Skip to contentSkip to conversation

samples/bpf: A couple s390 fixes

PATCH In review Ilya Leoshkevich 4 participants 21 messages Jul 28, 2020 – Aug 5, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover samples/bpf: A couple s390 fixes

This series contains small fixes that make samples work on s390:

Patch 1: Fix building M=samples/bpf with O=.
Patch 2: Use more portable __sys_connect instead of
SYSCALL(sys_connect).
Patch 3: Use bpf_probe_read_kernel instead of bpf_probe_read in
libbpf.

Ilya Leoshkevich (3):
samples/bpf: Fix building out of srctree
samples/bpf: Fix test_map_in_map on s390
libbpf: Use bpf_probe_read_kernel

 samples/bpf/Makefile               |  1 +
 samples/bpf/test_map_in_map_kern.c |  7 ++--
 tools/lib/bpf/bpf_core_read.h      | 51 ++++++++++++++++--------------
 tools/lib/bpf/bpf_tracing.h        | 15 ++++++---
 4 files changed, 41 insertions(+), 33 deletions(-)
Signature · 1 line
2.25.4
#1 lore raw

PATCH 1/3 samples/bpf: Fix building out of srctree

Building BPF samples out of srctree fails, because the output directory for progs shared with selftests (CGROUP_HELPERS, TRACE_HELPERS) is missing and the compiler cannot create output files.

Fix by creating the output directory in Makefile.

---
 samples/bpf/Makefile | 1 +
 1 file changed, 1 insertion(+)
samples/bpf/Makefile +1 -0
index f87ee02073ba..81ba0beca0a3 100644
@@ -301,6 +301,7 @@ $(obj)/hbm_edt_kern.o: $(src)/hbm.h $(src)/hbm_kern.h
 # But, there is no easy way to fix it, so just exclude it since it is
 # useless for BPF samples.
 $(obj)/%.o: $(src)/%.c
+	$(Q)mkdir -p $(@D)
 	@echo "  CLANG-bpf " $@
 	$(Q)$(CLANG) $(NOSTDINC_FLAGS) $(LINUXINCLUDE) $(BPF_EXTRA_CFLAGS) \
 		-I$(obj) -I$(srctree)/tools/testing/selftests/bpf/ \
Signature · 1 line
2.25.4
reviewer
#2 lore raw

Re: [PATCH bpf-next 1/3] samples/bpf: Fix building out of srctree

Quoted 6 lines · Ilya Leoshkevich

Building BPF samples out of srctree fails, because the output directory for progs shared with selftests (CGROUP_HELPERS, TRACE_HELPERS) is missing and the compiler cannot create output files.

Fix by creating the output directory in Makefile.

What is the make command line we use here? I am trying:

make M=samples/bpf O=./xxx

w/o this patch, make created ./xxx/samples/bpf.

Did I miss something?

Thanks,
Song

#3 lore raw

Re: [PATCH bpf-next 1/3] samples/bpf: Fix building out of srctree

Quoted 9 lines · Song Liu
Quoted 6 lines · Ilya Leoshkevich

Building BPF samples out of srctree fails, because the output directory
for progs shared with selftests (CGROUP_HELPERS, TRACE_HELPERS) is missing and the compiler cannot create output files.

Fix by creating the output directory in Makefile.

Quoted 7 lines

What is the make command line we use here? I am trying:

make M=samples/bpf O=./xxx

w/o this patch, make created ./xxx/samples/bpf.

Did I miss something?

I'm using

make O=$HOME/linux-build 'CC=ccache gcc' M=samples/bpf -j12

My make version is GNU Make 4.2.1.

Best regards,
Ilya

#4 lore raw

Re: [PATCH bpf-next 1/3] samples/bpf: Fix building out of srctree

Quoted 18 lines · Ilya Leoshkevich
Quoted 17 lines · Song Liu

On Tue, Jul 28, 2020 at 5:15 AM Ilya Leoshkevich <[email protected]

Quoted 1 line
Quoted 7 lines · wrote:

Building BPF samples out of srctree fails, because the output directory
for progs shared with selftests (CGROUP_HELPERS, TRACE_HELPERS) is
missing and the compiler cannot create output files.

Fix by creating the output directory in Makefile.

What is the make command line we use here? I am trying:

make M=samples/bpf O=./xxx

w/o this patch, make created ./xxx/samples/bpf.

Quoted 7 lines
Quoted 1 line

Did I miss something?

I'm using

make O=$HOME/linux-build 'CC=ccache gcc' M=samples/bpf -j12

My make version is GNU Make 4.2.1.

Hmm, I wanted to elaborate a little bit more on why this fix should work, and realized that it works only by accident: cgroup_helpers and trace_helpers are userspace code, and my fix modifies the recipe for building bpf programs.

Please disregard this patch - I will have to come up with a better fix.

#5 lore raw

PATCH 2/3 samples/bpf: Fix test_map_in_map on s390

In reply to Ilya Leoshkevich

s390 uses socketcall multiplexer instead of individual socket syscalls. Therefore, "kprobe/" SYSCALL(sys_connect) does not trigger and test_map_in_map fails. Fix by using "kprobe/__sys_connect" instead.

---
 samples/bpf/test_map_in_map_kern.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)
samples/bpf/test_map_in_map_kern.c +3 -4
index 8def45c5b697..b0200c8eac09 100644
@@ -103,10 +103,9 @@ static __always_inline int do_inline_hash_lookup(void *inner_map, u32 port)
 	return result ? *result : -ENOENT;
 }
 
-SEC("kprobe/" SYSCALL(sys_connect))
+SEC("kprobe/__sys_connect")
 int trace_sys_connect(struct pt_regs *ctx)
 {
-	struct pt_regs *real_regs = (struct pt_regs *)PT_REGS_PARM1_CORE(ctx);
 	struct sockaddr_in6 *in6;
 	u16 test_case, port, dst6[8];
 	int addrlen, ret, inline_ret, ret_key = 0;
@@ -114,8 +113,8 @@ int trace_sys_connect(struct pt_regs *ctx)
 	void *outer_map, *inner_map;
 	bool inline_hash = false;
 
-	in6 = (struct sockaddr_in6 *)PT_REGS_PARM2_CORE(real_regs);
-	addrlen = (int)PT_REGS_PARM3_CORE(real_regs);
+	in6 = (struct sockaddr_in6 *)PT_REGS_PARM2_CORE(ctx);
+	addrlen = (int)PT_REGS_PARM3_CORE(ctx);
 
 	if (addrlen != sizeof(*in6))
 		return 0;
Signature · 1 line
2.25.4
reviewer
#6 lore raw

Re: [PATCH bpf-next 2/3] samples/bpf: Fix test_map_in_map on s390

Quoted 4 lines · Ilya Leoshkevich

s390 uses socketcall multiplexer instead of individual socket syscalls. Therefore, "kprobe/" SYSCALL(sys_connect) does not trigger and test_map_in_map fails. Fix by using "kprobe/__sys_connect" instead.

samples/bpf is in semi-deprecated state. I tried for quite some time, but still cannot build it all successfully. So I apologize for bounding the question to you...

From the code, we do the SYSCALL() trick to change the exact name for different architecture. Would this change break the same file for x86?

Thanks,
Song

Quoted 36 lines
---
 samples/bpf/test_map_in_map_kern.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)
samples/bpf/test_map_in_map_kern.c +3 -5
index 8def45c5b697..b0200c8eac09 100644
@@ -103,10 +103,9 @@ static __always_inline int do_inline_hash_lookup(void *inner_map, u32 port)
        return result ? *result : -ENOENT;
 }
 
-SEC("kprobe/" SYSCALL(sys_connect))
+SEC("kprobe/__sys_connect")
 int trace_sys_connect(struct pt_regs *ctx)
 {
-       struct pt_regs *real_regs = (struct pt_regs *)PT_REGS_PARM1_CORE(ctx);
        struct sockaddr_in6 *in6;
        u16 test_case, port, dst6[8];
        int addrlen, ret, inline_ret, ret_key = 0;
@@ -114,8 +113,8 @@ int trace_sys_connect(struct pt_regs *ctx)
        void *outer_map, *inner_map;
        bool inline_hash = false;
 
-       in6 = (struct sockaddr_in6 *)PT_REGS_PARM2_CORE(real_regs);
-       addrlen = (int)PT_REGS_PARM3_CORE(real_regs);
+       in6 = (struct sockaddr_in6 *)PT_REGS_PARM2_CORE(ctx);
+       addrlen = (int)PT_REGS_PARM3_CORE(ctx);
 
        if (addrlen != sizeof(*in6))
                return 0;
--

2.25.4

#7 lore raw

Re: [PATCH bpf-next 2/3] samples/bpf: Fix test_map_in_map on s390

Quoted 7 lines · Song Liu
Quoted 4 lines · Ilya Leoshkevich

s390 uses socketcall multiplexer instead of individual socket syscalls.
Therefore, "kprobe/" SYSCALL(sys_connect) does not trigger and test_map_in_map fails. Fix by using "kprobe/__sys_connect" instead.

Quoted 8 lines

samples/bpf is in semi-deprecated state. I tried for quite some time, but still
cannot build it all successfully. So I apologize for bounding the question to you...

From the code, we do the SYSCALL() trick to change the exact name for different architecture. Would this change break the same file for x86?

No, it shouldn't - __sys_connect exists on all architectures and gets control from both regular socket syscalls and socketcall multiplexer. I tested it on x86 and it worked for me.
It's also already used by tools/testing/selftests/bpf/progs/test_probe_user.c

Best regards,
Ilya

#8 lore raw

PATCH 3/3 libbpf: Use bpf_probe_read_kernel

In reply to Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

---
 tools/lib/bpf/bpf_core_read.h | 51 ++++++++++++++++++-----------------
 tools/lib/bpf/bpf_tracing.h   | 15 +++++++----
 2 files changed, 37 insertions(+), 29 deletions(-)
tools/lib/bpf/bpf_core_read.h +27 -24
index eae5cccff761..b108381fbec4 100644
@@ -23,28 +23,29 @@ enum bpf_field_info_kind {
 	__builtin_preserve_field_info((src)->field, BPF_FIELD_##info)
 
 #if __BYTE_ORDER == __LITTLE_ENDIAN
-#define __CORE_BITFIELD_PROBE_READ(dst, src, fld)			      \
-	bpf_probe_read((void *)dst,					      \
-		       __CORE_RELO(src, fld, BYTE_SIZE),		      \
-		       (const void *)src + __CORE_RELO(src, fld, BYTE_OFFSET))
+#define __CORE_BITFIELD_PROBE_READ(dst, src, fld)                              \
+	bpf_probe_read_kernel((void *)dst, __CORE_RELO(src, fld, BYTE_SIZE),   \
+			      (const void *)src +                              \
+				      __CORE_RELO(src, fld, BYTE_OFFSET))
 #else
 /* semantics of LSHIFT_64 assumes loading values into low-ordered bytes, so
  * for big-endian we need to adjust destination pointer accordingly, based on
  * field byte size
  */
-#define __CORE_BITFIELD_PROBE_READ(dst, src, fld)			      \
-	bpf_probe_read((void *)dst + (8 - __CORE_RELO(src, fld, BYTE_SIZE)),  \
-		       __CORE_RELO(src, fld, BYTE_SIZE),		      \
-		       (const void *)src + __CORE_RELO(src, fld, BYTE_OFFSET))
+#define __CORE_BITFIELD_PROBE_READ(dst, src, fld)                              \
+	bpf_probe_read_kernel(                                                 \
+		(void *)dst + (8 - __CORE_RELO(src, fld, BYTE_SIZE)),          \
+		__CORE_RELO(src, fld, BYTE_SIZE),                              \
+		(const void *)src + __CORE_RELO(src, fld, BYTE_OFFSET))
 #endif
 
 /*
  * Extract bitfield, identified by s->field, and return its value as u64.
  * All this is done in relocatable manner, so bitfield changes such as
  * signedness, bit size, offset changes, this will be handled automatically.
- * This version of macro is using bpf_probe_read() to read underlying integer
- * storage. Macro functions as an expression and its return type is
- * bpf_probe_read()'s return value: 0, on success, <0 on error.
+ * This version of macro is using bpf_probe_read_kernel() to read underlying
+ * integer storage. Macro functions as an expression and its return type is
+ * bpf_probe_read_kernel()'s return value: 0, on success, <0 on error.
  */
 #define BPF_CORE_READ_BITFIELD_PROBED(s, field) ({			      \
 	unsigned long long val = 0;					      \
@@ -99,9 +100,9 @@ enum bpf_field_info_kind {
 	__builtin_preserve_field_info(field, BPF_FIELD_BYTE_SIZE)
 
 /*
- * bpf_core_read() abstracts away bpf_probe_read() call and captures offset
- * relocation for source address using __builtin_preserve_access_index()
- * built-in, provided by Clang.
+ * bpf_core_read() abstracts away bpf_probe_read_kernel() call and captures
+ * offset relocation for source address using
+ * __builtin_preserve_access_index() built-in, provided by Clang.
  *
  * __builtin_preserve_access_index() takes as an argument an expression of
  * taking an address of a field within struct/union. It makes compiler emit
@@ -114,18 +115,18 @@ enum bpf_field_info_kind {
  * actual field offset, based on target kernel BTF type that matches original
  * (local) BTF, used to record relocation.
  */
-#define bpf_core_read(dst, sz, src)					    \
-	bpf_probe_read(dst, sz,						    \
-		       (const void *)__builtin_preserve_access_index(src))
+#define bpf_core_read(dst, sz, src)                                            \
+	bpf_probe_read_kernel(                                                 \
+		dst, sz, (const void *)__builtin_preserve_access_index(src))
 
 /*
- * bpf_core_read_str() is a thin wrapper around bpf_probe_read_str()
+ * bpf_core_read_str() is a thin wrapper around bpf_probe_read_kernel_str()
  * additionally emitting BPF CO-RE field relocation for specified source
  * argument.
  */
-#define bpf_core_read_str(dst, sz, src)					    \
-	bpf_probe_read_str(dst, sz,					    \
-			   (const void *)__builtin_preserve_access_index(src))
+#define bpf_core_read_str(dst, sz, src)                                        \
+	bpf_probe_read_kernel_str(                                             \
+		dst, sz, (const void *)__builtin_preserve_access_index(src))
 
 #define ___concat(a, b) a ## b
 #define ___apply(fn, n) ___concat(fn, n)
@@ -239,15 +240,17 @@ enum bpf_field_info_kind {
  *	int x = BPF_CORE_READ(s, a.b.c, d.e, f, g);
  *
  * BPF_CORE_READ will decompose above statement into 4 bpf_core_read (BPF
- * CO-RE relocatable bpf_probe_read() wrapper) calls, logically equivalent to:
+ * CO-RE relocatable bpf_probe_read_kernel() wrapper) calls, logically
+ * equivalent to:
  * 1. const void *__t = s->a.b.c;
  * 2. __t = __t->d.e;
  * 3. __t = __t->f;
  * 4. return __t->g;
  *
  * Equivalence is logical, because there is a heavy type casting/preservation
- * involved, as well as all the reads are happening through bpf_probe_read()
- * calls using __builtin_preserve_access_index() to emit CO-RE relocations.
+ * involved, as well as all the reads are happening through
+ * bpf_probe_read_kernel() calls using __builtin_preserve_access_index() to
+ * emit CO-RE relocations.
  *
  * N.B. Only up to 9 "field accessors" are supported, which should be more
  * than enough for any practical purpose.
tools/lib/bpf/bpf_tracing.h +10 -5
index 58eceb884df3..4eb3be4130f0 100644
@@ -288,11 +288,16 @@ struct pt_regs;
 #define BPF_KPROBE_READ_RET_IP(ip, ctx)		({ (ip) = PT_REGS_RET(ctx); })
 #define BPF_KRETPROBE_READ_RET_IP		BPF_KPROBE_READ_RET_IP
 #else
-#define BPF_KPROBE_READ_RET_IP(ip, ctx)					    \
-	({ bpf_probe_read(&(ip), sizeof(ip), (void *)PT_REGS_RET(ctx)); })
-#define BPF_KRETPROBE_READ_RET_IP(ip, ctx)				    \
-	({ bpf_probe_read(&(ip), sizeof(ip),				    \
-			  (void *)(PT_REGS_FP(ctx) + sizeof(ip))); })
+#define BPF_KPROBE_READ_RET_IP(ip, ctx)                                        \
+	({                                                                     \
+		bpf_probe_read_kernel(&(ip), sizeof(ip),                       \
+				      (void *)PT_REGS_RET(ctx));               \
+	})
+#define BPF_KRETPROBE_READ_RET_IP(ip, ctx)                                     \
+	({                                                                     \
+		bpf_probe_read_kernel(&(ip), sizeof(ip),                       \
+				      (void *)(PT_REGS_FP(ctx) + sizeof(ip))); \
+	})
 #endif
 
 #define ___bpf_concat(a, b) a ## b
Signature · 1 line
2.25.4
#9 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

Quoted 7 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

---

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Quoted 4 lines
 tools/lib/bpf/bpf_core_read.h | 51 ++++++++++++++++++-----------------
 tools/lib/bpf/bpf_tracing.h   | 15 +++++++----
 2 files changed, 37 insertions(+), 29 deletions(-)

[...]

maintainer
#10 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

Quoted 8 lines · Andrii Nakryiko
Quoted 6 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

Quoted 8 lines

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Could you elaborate what this means wrt dependency on latest clang? Given clang releases have a rather long cadence, what about existing users with current clang releases?

Quoted 7 lines
Quoted 4 lines
  tools/lib/bpf/bpf_core_read.h | 51 ++++++++++++++++++-----------------
  tools/lib/bpf/bpf_tracing.h   | 15 +++++++----
  2 files changed, 37 insertions(+), 29 deletions(-)

[...]

#11 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

Quoted 19 lines · Daniel Borkmann
Quoted 8 lines · Andrii Nakryiko
Quoted 6 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

Quoted 8 lines

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Quoted 3 lines

Could you elaborate what this means wrt dependency on latest clang? Given clang releases have a rather long cadence, what about existing users with current clang releases?

So the overall idea is to use something like this to do kernel reads:

static __always_inline int bpf_probe_read_universal(void *dst, u32 sz, const void *src)

{
    if (bpf_core_type_exists(btf_bpf_probe_read_kernel))
        return bpf_probe_read_kernel(dst, sz, src);
    else
        return bpf_probe_read(dst, sz, src);
}

And then use bpf_probe_read_universal() in BPF_CORE_READ and family.

This approach relies on few things:

1. each BPF helper has a corresponding btf_<helper-name> type defined for it
2. bpf_core_type_exists(some_type) returns 0 or 1, depending if specified type is found in kernel BTF (so needs kernel BTF, of course). This is the part me and Yonghong are working on at the moment.
3. verifier's dead code elimination, which will leave only bpf_probe_read() or bpf_probe_read_kernel() calls and will remove the other one. So on older kernels, there will never be unsupoorted call to bpf_probe_read_kernel().

The new type existence relocation requires the latest Clang. So the way to deal with older Clangs would be to just fallback to bpf_probe_read, if we detect that Clang is too old and can't emit necessary relocation.

If that's not an acceptable plan, then one can "parameterize" BPF_CORE_READ macro family by re-defining bpf_core_read() macro. Right now it's defined as:

#define bpf_core_read(dst, sz, src) \
bpf_probe_read(dst, sz, (const void *)__builtin_preserve_access_index(src))

Re-defining it in terms of bpf_probe_read_kernel is trivial, but I can't do it for BPF_CORE_READ, because it will break all the users of bpf_core_read.h that run on older kernels.

Quoted 9 lines
Quoted 8 lines
Quoted 5 lines
  tools/lib/bpf/bpf_core_read.h | 51 ++++++++++++++++++-----------------
  tools/lib/bpf/bpf_tracing.h   | 15 +++++++----
  2 files changed, 37 insertions(+), 29 deletions(-)

[...]

maintainer
#12 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

Quoted 48 lines · Andrii Nakryiko
Quoted 18 lines · Daniel Borkmann
Quoted 8 lines · Andrii Nakryiko
Quoted 6 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

Quoted 8 lines

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Quoted 3 lines

Could you elaborate what this means wrt dependency on latest clang? Given clang releases have a rather long cadence, what about existing users with current clang releases?

So the overall idea is to use something like this to do kernel reads:

static __always_inline int bpf_probe_read_universal(void *dst, u32 sz, const void *src)

{
     if (bpf_core_type_exists(btf_bpf_probe_read_kernel))
         return bpf_probe_read_kernel(dst, sz, src);
     else
         return bpf_probe_read(dst, sz, src);
}

And then use bpf_probe_read_universal() in BPF_CORE_READ and family.

This approach relies on few things:

1. each BPF helper has a corresponding btf_<helper-name> type defined for it
2. bpf_core_type_exists(some_type) returns 0 or 1, depending if specified type is found in kernel BTF (so needs kernel BTF, of course). This is the part me and Yonghong are working on at the moment.
3. verifier's dead code elimination, which will leave only bpf_probe_read() or bpf_probe_read_kernel() calls and will remove the other one. So on older kernels, there will never be unsupoorted call to bpf_probe_read_kernel().

Quoted 4 lines

The new type existence relocation requires the latest Clang. So the way to deal with older Clangs would be to just fallback to bpf_probe_read, if we detect that Clang is too old and can't emit necessary relocation.

Okay, seems reasonable overall. One question though: couldn't libbpf transparently fix up the selection of bpf_probe_read() vs bpf_probe_read_kernel()? E.g. it would probe the kernel whether bpf_probe_read_kernel() is available and if it is then it would rewrite the raw call number from the instruction from bpf_probe_read() into the one for bpf_probe_read_kernel()? I guess the question then becomes whether the original use for bpf_probe_read() was related to CO-RE. But I think this could also be overcome by adding a fake helper signature in libbpf with a unreasonable high number that is dedicated to probing mem via CO-RE and then libbpf picks the right underlying helper call number for the insn. That avoids fiddling with macros and need for new clang version, no (unless I'm missing something)?

Quoted 21 lines

If that's not an acceptable plan, then one can "parameterize" BPF_CORE_READ macro family by re-defining bpf_core_read() macro. Right now it's defined as:

#define bpf_core_read(dst, sz, src) \
bpf_probe_read(dst, sz, (const void *)__builtin_preserve_access_index(src))

Re-defining it in terms of bpf_probe_read_kernel is trivial, but I can't do it for BPF_CORE_READ, because it will break all the users of bpf_core_read.h that run on older kernels.

Quoted 9 lines
Quoted 8 lines
Quoted 5 lines
   tools/lib/bpf/bpf_core_read.h | 51 ++++++++++++++++++-----------------
   tools/lib/bpf/bpf_tracing.h   | 15 +++++++----
   2 files changed, 37 insertions(+), 29 deletions(-)

[...]

#13 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

Quoted 55 lines · Daniel Borkmann
Quoted 48 lines · Andrii Nakryiko
Quoted 18 lines · Daniel Borkmann
Quoted 8 lines · Andrii Nakryiko
Quoted 6 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

Quoted 8 lines

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Quoted 3 lines

Could you elaborate what this means wrt dependency on latest clang? Given clang releases have a rather long cadence, what about existing users with current clang releases?

So the overall idea is to use something like this to do kernel reads:

static __always_inline int bpf_probe_read_universal(void *dst, u32 sz, const void *src)

{
     if (bpf_core_type_exists(btf_bpf_probe_read_kernel))
         return bpf_probe_read_kernel(dst, sz, src);
     else
         return bpf_probe_read(dst, sz, src);
}

And then use bpf_probe_read_universal() in BPF_CORE_READ and family.

This approach relies on few things:

1. each BPF helper has a corresponding btf_<helper-name> type defined for it
2. bpf_core_type_exists(some_type) returns 0 or 1, depending if specified type is found in kernel BTF (so needs kernel BTF, of course). This is the part me and Yonghong are working on at the moment.
3. verifier's dead code elimination, which will leave only bpf_probe_read() or bpf_probe_read_kernel() calls and will remove the other one. So on older kernels, there will never be unsupoorted call to bpf_probe_read_kernel().

Quoted 4 lines

The new type existence relocation requires the latest Clang. So the way to deal with older Clangs would be to just fallback to bpf_probe_read, if we detect that Clang is too old and can't emit necessary relocation.

Quoted 10 lines

Okay, seems reasonable overall. One question though: couldn't libbpf transparently fix up the selection of bpf_probe_read() vs bpf_probe_read_kernel()? E.g. it would probe the kernel whether bpf_probe_read_kernel() is available and if it is then it would rewrite the raw call number from the instruction from bpf_probe_read() into the one for bpf_probe_read_kernel()? I guess the question then becomes whether the original use for bpf_probe_read() was related to CO-RE. But I think this could also be overcome by adding a fake helper signature in libbpf with a unreasonable high number that is dedicated to probing mem via CO-RE and then libbpf picks the right underlying helper call number for the insn. That avoids fiddling with macros and need for new clang version, no (unless I'm missing something)?

Libbpf could do it, but I'm a bit worried that unconditionally changing bpf_probe_read() into bpf_probe_read_kernel() is going to be wrong in some cases. If that wasn't the case, why wouldn't we just re-purpose bpf_probe_read() into bpf_probe_read_kernel() in kernel itself, right?

But fear not about old Clang support. The bpf_core_type_exists() will use a new built-in, and I'll be able to detect its presence with __has_builtin(X) check in Clang. So it will be completely transparent to users in the end.

Quoted 23 lines
Quoted 22 lines

If that's not an acceptable plan, then one can "parameterize" BPF_CORE_READ macro family by re-defining bpf_core_read() macro. Right now it's defined as:

#define bpf_core_read(dst, sz, src) \
bpf_probe_read(dst, sz, (const void *)__builtin_preserve_access_index(src))

Re-defining it in terms of bpf_probe_read_kernel is trivial, but I can't do it for BPF_CORE_READ, because it will break all the users of bpf_core_read.h that run on older kernels.

Quoted 9 lines
Quoted 8 lines
Quoted 5 lines
   tools/lib/bpf/bpf_core_read.h | 51 ++++++++++++++++++-----------------
   tools/lib/bpf/bpf_tracing.h   | 15 +++++++----
   2 files changed, 37 insertions(+), 29 deletions(-)

[...]

maintainer
#14 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

Quoted 66 lines · Andrii Nakryiko
Quoted 54 lines · Daniel Borkmann
Quoted 48 lines · Andrii Nakryiko
Quoted 18 lines · Daniel Borkmann
Quoted 8 lines · Andrii Nakryiko
Quoted 6 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

Quoted 8 lines

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Quoted 3 lines

Could you elaborate what this means wrt dependency on latest clang? Given clang releases have a rather long cadence, what about existing users with current clang releases?

So the overall idea is to use something like this to do kernel reads:

static __always_inline int bpf_probe_read_universal(void *dst, u32 sz, const void *src)

{
      if (bpf_core_type_exists(btf_bpf_probe_read_kernel))
          return bpf_probe_read_kernel(dst, sz, src);
      else
          return bpf_probe_read(dst, sz, src);
}

And then use bpf_probe_read_universal() in BPF_CORE_READ and family.

This approach relies on few things:

1. each BPF helper has a corresponding btf_<helper-name> type defined for it
2. bpf_core_type_exists(some_type) returns 0 or 1, depending if specified type is found in kernel BTF (so needs kernel BTF, of course). This is the part me and Yonghong are working on at the moment.
3. verifier's dead code elimination, which will leave only bpf_probe_read() or bpf_probe_read_kernel() calls and will remove the other one. So on older kernels, there will never be unsupoorted call to bpf_probe_read_kernel().

Quoted 4 lines

The new type existence relocation requires the latest Clang. So the way to deal with older Clangs would be to just fallback to bpf_probe_read, if we detect that Clang is too old and can't emit necessary relocation.

Quoted 10 lines

Okay, seems reasonable overall. One question though: couldn't libbpf transparently fix up the selection of bpf_probe_read() vs bpf_probe_read_kernel()? E.g. it would probe the kernel whether bpf_probe_read_kernel() is available and if it is then it would rewrite the raw call number from the instruction from bpf_probe_read() into the one for bpf_probe_read_kernel()? I guess the question then becomes whether the original use for bpf_probe_read() was related to CO-RE. But I think this could also be overcome by adding a fake helper signature in libbpf with a unreasonable high number that is dedicated to probing mem via CO-RE and then libbpf picks the right underlying helper call number for the insn. That avoids fiddling with macros and need for new clang version, no (unless I'm missing something)?

Quoted 5 lines

Libbpf could do it, but I'm a bit worried that unconditionally changing bpf_probe_read() into bpf_probe_read_kernel() is going to be wrong in some cases. If that wasn't the case, why wouldn't we just re-purpose bpf_probe_read() into bpf_probe_read_kernel() in kernel itself, right?

Yes, that is correct, but I mentioned above that this new 'fake' helper call number that libbpf would be fixing up would only be used for bpf_probe_read{,str}() inside bpf_core_read.h.

Small example, bpf_core_read.h would be changed to (just an extract):

tools/lib/bpf/bpf_core_read.h +2 -2
index eae5cccff761..4bddb2ddf3f0 100644
@@ -115,7 +115,7 @@ enum bpf_field_info_kind {
  * (local) BTF, used to record relocation.
  */
 #define bpf_core_read(dst, sz, src)                                        \
-       bpf_probe_read(dst, sz,                                             \
+       bpf_probe_read_selector(dst, sz,                                                    \
                       (const void *)__builtin_preserve_access_index(src))
 
 /*
@@ -124,7 +124,7 @@ enum bpf_field_info_kind {
  * argument.
  */
 #define bpf_core_read_str(dst, sz, src)                                            \
-       bpf_probe_read_str(dst, sz,                                         \
+       bpf_probe_read_str_selector(dst, sz,                                        \
                           (const void *)__builtin_preserve_access_index(src))
 
 #define ___concat(a, b) a ## b

And bpf_probe_read_{,str_}selector would be defined as e.g. ...

static long (*bpf_probe_read_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -1; static long (*bpf_probe_read_str_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -2;

... where libbpf would do the fix up to either 4 or 45 for insn->imm. But it's still confined to usage in bpf_core_read.h when the CO-RE macros are used.

Quoted 4 lines

But fear not about old Clang support. The bpf_core_type_exists() will use a new built-in, and I'll be able to detect its presence with __has_builtin(X) check in Clang. So it will be completely transparent to users in the end.

Ok.

Quoted 22 lines
Quoted 22 lines
Quoted 21 lines

If that's not an acceptable plan, then one can "parameterize" BPF_CORE_READ macro family by re-defining bpf_core_read() macro. Right now it's defined as:

#define bpf_core_read(dst, sz, src) \
bpf_probe_read(dst, sz, (const void *)__builtin_preserve_access_index(src))

Re-defining it in terms of bpf_probe_read_kernel is trivial, but I can't do it for BPF_CORE_READ, because it will break all the users of bpf_core_read.h that run on older kernels.

Quoted 9 lines
Quoted 8 lines
Quoted 5 lines
    tools/lib/bpf/bpf_core_read.h | 51 ++++++++++++++++++-----------------
    tools/lib/bpf/bpf_tracing.h   | 15 +++++++----
    2 files changed, 37 insertions(+), 29 deletions(-)

[...]

#15 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

Quoted 103 lines · Daniel Borkmann
Quoted 66 lines · Andrii Nakryiko
Quoted 54 lines · Daniel Borkmann
Quoted 48 lines · Andrii Nakryiko
Quoted 18 lines · Daniel Borkmann
Quoted 8 lines · Andrii Nakryiko
Quoted 6 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

Quoted 8 lines

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Quoted 3 lines

Could you elaborate what this means wrt dependency on latest clang? Given clang releases have a rather long cadence, what about existing users with current clang releases?

So the overall idea is to use something like this to do kernel reads:

static __always_inline int bpf_probe_read_universal(void *dst, u32 sz, const void *src)

{
      if (bpf_core_type_exists(btf_bpf_probe_read_kernel))
          return bpf_probe_read_kernel(dst, sz, src);
      else
          return bpf_probe_read(dst, sz, src);
}

And then use bpf_probe_read_universal() in BPF_CORE_READ and family.

This approach relies on few things:

1. each BPF helper has a corresponding btf_<helper-name> type defined for it
2. bpf_core_type_exists(some_type) returns 0 or 1, depending if specified type is found in kernel BTF (so needs kernel BTF, of course). This is the part me and Yonghong are working on at the moment.
3. verifier's dead code elimination, which will leave only bpf_probe_read() or bpf_probe_read_kernel() calls and will remove the other one. So on older kernels, there will never be unsupoorted call to bpf_probe_read_kernel().

Quoted 4 lines

The new type existence relocation requires the latest Clang. So the way to deal with older Clangs would be to just fallback to bpf_probe_read, if we detect that Clang is too old and can't emit necessary relocation.

Quoted 10 lines

Okay, seems reasonable overall. One question though: couldn't libbpf transparently fix up the selection of bpf_probe_read() vs bpf_probe_read_kernel()? E.g. it would probe the kernel whether bpf_probe_read_kernel() is available and if it is then it would rewrite the raw call number from the instruction from bpf_probe_read() into the one for bpf_probe_read_kernel()? I guess the question then becomes whether the original use for bpf_probe_read() was related to CO-RE. But I think this could also be overcome by adding a fake helper signature in libbpf with a unreasonable high number that is dedicated to probing mem via CO-RE and then libbpf picks the right underlying helper call number for the insn. That avoids fiddling with macros and need for new clang version, no (unless I'm missing something)?

Quoted 5 lines

Libbpf could do it, but I'm a bit worried that unconditionally changing bpf_probe_read() into bpf_probe_read_kernel() is going to be wrong in some cases. If that wasn't the case, why wouldn't we just re-purpose bpf_probe_read() into bpf_probe_read_kernel() in kernel itself, right?

Yes, that is correct, but I mentioned above that this new 'fake' helper call number that libbpf would be fixing up would only be used for bpf_probe_read{,str}() inside bpf_core_read.h.

Small example, bpf_core_read.h would be changed to (just an extract):

tools/lib/bpf/bpf_core_read.h +2 -2
index eae5cccff761..4bddb2ddf3f0 100644
@@ -115,7 +115,7 @@ enum bpf_field_info_kind {
   * (local) BTF, used to record relocation.
   */
  #define bpf_core_read(dst, sz, src)                                        \
-       bpf_probe_read(dst, sz,                                             \
+       bpf_probe_read_selector(dst, sz,                                                    \
                        (const void *)__builtin_preserve_access_index(src))
 
  /*
@@ -124,7 +124,7 @@ enum bpf_field_info_kind {
   * argument.
   */
  #define bpf_core_read_str(dst, sz, src)                                            \
-       bpf_probe_read_str(dst, sz,                                         \
+       bpf_probe_read_str_selector(dst, sz,                                        \
                            (const void *)__builtin_preserve_access_index(src))
 
  #define ___concat(a, b) a ## b
Quoted 7 lines

And bpf_probe_read_{,str_}selector would be defined as e.g. ...

static long (*bpf_probe_read_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -1; static long (*bpf_probe_read_str_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -2;

... where libbpf would do the fix up to either 4 or 45 for insn->imm. But it's still confined to usage in bpf_core_read.h when the CO-RE macros are used.

Ah, I see. Yeah, I suppose that would work as well. Do you prefer me to go this way?

Quoted 31 lines
Quoted 5 lines

But fear not about old Clang support. The bpf_core_type_exists() will use a new built-in, and I'll be able to detect its presence with __has_builtin(X) check in Clang. So it will be completely transparent to users in the end.

Ok.

Quoted 22 lines
Quoted 22 lines
Quoted 21 lines

If that's not an acceptable plan, then one can "parameterize" BPF_CORE_READ macro family by re-defining bpf_core_read() macro. Right now it's defined as:

#define bpf_core_read(dst, sz, src) \
bpf_probe_read(dst, sz, (const void *)__builtin_preserve_access_index(src))

Re-defining it in terms of bpf_probe_read_kernel is trivial, but I can't do it for BPF_CORE_READ, because it will break all the users of bpf_core_read.h that run on older kernels.

Quoted 9 lines
Quoted 8 lines
Quoted 5 lines
    tools/lib/bpf/bpf_core_read.h | 51 ++++++++++++++++++-----------------
    tools/lib/bpf/bpf_tracing.h   | 15 +++++++----
    2 files changed, 37 insertions(+), 29 deletions(-)

[...]

maintainer
#16 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

Quoted 105 lines · Andrii Nakryiko
Quoted 104 lines · Daniel Borkmann
Quoted 66 lines · Andrii Nakryiko
Quoted 54 lines · Daniel Borkmann
Quoted 48 lines · Andrii Nakryiko
Quoted 18 lines · Daniel Borkmann
Quoted 8 lines · Andrii Nakryiko
Quoted 6 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

Quoted 8 lines

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Quoted 3 lines

Could you elaborate what this means wrt dependency on latest clang? Given clang releases have a rather long cadence, what about existing users with current clang releases?

So the overall idea is to use something like this to do kernel reads:

static __always_inline int bpf_probe_read_universal(void *dst, u32 sz, const void *src)

{
       if (bpf_core_type_exists(btf_bpf_probe_read_kernel))
           return bpf_probe_read_kernel(dst, sz, src);
       else
           return bpf_probe_read(dst, sz, src);
}

And then use bpf_probe_read_universal() in BPF_CORE_READ and family.

This approach relies on few things:

1. each BPF helper has a corresponding btf_<helper-name> type defined for it
2. bpf_core_type_exists(some_type) returns 0 or 1, depending if specified type is found in kernel BTF (so needs kernel BTF, of course). This is the part me and Yonghong are working on at the moment.
3. verifier's dead code elimination, which will leave only bpf_probe_read() or bpf_probe_read_kernel() calls and will remove the other one. So on older kernels, there will never be unsupoorted call to bpf_probe_read_kernel().

Quoted 4 lines

The new type existence relocation requires the latest Clang. So the way to deal with older Clangs would be to just fallback to bpf_probe_read, if we detect that Clang is too old and can't emit necessary relocation.

Quoted 10 lines

Okay, seems reasonable overall. One question though: couldn't libbpf transparently fix up the selection of bpf_probe_read() vs bpf_probe_read_kernel()? E.g. it would probe the kernel whether bpf_probe_read_kernel() is available and if it is then it would rewrite the raw call number from the instruction from bpf_probe_read() into the one for bpf_probe_read_kernel()? I guess the question then becomes whether the original use for bpf_probe_read() was related to CO-RE. But I think this could also be overcome by adding a fake helper signature in libbpf with a unreasonable high number that is dedicated to probing mem via CO-RE and then libbpf picks the right underlying helper call number for the insn. That avoids fiddling with macros and need for new clang version, no (unless I'm missing something)?

Quoted 5 lines

Libbpf could do it, but I'm a bit worried that unconditionally changing bpf_probe_read() into bpf_probe_read_kernel() is going to be wrong in some cases. If that wasn't the case, why wouldn't we just re-purpose bpf_probe_read() into bpf_probe_read_kernel() in kernel itself, right?

Yes, that is correct, but I mentioned above that this new 'fake' helper call number that libbpf would be fixing up would only be used for bpf_probe_read{,str}() inside bpf_core_read.h.

Small example, bpf_core_read.h would be changed to (just an extract):

tools/lib/bpf/bpf_core_read.h +2 -2
index eae5cccff761..4bddb2ddf3f0 100644
@@ -115,7 +115,7 @@ enum bpf_field_info_kind {
    * (local) BTF, used to record relocation.
    */
   #define bpf_core_read(dst, sz, src)                                        \
-       bpf_probe_read(dst, sz,                                             \
+       bpf_probe_read_selector(dst, sz,                                                    \
                         (const void *)__builtin_preserve_access_index(src))
 
   /*
@@ -124,7 +124,7 @@ enum bpf_field_info_kind {
    * argument.
    */
   #define bpf_core_read_str(dst, sz, src)                                            \
-       bpf_probe_read_str(dst, sz,                                         \
+       bpf_probe_read_str_selector(dst, sz,                                        \
                             (const void *)__builtin_preserve_access_index(src))
 
   #define ___concat(a, b) a ## b

And bpf_probe_read_{,str_}selector would be defined as e.g. ...

Quoted 8 lines
Quoted 5 lines

static long (*bpf_probe_read_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -1; static long (*bpf_probe_read_str_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -2;

... where libbpf would do the fix up to either 4 or 45 for insn->imm. But it's still confined to usage in bpf_core_read.h when the CO-RE macros are used.

Ah, I see. Yeah, I suppose that would work as well. Do you prefer me to go this way?

I would suggest we should try this path given this can be used with any clang version that has the BPF backend, not just latest upstream git.

Thanks,
Daniel

#17 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

Quoted 110 lines · Daniel Borkmann
Quoted 108 lines · Andrii Nakryiko
Quoted 107 lines · Daniel Borkmann
Quoted 66 lines · Andrii Nakryiko
Quoted 54 lines · Daniel Borkmann
Quoted 48 lines · Andrii Nakryiko
Quoted 18 lines · Daniel Borkmann
Quoted 8 lines · Andrii Nakryiko
Quoted 6 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

Quoted 8 lines

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Quoted 3 lines

Could you elaborate what this means wrt dependency on latest clang? Given clang releases have a rather long cadence, what about existing users with current clang releases?

So the overall idea is to use something like this to do kernel reads:

static __always_inline int bpf_probe_read_universal(void *dst, u32 sz, const void *src)

{
       if (bpf_core_type_exists(btf_bpf_probe_read_kernel))
           return bpf_probe_read_kernel(dst, sz, src);
       else
           return bpf_probe_read(dst, sz, src);
}

And then use bpf_probe_read_universal() in BPF_CORE_READ and family.

This approach relies on few things:

1. each BPF helper has a corresponding btf_<helper-name> type defined for it
2. bpf_core_type_exists(some_type) returns 0 or 1, depending if specified type is found in kernel BTF (so needs kernel BTF, of course). This is the part me and Yonghong are working on at the moment.
3. verifier's dead code elimination, which will leave only bpf_probe_read() or bpf_probe_read_kernel() calls and will remove the other one. So on older kernels, there will never be unsupoorted call to bpf_probe_read_kernel().

Quoted 4 lines

The new type existence relocation requires the latest Clang. So the way to deal with older Clangs would be to just fallback to bpf_probe_read, if we detect that Clang is too old and can't emit necessary relocation.

Quoted 10 lines

Okay, seems reasonable overall. One question though: couldn't libbpf transparently fix up the selection of bpf_probe_read() vs bpf_probe_read_kernel()? E.g. it would probe the kernel whether bpf_probe_read_kernel() is available and if it is then it would rewrite the raw call number from the instruction from bpf_probe_read() into the one for bpf_probe_read_kernel()? I guess the question then becomes whether the original use for bpf_probe_read() was related to CO-RE. But I think this could also be overcome by adding a fake helper signature in libbpf with a unreasonable high number that is dedicated to probing mem via CO-RE and then libbpf picks the right underlying helper call number for the insn. That avoids fiddling with macros and need for new clang version, no (unless I'm missing something)?

Quoted 5 lines

Libbpf could do it, but I'm a bit worried that unconditionally changing bpf_probe_read() into bpf_probe_read_kernel() is going to be wrong in some cases. If that wasn't the case, why wouldn't we just re-purpose bpf_probe_read() into bpf_probe_read_kernel() in kernel itself, right?

Yes, that is correct, but I mentioned above that this new 'fake' helper call number that libbpf would be fixing up would only be used for bpf_probe_read{,str}() inside bpf_core_read.h.

Small example, bpf_core_read.h would be changed to (just an extract):

tools/lib/bpf/bpf_core_read.h +2 -2
index eae5cccff761..4bddb2ddf3f0 100644
@@ -115,7 +115,7 @@ enum bpf_field_info_kind {
    * (local) BTF, used to record relocation.
    */
   #define bpf_core_read(dst, sz, src)                                        \
-       bpf_probe_read(dst, sz,                                             \
+       bpf_probe_read_selector(dst, sz,                                                    \
                         (const void *)__builtin_preserve_access_index(src))
 
   /*
@@ -124,7 +124,7 @@ enum bpf_field_info_kind {
    * argument.
    */
   #define bpf_core_read_str(dst, sz, src)                                            \
-       bpf_probe_read_str(dst, sz,                                         \
+       bpf_probe_read_str_selector(dst, sz,                                        \
                             (const void *)__builtin_preserve_access_index(src))
 
   #define ___concat(a, b) a ## b

And bpf_probe_read_{,str_}selector would be defined as e.g. ...

static long (*bpf_probe_read_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -1; static long (*bpf_probe_read_str_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -2;

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

... where libbpf would do the fix up to either 4 or 45 for insn->imm. But it's still confined to usage in bpf_core_read.h when the CO-RE macros are used.

Ah, I see. Yeah, I suppose that would work as well. Do you prefer me to go this way?

I would suggest we should try this path given this can be used with any clang version that has the BPF backend, not just latest upstream git.

Sure, sounds good.

Quoted 3 lines

Thanks,
Daniel

#18 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

In reply to Daniel Borkmann
Quoted 110 lines · Daniel Borkmann
Quoted 108 lines · Andrii Nakryiko
Quoted 107 lines · Daniel Borkmann
Quoted 66 lines · Andrii Nakryiko
Quoted 54 lines · Daniel Borkmann
Quoted 48 lines · Andrii Nakryiko
Quoted 18 lines · Daniel Borkmann
Quoted 8 lines · Andrii Nakryiko
Quoted 6 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

Quoted 8 lines

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Quoted 3 lines

Could you elaborate what this means wrt dependency on latest clang? Given clang releases have a rather long cadence, what about existing users with current clang releases?

So the overall idea is to use something like this to do kernel reads:

static __always_inline int bpf_probe_read_universal(void *dst, u32 sz, const void *src)

{
       if (bpf_core_type_exists(btf_bpf_probe_read_kernel))
           return bpf_probe_read_kernel(dst, sz, src);
       else
           return bpf_probe_read(dst, sz, src);
}

And then use bpf_probe_read_universal() in BPF_CORE_READ and family.

This approach relies on few things:

1. each BPF helper has a corresponding btf_<helper-name> type defined for it
2. bpf_core_type_exists(some_type) returns 0 or 1, depending if specified type is found in kernel BTF (so needs kernel BTF, of course). This is the part me and Yonghong are working on at the moment.
3. verifier's dead code elimination, which will leave only bpf_probe_read() or bpf_probe_read_kernel() calls and will remove the other one. So on older kernels, there will never be unsupoorted call to bpf_probe_read_kernel().

Quoted 4 lines

The new type existence relocation requires the latest Clang. So the way to deal with older Clangs would be to just fallback to bpf_probe_read, if we detect that Clang is too old and can't emit necessary relocation.

Quoted 10 lines

Okay, seems reasonable overall. One question though: couldn't libbpf transparently fix up the selection of bpf_probe_read() vs bpf_probe_read_kernel()? E.g. it would probe the kernel whether bpf_probe_read_kernel() is available and if it is then it would rewrite the raw call number from the instruction from bpf_probe_read() into the one for bpf_probe_read_kernel()? I guess the question then becomes whether the original use for bpf_probe_read() was related to CO-RE. But I think this could also be overcome by adding a fake helper signature in libbpf with a unreasonable high number that is dedicated to probing mem via CO-RE and then libbpf picks the right underlying helper call number for the insn. That avoids fiddling with macros and need for new clang version, no (unless I'm missing something)?

Quoted 5 lines

Libbpf could do it, but I'm a bit worried that unconditionally changing bpf_probe_read() into bpf_probe_read_kernel() is going to be wrong in some cases. If that wasn't the case, why wouldn't we just re-purpose bpf_probe_read() into bpf_probe_read_kernel() in kernel itself, right?

Yes, that is correct, but I mentioned above that this new 'fake' helper call number that libbpf would be fixing up would only be used for bpf_probe_read{,str}() inside bpf_core_read.h.

Small example, bpf_core_read.h would be changed to (just an extract):

tools/lib/bpf/bpf_core_read.h +2 -2
index eae5cccff761..4bddb2ddf3f0 100644
@@ -115,7 +115,7 @@ enum bpf_field_info_kind {
    * (local) BTF, used to record relocation.
    */
   #define bpf_core_read(dst, sz, src)                                        \
-       bpf_probe_read(dst, sz,                                             \
+       bpf_probe_read_selector(dst, sz,                                                    \
                         (const void *)__builtin_preserve_access_index(src))
 
   /*
@@ -124,7 +124,7 @@ enum bpf_field_info_kind {
    * argument.
    */
   #define bpf_core_read_str(dst, sz, src)                                            \
-       bpf_probe_read_str(dst, sz,                                         \
+       bpf_probe_read_str_selector(dst, sz,                                        \
                             (const void *)__builtin_preserve_access_index(src))
 
   #define ___concat(a, b) a ## b

And bpf_probe_read_{,str_}selector would be defined as e.g. ...

static long (*bpf_probe_read_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -1; static long (*bpf_probe_read_str_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -2;

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

... where libbpf would do the fix up to either 4 or 45 for insn->imm. But it's still confined to usage in bpf_core_read.h when the CO-RE macros are used.

Ah, I see. Yeah, I suppose that would work as well. Do you prefer me to go this way?

I would suggest we should try this path given this can be used with any clang version that has the BPF backend, not just latest upstream git.

I have an even better solution, I think. Convert everything to bpf_probe_read_kernel() or bpf_probe_read_user() unconditionally, but let libbpf switch those two to bpf_probe_read() if _kernel()/_user() variants are not yet in the kernel. That should handle both CO-RE helpers and just pretty much any use case that was converted.

Quoted 3 lines

Thanks,
Daniel

maintainer
#19 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

Quoted 116 lines · Andrii Nakryiko
Quoted 115 lines · Daniel Borkmann
Quoted 113 lines · Andrii Nakryiko
Quoted 102 lines · Daniel Borkmann
Quoted 66 lines · Andrii Nakryiko
Quoted 54 lines · Daniel Borkmann
Quoted 48 lines · Andrii Nakryiko
Quoted 18 lines · Daniel Borkmann
Quoted 8 lines · Andrii Nakryiko
Quoted 6 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

Quoted 8 lines

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Quoted 3 lines

Could you elaborate what this means wrt dependency on latest clang? Given clang releases have a rather long cadence, what about existing users with current clang releases?

So the overall idea is to use something like this to do kernel reads:

static __always_inline int bpf_probe_read_universal(void *dst, u32 sz, const void *src)

{
        if (bpf_core_type_exists(btf_bpf_probe_read_kernel))
            return bpf_probe_read_kernel(dst, sz, src);
        else
            return bpf_probe_read(dst, sz, src);
}

And then use bpf_probe_read_universal() in BPF_CORE_READ and family.

This approach relies on few things:

1. each BPF helper has a corresponding btf_<helper-name> type defined for it
2. bpf_core_type_exists(some_type) returns 0 or 1, depending if specified type is found in kernel BTF (so needs kernel BTF, of course). This is the part me and Yonghong are working on at the moment.
3. verifier's dead code elimination, which will leave only bpf_probe_read() or bpf_probe_read_kernel() calls and will remove the other one. So on older kernels, there will never be unsupoorted call to bpf_probe_read_kernel().

Quoted 4 lines

The new type existence relocation requires the latest Clang. So the way to deal with older Clangs would be to just fallback to bpf_probe_read, if we detect that Clang is too old and can't emit necessary relocation.

Quoted 10 lines

Okay, seems reasonable overall. One question though: couldn't libbpf transparently fix up the selection of bpf_probe_read() vs bpf_probe_read_kernel()? E.g. it would probe the kernel whether bpf_probe_read_kernel() is available and if it is then it would rewrite the raw call number from the instruction from bpf_probe_read() into the one for bpf_probe_read_kernel()? I guess the question then becomes whether the original use for bpf_probe_read() was related to CO-RE. But I think this could also be overcome by adding a fake helper signature in libbpf with a unreasonable high number that is dedicated to probing mem via CO-RE and then libbpf picks the right underlying helper call number for the insn. That avoids fiddling with macros and need for new clang version, no (unless I'm missing something)?

Quoted 5 lines

Libbpf could do it, but I'm a bit worried that unconditionally changing bpf_probe_read() into bpf_probe_read_kernel() is going to be wrong in some cases. If that wasn't the case, why wouldn't we just re-purpose bpf_probe_read() into bpf_probe_read_kernel() in kernel itself, right?

Yes, that is correct, but I mentioned above that this new 'fake' helper call number that libbpf would be fixing up would only be used for bpf_probe_read{,str}() inside bpf_core_read.h.

Small example, bpf_core_read.h would be changed to (just an extract):

tools/lib/bpf/bpf_core_read.h +2 -2
index eae5cccff761..4bddb2ddf3f0 100644
@@ -115,7 +115,7 @@ enum bpf_field_info_kind {
     * (local) BTF, used to record relocation.
     */
    #define bpf_core_read(dst, sz, src)                                        \
-       bpf_probe_read(dst, sz,                                             \
+       bpf_probe_read_selector(dst, sz,                                                    \
                          (const void *)__builtin_preserve_access_index(src))
 
    /*
@@ -124,7 +124,7 @@ enum bpf_field_info_kind {
     * argument.
     */
    #define bpf_core_read_str(dst, sz, src)                                            \
-       bpf_probe_read_str(dst, sz,                                         \
+       bpf_probe_read_str_selector(dst, sz,                                        \
                              (const void *)__builtin_preserve_access_index(src))
 
    #define ___concat(a, b) a ## b
Quoted 7 lines

And bpf_probe_read_{,str_}selector would be defined as e.g. ...

static long (*bpf_probe_read_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -1; static long (*bpf_probe_read_str_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -2;

... where libbpf would do the fix up to either 4 or 45 for insn->imm. But it's still confined to usage in bpf_core_read.h when the CO-RE macros are used.

Ah, I see. Yeah, I suppose that would work as well. Do you prefer me to go this way?

Quoted 8 lines
Quoted 2 lines

I would suggest we should try this path given this can be used with any clang version that has the BPF backend, not just latest upstream git.

I have an even better solution, I think. Convert everything to bpf_probe_read_kernel() or bpf_probe_read_user() unconditionally, but let libbpf switch those two to bpf_probe_read() if _kernel()/_user() variants are not yet in the kernel. That should handle both CO-RE helpers and just pretty much any use case that was converted.

Yes, agree, that is an even cleaner solution and avoids to 'pollute' the helper ID space with such remapping. The user intent with bpf_probe_read_kernel() or bpf_probe_read_user() is rather clear so bpf_probe_read() can be a fallback for this direction. Lets go with that.

Thanks,
Daniel

#20 lore raw

Re: [PATCH bpf-next 3/3] libbpf: Use bpf_probe_read_kernel

Quoted 127 lines · Daniel Borkmann
Quoted 116 lines · Andrii Nakryiko
Quoted 115 lines · Daniel Borkmann
Quoted 113 lines · Andrii Nakryiko
Quoted 102 lines · Daniel Borkmann
Quoted 66 lines · Andrii Nakryiko
Quoted 54 lines · Daniel Borkmann
Quoted 48 lines · Andrii Nakryiko
Quoted 18 lines · Daniel Borkmann
Quoted 8 lines · Andrii Nakryiko
Quoted 6 lines · Ilya Leoshkevich

Yet another adaptation to commit 0ebeea8ca8a4 ("bpf: Restrict bpf_probe_read{, str}() only to archs where they work") that makes more samples compile on s390.

Quoted 8 lines

Sorry, we can't do this yet. This will break on older kernels that don't yet have bpf_probe_read_kernel() implemented. Met and Yonghong are working on extending a set of CO-RE relocations, that would allow to do bpf_probe_read_kernel() detection on BPF side, transparently for an application, and will pick either bpf_probe_read() or bpf_probe_read_kernel(). It should be ready soon (this or next week, most probably), though it will have dependency on the latest Clang. But for now, please don't change this.

Quoted 3 lines

Could you elaborate what this means wrt dependency on latest clang? Given clang releases have a rather long cadence, what about existing users with current clang releases?

So the overall idea is to use something like this to do kernel reads:

static __always_inline int bpf_probe_read_universal(void *dst, u32 sz, const void *src)

{
        if (bpf_core_type_exists(btf_bpf_probe_read_kernel))
            return bpf_probe_read_kernel(dst, sz, src);
        else
            return bpf_probe_read(dst, sz, src);
}

And then use bpf_probe_read_universal() in BPF_CORE_READ and family.

This approach relies on few things:

1. each BPF helper has a corresponding btf_<helper-name> type defined for it
2. bpf_core_type_exists(some_type) returns 0 or 1, depending if specified type is found in kernel BTF (so needs kernel BTF, of course). This is the part me and Yonghong are working on at the moment.
3. verifier's dead code elimination, which will leave only bpf_probe_read() or bpf_probe_read_kernel() calls and will remove the other one. So on older kernels, there will never be unsupoorted call to bpf_probe_read_kernel().

Quoted 4 lines

The new type existence relocation requires the latest Clang. So the way to deal with older Clangs would be to just fallback to bpf_probe_read, if we detect that Clang is too old and can't emit necessary relocation.

Quoted 10 lines

Okay, seems reasonable overall. One question though: couldn't libbpf transparently fix up the selection of bpf_probe_read() vs bpf_probe_read_kernel()? E.g. it would probe the kernel whether bpf_probe_read_kernel() is available and if it is then it would rewrite the raw call number from the instruction from bpf_probe_read() into the one for bpf_probe_read_kernel()? I guess the question then becomes whether the original use for bpf_probe_read() was related to CO-RE. But I think this could also be overcome by adding a fake helper signature in libbpf with a unreasonable high number that is dedicated to probing mem via CO-RE and then libbpf picks the right underlying helper call number for the insn. That avoids fiddling with macros and need for new clang version, no (unless I'm missing something)?

Quoted 5 lines

Libbpf could do it, but I'm a bit worried that unconditionally changing bpf_probe_read() into bpf_probe_read_kernel() is going to be wrong in some cases. If that wasn't the case, why wouldn't we just re-purpose bpf_probe_read() into bpf_probe_read_kernel() in kernel itself, right?

Yes, that is correct, but I mentioned above that this new 'fake' helper call number that libbpf would be fixing up would only be used for bpf_probe_read{,str}() inside bpf_core_read.h.

Small example, bpf_core_read.h would be changed to (just an extract):

tools/lib/bpf/bpf_core_read.h +2 -2
index eae5cccff761..4bddb2ddf3f0 100644
@@ -115,7 +115,7 @@ enum bpf_field_info_kind {
     * (local) BTF, used to record relocation.
     */
    #define bpf_core_read(dst, sz, src)                                        \
-       bpf_probe_read(dst, sz,                                             \
+       bpf_probe_read_selector(dst, sz,                                                    \
                          (const void *)__builtin_preserve_access_index(src))
 
    /*
@@ -124,7 +124,7 @@ enum bpf_field_info_kind {
     * argument.
     */
    #define bpf_core_read_str(dst, sz, src)                                            \
-       bpf_probe_read_str(dst, sz,                                         \
+       bpf_probe_read_str_selector(dst, sz,                                        \
                              (const void *)__builtin_preserve_access_index(src))
 
    #define ___concat(a, b) a ## b
Quoted 7 lines

And bpf_probe_read_{,str_}selector would be defined as e.g. ...

static long (*bpf_probe_read_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -1; static long (*bpf_probe_read_str_selector)(void *dst, __u32 size, const void *unsafe_ptr) = (void *) -2;

... where libbpf would do the fix up to either 4 or 45 for insn->imm. But it's still confined to usage in bpf_core_read.h when the CO-RE macros are used.

Ah, I see. Yeah, I suppose that would work as well. Do you prefer me to go this way?

Quoted 8 lines
Quoted 2 lines

I would suggest we should try this path given this can be used with any clang version that has the BPF backend, not just latest upstream git.

I have an even better solution, I think. Convert everything to bpf_probe_read_kernel() or bpf_probe_read_user() unconditionally, but let libbpf switch those two to bpf_probe_read() if _kernel()/_user() variants are not yet in the kernel. That should handle both CO-RE helpers and just pretty much any use case that was converted.

Quoted 5 lines

Yes, agree, that is an even cleaner solution and avoids to 'pollute' the helper ID space with such remapping. The user intent with bpf_probe_read_kernel() or bpf_probe_read_user() is rather clear so bpf_probe_read() can be a fallback for this direction. Lets go with that.

Ok, I have all this working locally. I'll post patches once bpf-next re-opens.

Quoted 2 lines

Thanks,
Daniel