 12/16/22 10:31 AM, Dave Marchevsky wrote: > This patch adds a test exercising logic that was fixed / improved in > the previous patch in the series, as well as general sanity checking for > jit's PROBE_MEM logic which should've been unaffected by the previous > patch. > > The added verifier test does the following: > * Acquire a referenced kptr to struct prog_test_ref_kfunc using > existing net/bpf/test_run.c kfunc > * Helper returns ptr to a specific prog_test_ref_kfunc whose first > two fields - both ints - have been prepopulated w/ vals 42 and > 108, respectively > * kptr_xchg the acquired ptr into an arraymap > * Do a direct map_value load of the just-added ptr > * Goal of all this setup is to get an unreferenced kptr pointing to > struct with ints of known value, which is the result of this step > * Using unreferenced kptr obtained in previous step, do loads of > prog_test_ref_kfunc.a (offset 0) and .b (offset 4) > * Then incr the kptr by 8 and load prog_test_ref_kfunc.a again (this > time at offset -8) > * Add all the loaded ints together and return > > Before the PROBE_MEM fixes in previous patch, the loads at offset 0 and > 4 would succeed, while the load at offset -8 would incorrectly fail > runtime check emitted by the JIT and 0 out dst reg as a result. This > confirmed by retval of 150 for this test before previous patch - since > second .a read is 0'd out - and a retval of 192 with the fixed logic. > > The test exercises the two optimizations to fixed logic added in last > patch as well: > * First load, with insn "r8 = *(u32 *)(r9 + 0)" exercises "insn->off > is 0, no need to add / sub from src_reg" optimization > * Third load, with insn "r9 = *(u32 *)(r9 - 8)" exercises "src_reg == > dst_reg, no need to restore src_reg after load" optimization > > Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com> Ack with one nit below. Acked-by: Yonghong Song <yhs@fb.com> > --- > v1 -> v2: lore.kernel.org/bpf/20221213182726.325137-2-davemarchevsky@fb.com > * Rewrite the test to be a "normal" C prog in selftests/bpf/progs. Result > is a much easier-to-understand test with assembly used only for the 3 > loads. (Yonghong) > > .../selftests/bpf/prog_tests/jit_probe_mem.c | 28 +++++++++ > .../selftests/bpf/progs/jit_probe_mem.c | 61 +++++++++++++++++++ > 2 files changed, 89 insertions(+) > create mode 100644 tools/testing/selftests/bpf/prog_tests/jit_probe_mem.c > create mode 100644 tools/testing/selftests/bpf/progs/jit_probe_mem.c > > diff --git a/tools/testing/selftests/bpf/prog_tests/jit_probe_mem.c b/tools/testing/selftests/bpf/prog_tests/jit_probe_mem.c > new file mode 100644 > index 000000000000..5639428607e6 > --- /dev/null > +++ b/tools/testing/selftests/bpf/prog_tests/jit_probe_mem.c > @@ -0,0 +1,28 @@ > +// SPDX-License-Identifier: GPL-2.0 > +/* Copyright (c) 2022 Meta Platforms, Inc. and affiliates. */ > +#include <test_progs.h> > +#include <network_helpers.h> > + > +#include "jit_probe_mem.skel.h" > + > +void test_jit_probe_mem(void) > +{ > + LIBBPF_OPTS(bpf_test_run_opts, opts, > + .data_in = &pkt_v4, > + .data_size_in = sizeof(pkt_v4), > + .repeat = 1, > + ); > + struct jit_probe_mem *skel; > + int ret; > + > + skel = jit_probe_mem__open_and_load(); > + if (!ASSERT_OK_PTR(skel, "jit_probe_mem__open_and_load")) > + return; > + > + ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_jit_probe_mem), &opts); > + ASSERT_OK(ret, "jit_probe_mem ret"); > + ASSERT_OK(opts.retval, "jit_probe_mem opts.retval"); > + ASSERT_EQ(skel->data->total_sum, 192, "jit_probe_mem total_sum"); > + > + jit_probe_mem__destroy(skel); > +} > diff --git a/tools/testing/selftests/bpf/progs/jit_probe_mem.c b/tools/testing/selftests/bpf/progs/jit_probe_mem.c > new file mode 100644 > index 000000000000..3bb8af4df837 > --- /dev/null > +++ b/tools/testing/selftests/bpf/progs/jit_probe_mem.c > @@ -0,0 +1,61 @@ > +// SPDX-License-Identifier: GPL-2.0 > +/* Copyright (c) 2022 Meta Platforms, Inc. and affiliates. */ > +#include <vmlinux.h> > +#include <bpf/bpf_tracing.h> > +#include <bpf/bpf_helpers.h> > + > +static struct prog_test_ref_kfunc __kptr_ref *v; > +long total_sum = -1; > + > +extern struct prog_test_ref_kfunc *bpf_kfunc_call_test_acquire(unsigned long *sp) __ksym; > +extern void bpf_kfunc_call_test_release(struct prog_test_ref_kfunc *p) __ksym; > + > +SEC("tc") > +int test_jit_probe_mem(struct __sk_buff *ctx) > +{ > + struct prog_test_ref_kfunc *p; > + unsigned long zero = 0, sum; > + > + p = bpf_kfunc_call_test_acquire(&zero); > + if (!p) > + return 1; > + > + p = bpf_kptr_xchg(&v, p); > + if (p) > + goto release_out; > + > + /* Direct map value access of kptr, should be PTR_UNTRUSTED */ > + p = v; > + if (!p) > + return 1; > + > + asm volatile ( > + "r9 = %[p];\n" > + "%[sum] = 0;\n" > + > + /* r8 = p->a */ > + "r8 = *(u32 *)(r9 + 0);\n" > + "%[sum] += r8;\n" > + > + /* r8 = p->b */ > + "r8 = *(u32 *)(r9 + 4);\n" > + "%[sum] += r8;\n" > + > + "r9 += 8;\n" > + /* r9 = p->a */ > + "r9 = *(u32 *)(r9 - 8);\n" > + "%[sum] += r9;\n" All these '\n' are not necessary. > + > + : [sum] "=r"(sum) > + : [p] "r"(p) > + : "r8", "r9" > + ); > + > + total_sum = sum; > + return 0; > +release_out: > + bpf_kfunc_call_test_release(p); > + return 1; > +} > + > +char _license[] SEC("license") = "GPL";
diff --git a/tools/testing/selftests/bpf/prog_tests/jit_probe_mem.c b/tools/testing/selftests/bpf/prog_tests/jit_probe_mem.c new file mode 100644 index 000000000000..5639428607e6 --- /dev/null +++ b/tools/testing/selftests/bpf/prog_tests/jit_probe_mem.c @@ -0,0 +1,28 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) 2022 Meta Platforms, Inc. and affiliates. */ +#include <test_progs.h> +#include <network_helpers.h> + +#include "jit_probe_mem.skel.h" + +void test_jit_probe_mem(void) +{ + LIBBPF_OPTS(bpf_test_run_opts, opts, + .data_in = &pkt_v4, + .data_size_in = sizeof(pkt_v4), + .repeat = 1, + ); + struct jit_probe_mem *skel; + int ret; + + skel = jit_probe_mem__open_and_load(); + if (!ASSERT_OK_PTR(skel, "jit_probe_mem__open_and_load")) + return; + + ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_jit_probe_mem), &opts); + ASSERT_OK(ret, "jit_probe_mem ret"); + ASSERT_OK(opts.retval, "jit_probe_mem opts.retval"); + ASSERT_EQ(skel->data->total_sum, 192, "jit_probe_mem total_sum"); + + jit_probe_mem__destroy(skel); +} diff --git a/tools/testing/selftests/bpf/progs/jit_probe_mem.c b/tools/testing/selftests/bpf/progs/jit_probe_mem.c new file mode 100644 index 000000000000..3bb8af4df837 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/jit_probe_mem.c @@ -0,0 +1,61 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) 2022 Meta Platforms, Inc. and affiliates. */ +#include <vmlinux.h> +#include <bpf/bpf_tracing.h> +#include <bpf/bpf_helpers.h> + +static struct prog_test_ref_kfunc __kptr_ref *v; +long total_sum = -1; + +extern struct prog_test_ref_kfunc *bpf_kfunc_call_test_acquire(unsigned long *sp) __ksym; +extern void bpf_kfunc_call_test_release(struct prog_test_ref_kfunc *p) __ksym; + +SEC("tc") +int test_jit_probe_mem(struct __sk_buff *ctx) +{ + struct prog_test_ref_kfunc *p; + unsigned long zero = 0, sum; + + p = bpf_kfunc_call_test_acquire(&zero); + if (!p) + return 1; + + p = bpf_kptr_xchg(&v, p); + if (p) + goto release_out; + + /* Direct map value access of kptr, should be PTR_UNTRUSTED */ + p = v; + if (!p) + return 1; + + asm volatile ( + "r9 = %[p];\n" + "%[sum] = 0;\n" + + /* r8 = p->a */ + "r8 = *(u32 *)(r9 + 0);\n" + "%[sum] += r8;\n" + + /* r8 = p->b */ + "r8 = *(u32 *)(r9 + 4);\n" + "%[sum] += r8;\n" + + "r9 += 8;\n" + /* r9 = p->a */ + "r9 = *(u32 *)(r9 - 8);\n" + "%[sum] += r9;\n" + + : [sum] "=r"(sum) + : [p] "r"(p) + : "r8", "r9" + ); + + total_sum = sum; + return 0; +release_out: + bpf_kfunc_call_test_release(p); + return 1; +} + +char _license[] SEC("license") = "GPL";
This patch adds a test exercising logic that was fixed / improved in the previous patch in the series, as well as general sanity checking for jit's PROBE_MEM logic which should've been unaffected by the previous patch. The added verifier test does the following: * Acquire a referenced kptr to struct prog_test_ref_kfunc using existing net/bpf/test_run.c kfunc * Helper returns ptr to a specific prog_test_ref_kfunc whose first two fields - both ints - have been prepopulated w/ vals 42 and 108, respectively * kptr_xchg the acquired ptr into an arraymap * Do a direct map_value load of the just-added ptr * Goal of all this setup is to get an unreferenced kptr pointing to struct with ints of known value, which is the result of this step * Using unreferenced kptr obtained in previous step, do loads of prog_test_ref_kfunc.a (offset 0) and .b (offset 4) * Then incr the kptr by 8 and load prog_test_ref_kfunc.a again (this time at offset -8) * Add all the loaded ints together and return Before the PROBE_MEM fixes in previous patch, the loads at offset 0 and 4 would succeed, while the load at offset -8 would incorrectly fail runtime check emitted by the JIT and 0 out dst reg as a result. This confirmed by retval of 150 for this test before previous patch - since second .a read is 0'd out - and a retval of 192 with the fixed logic. The test exercises the two optimizations to fixed logic added in last patch as well: * First load, with insn "r8 = *(u32 *)(r9 + 0)" exercises "insn->off is 0, no need to add / sub from src_reg" optimization * Third load, with insn "r9 = *(u32 *)(r9 - 8)" exercises "src_reg == dst_reg, no need to restore src_reg after load" optimization Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com> --- v1 -> v2: lore.kernel.org/bpf/20221213182726.325137-2-davemarchevsky@fb.com * Rewrite the test to be a "normal" C prog in selftests/bpf/progs. Result is a much easier-to-understand test with assembly used only for the 3 loads. (Yonghong) .../selftests/bpf/prog_tests/jit_probe_mem.c | 28 +++++++++ .../selftests/bpf/progs/jit_probe_mem.c | 61 +++++++++++++++++++ 2 files changed, 89 insertions(+) create mode 100644 tools/testing/selftests/bpf/prog_tests/jit_probe_mem.c create mode 100644 tools/testing/selftests/bpf/progs/jit_probe_mem.c