[PATCH bpf-next] selftests/bpf: Fix building error when using userspace pt_regs

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

Signed-off-by: Pu Lehui <pulehui@huawei.com>
---
 tools/testing/selftests/bpf/progs/loop1.c     |  8 ++------
 tools/testing/selftests/bpf/progs/loop2.c     |  8 ++------
 tools/testing/selftests/bpf/progs/loop3.c     |  8 ++------
 tools/testing/selftests/bpf/progs/loop6.c     | 20 ++++++-------------
 .../selftests/bpf/progs/test_overhead.c       |  8 ++------
 .../selftests/bpf/progs/test_probe_user.c     |  6 +-----
 6 files changed, 15 insertions(+), 43 deletions(-)

diff --git a/tools/testing/selftests/bpf/progs/loop1.c b/tools/testing/selftests/bpf/progs/loop1.c
index 50e66772c046..ea04c035719c 100644
--- a/tools/testing/selftests/bpf/progs/loop1.c
+++ b/tools/testing/selftests/bpf/progs/loop1.c

@@ -1,11 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 // Copyright (c) 2019 Facebook
-#include <linux/sched.h>
-#include <linux/ptrace.h>
-#include <stdint.h>
-#include <stddef.h>
-#include <stdbool.h>
-#include <linux/bpf.h>
+
+#include "vmlinux.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 

diff --git a/tools/testing/selftests/bpf/progs/loop2.c b/tools/testing/selftests/bpf/progs/loop2.c
index 947bb7e988c2..edf07b1d310e 100644
--- a/tools/testing/selftests/bpf/progs/loop2.c
+++ b/tools/testing/selftests/bpf/progs/loop2.c

@@ -1,11 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 // Copyright (c) 2019 Facebook
-#include <linux/sched.h>
-#include <linux/ptrace.h>
-#include <stdint.h>
-#include <stddef.h>
-#include <stdbool.h>
-#include <linux/bpf.h>
+
+#include "vmlinux.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 

diff --git a/tools/testing/selftests/bpf/progs/loop3.c b/tools/testing/selftests/bpf/progs/loop3.c
index 76e93b31c14b..c8d2f2c55547 100644
--- a/tools/testing/selftests/bpf/progs/loop3.c
+++ b/tools/testing/selftests/bpf/progs/loop3.c

@@ -1,11 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 // Copyright (c) 2019 Facebook
-#include <linux/sched.h>
-#include <linux/ptrace.h>
-#include <stdint.h>
-#include <stddef.h>
-#include <stdbool.h>
-#include <linux/bpf.h>
+
+#include "vmlinux.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 

diff --git a/tools/testing/selftests/bpf/progs/loop6.c b/tools/testing/selftests/bpf/progs/loop6.c
index 38de0331e6b4..17ac4da5664d 100644
--- a/tools/testing/selftests/bpf/progs/loop6.c
+++ b/tools/testing/selftests/bpf/progs/loop6.c

@@ -1,8 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0
 
-#include <linux/ptrace.h>
-#include <stddef.h>
-#include <linux/bpf.h>
+#include "vmlinux.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 

@@ -25,12 +23,6 @@ char _license[] SEC("license") = "GPL";
 #define SG_CHAIN	0x01UL
 #define SG_END		0x02UL
 
-struct scatterlist {
-	unsigned long   page_link;
-	unsigned int    offset;
-	unsigned int    length;
-};
-
 #define sg_is_chain(sg)		((sg)->page_link & SG_CHAIN)
 #define sg_is_last(sg)		((sg)->page_link & SG_END)
 #define sg_chain_ptr(sg)	\

@@ -61,8 +53,8 @@ static inline struct scatterlist *get_sgp(struct scatterlist **sgs, int i)
 	return sgp;
 }
 
-int config = 0;
-int result = 0;
+int g_config = 0;
+int g_result = 0;
 
 SEC("kprobe/virtqueue_add_sgs")
 int BPF_KPROBE(trace_virtqueue_add_sgs, void *unused, struct scatterlist **sgs,

@@ -72,7 +64,7 @@ int BPF_KPROBE(trace_virtqueue_add_sgs, void *unused, struct scatterlist **sgs,
 	__u64 length1 = 0, length2 = 0;
 	unsigned int i, n, len;
 
-	if (config != 0)
+	if (g_config != 0)
 		return 0;
 
 	for (i = 0; (i < VIRTIO_MAX_SGS) && (i < out_sgs); i++) {

@@ -93,7 +85,7 @@ int BPF_KPROBE(trace_virtqueue_add_sgs, void *unused, struct scatterlist **sgs,
 		}
 	}
 
-	config = 1;
-	result = length2 - length1;
+	g_config = 1;
+	g_result = length2 - length1;
 	return 0;
 }

diff --git a/tools/testing/selftests/bpf/progs/test_overhead.c b/tools/testing/selftests/bpf/progs/test_overhead.c
index abb7344b531f..df035e6a3143 100644
--- a/tools/testing/selftests/bpf/progs/test_overhead.c
+++ b/tools/testing/selftests/bpf/progs/test_overhead.c

@@ -1,14 +1,10 @@
 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2019 Facebook */
-#include <stdbool.h>
-#include <stddef.h>
-#include <linux/bpf.h>
-#include <linux/ptrace.h>
+
+#include "vmlinux.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 
-struct task_struct;
-
 SEC("kprobe/__set_task_comm")
 int BPF_KPROBE(prog1, struct task_struct *tsk, const char *buf, bool exec)
 {

diff --git a/tools/testing/selftests/bpf/progs/test_probe_user.c b/tools/testing/selftests/bpf/progs/test_probe_user.c
index 8812a90da4eb..bf6c0b864ace 100644
--- a/tools/testing/selftests/bpf/progs/test_probe_user.c
+++ b/tools/testing/selftests/bpf/progs/test_probe_user.c

@@ -1,10 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0
 
-#include <linux/ptrace.h>
-#include <linux/bpf.h>
-
-#include <netinet/in.h>
-
+#include "vmlinux.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 

-- 
2.25.1

On 12/14/21 2:55 PM, Pu Lehui wrote:

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

Signed-off-by: Pu Lehui <pulehui@huawei.com>

Looks like this lets CI fail, did you run the selftests also with vmtest.sh to
double check?

https://github.com/kernel-patches/bpf/runs/4521708490?check_suite_focus=true :

[...]
#189 verif_scale_loop6:FAIL
libbpf: prog 'trace_virtqueue_add_sgs': BPF program load failed: Argument list too long
libbpf: prog 'trace_virtqueue_add_sgs': -- BEGIN PROG LOAD LOG --
R1 type=ctx expected=fp
BPF program is too large. Processed 1000001 insn
verification time 12250995 usec
stack depth 88
processed 1000001 insns (limit 1000000) max_states_per_insn 107 total_states 21739 peak_states 2271 mark_read 6
-- END PROG LOAD LOG --
libbpf: failed to load program 'trace_virtqueue_add_sgs'
libbpf: failed to load object 'loop6.o'
scale_test:FAIL:expect_success unexpected error: -7 (errno 7)
Summary: 221/986 PASSED, 8 SKIPPED, 1 FAILED
[...]

Please take a look and fix in your patch, thanks!

On 2021/12/15 4:01, Daniel Borkmann wrote:

On 12/14/21 2:55 PM, Pu Lehui wrote:

quoted

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

Signed-off-by: Pu Lehui <pulehui@huawei.com>

Looks like this lets CI fail, did you run the selftests also with vmtest.sh to
double check?

https://github.com/kernel-patches/bpf/runs/4521708490?check_suite_focus=true :

[...]
#189 verif_scale_loop6:FAIL
libbpf: prog 'trace_virtqueue_add_sgs': BPF program load failed: Argument list too long
libbpf: prog 'trace_virtqueue_add_sgs': -- BEGIN PROG LOAD LOG --
R1 type=ctx expected=fp
BPF program is too large. Processed 1000001 insn
verification time 12250995 usec
stack depth 88
processed 1000001 insns (limit 1000000) max_states_per_insn 107 total_states 21739 peak_states 2271 mark_read 6
-- END PROG LOAD LOG --
libbpf: failed to load program 'trace_virtqueue_add_sgs'
libbpf: failed to load object 'loop6.o'
scale_test:FAIL:expect_success unexpected error: -7 (errno 7)
Summary: 221/986 PASSED, 8 SKIPPED, 1 FAILED
[...]

Please take a look and fix in your patch, thanks!
.

Sorry for my negligence, I'll take a look and fix it.

On Tue, Dec 14, 2021 at 5:54 AM Pu Lehui [off-list ref] wrote:

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

We could probably also extend bpf_tracing.h to work with
kernel-defined pt_regs, just like we do for x86 (see __KERNEL__ and
__VMLINUX_H__ checks). It's more work, but will benefit other end
users, not just selftests.

Signed-off-by: Pu Lehui <pulehui@huawei.com>
---
 tools/testing/selftests/bpf/progs/loop1.c     |  8 ++------
 tools/testing/selftests/bpf/progs/loop2.c     |  8 ++------
 tools/testing/selftests/bpf/progs/loop3.c     |  8 ++------
 tools/testing/selftests/bpf/progs/loop6.c     | 20 ++++++-------------
 .../selftests/bpf/progs/test_overhead.c       |  8 ++------
 .../selftests/bpf/progs/test_probe_user.c     |  6 +-----
 6 files changed, 15 insertions(+), 43 deletions(-)

[...]

On 2021/12/15 9:20, Pu Lehui wrote:

On 2021/12/15 4:01, Daniel Borkmann wrote:

quoted

On 12/14/21 2:55 PM, Pu Lehui wrote:

quoted

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

Signed-off-by: Pu Lehui <pulehui@huawei.com>

Looks like this lets CI fail, did you run the selftests also with vmtest.sh to
double check?

https://github.com/kernel-patches/bpf/runs/4521708490?check_suite_focus=true :

[...]
#189 verif_scale_loop6:FAIL
libbpf: prog 'trace_virtqueue_add_sgs': BPF program load failed: Argument list too long
libbpf: prog 'trace_virtqueue_add_sgs': -- BEGIN PROG LOAD LOG --
R1 type=ctx expected=fp
BPF program is too large. Processed 1000001 insn
verification time 12250995 usec
stack depth 88
processed 1000001 insns (limit 1000000) max_states_per_insn 107 total_states 21739 peak_states 2271 mark_read 6
-- END PROG LOAD LOG --
libbpf: failed to load program 'trace_virtqueue_add_sgs'
libbpf: failed to load object 'loop6.o'
scale_test:FAIL:expect_success unexpected error: -7 (errno 7)
Summary: 221/986 PASSED, 8 SKIPPED, 1 FAILED
[...]

Please take a look and fix in your patch, thanks!
.

Sorry for my negligence, I'll take a look and fix it.
.

It seems strange that verifier think the loop can execute up to u64_max while I just replace the header file.
This looks very similar to the previous llvm issue, https://github.com/iovisor/bcc/pull/3270, but I have no idea how to locate.

Back to arm64 bpf selftest compiling problem, we can use header file directory generated by "make headers_install" to fix it.

--- a/tools/testing/selftests/bpf/Makefile
+++ b/tools/testing/selftests/bpf/Makefile

@@ -294,7 +294,8 @@ MENDIAN=$(if $(IS_LITTLE_ENDIAN),-mlittle-endian,-mbig-endian)
 CLANG_SYS_INCLUDES = $(call get_sys_includes,$(CLANG))
 BPF_CFLAGS = -g -D__TARGET_ARCH_$(SRCARCH) $(MENDIAN) \
 	     -I$(INCLUDE_DIR) -I$(CURDIR) -I$(APIDIR) \
-	     -I$(abspath $(OUTPUT)/../usr/include)
+	     -I$(abspath $(OUTPUT)/../usr/include) \
+	     -I../../../../usr/include

On 2021/12/16 12:06, Andrii Nakryiko wrote:

On Tue, Dec 14, 2021 at 5:54 AM Pu Lehui [off-list ref] wrote:

quoted

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

We could probably also extend bpf_tracing.h to work with
kernel-defined pt_regs, just like we do for x86 (see __KERNEL__ and
__VMLINUX_H__ checks). It's more work, but will benefit other end
users, not just selftests.

It might change a lot. We can use header file directory generated by "make headers_install" to fix it.

--- a/tools/testing/selftests/bpf/Makefile
+++ b/tools/testing/selftests/bpf/Makefile

@@ -294,7 +294,8 @@ MENDIAN=$(if $(IS_LITTLE_ENDIAN),-mlittle-endian,-mbig-endian)
 CLANG_SYS_INCLUDES = $(call get_sys_includes,$(CLANG))
 BPF_CFLAGS = -g -D__TARGET_ARCH_$(SRCARCH) $(MENDIAN) \
 	     -I$(INCLUDE_DIR) -I$(CURDIR) -I$(APIDIR) \
-	     -I$(abspath $(OUTPUT)/../usr/include)
+	     -I$(abspath $(OUTPUT)/../usr/include) \
+	     -I../../../../usr/include

quoted

Signed-off-by: Pu Lehui <pulehui@huawei.com>
---
  tools/testing/selftests/bpf/progs/loop1.c     |  8 ++------
  tools/testing/selftests/bpf/progs/loop2.c     |  8 ++------
  tools/testing/selftests/bpf/progs/loop3.c     |  8 ++------
  tools/testing/selftests/bpf/progs/loop6.c     | 20 ++++++-------------
  .../selftests/bpf/progs/test_overhead.c       |  8 ++------
  .../selftests/bpf/progs/test_probe_user.c     |  6 +-----
  6 files changed, 15 insertions(+), 43 deletions(-)

[...]
.

On Thu, Dec 16, 2021 at 6:25 PM Pu Lehui [off-list ref] wrote:

On 2021/12/16 12:06, Andrii Nakryiko wrote:

quoted

On Tue, Dec 14, 2021 at 5:54 AM Pu Lehui [off-list ref] wrote:

quoted

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

We could probably also extend bpf_tracing.h to work with
kernel-defined pt_regs, just like we do for x86 (see __KERNEL__ and
__VMLINUX_H__ checks). It's more work, but will benefit other end
users, not just selftests.

It might change a lot. We can use header file directory generated by
"make headers_install" to fix it.

We don't have dependency on "make headers_install" and I'd rather not add it.

What do you mean by "change a lot"?

On 2021/12/18 0:45, Andrii Nakryiko wrote:

On Thu, Dec 16, 2021 at 6:25 PM Pu Lehui [off-list ref] wrote:

quoted

On 2021/12/16 12:06, Andrii Nakryiko wrote:

quoted

On Tue, Dec 14, 2021 at 5:54 AM Pu Lehui [off-list ref] wrote:

quoted

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

We could probably also extend bpf_tracing.h to work with
kernel-defined pt_regs, just like we do for x86 (see __KERNEL__ and
__VMLINUX_H__ checks). It's more work, but will benefit other end
users, not just selftests.

It might change a lot. We can use header file directory generated by
"make headers_install" to fix it.

We don't have dependency on "make headers_install" and I'd rather not add it.

What do you mean by "change a lot"?

Maybe I misunderstood your advice. Your suggestion might be to extend bpf_tracing.h to kernel-space pt_regs, while some archs, like arm64, only support user-space. So the patch might be like this:

diff --git a/tools/lib/bpf/bpf_tracing.h b/tools/lib/bpf/bpf_tracing.h
index db05a5937105..2c3cb8e9ae92 100644
--- a/tools/lib/bpf/bpf_tracing.h
+++ b/tools/lib/bpf/bpf_tracing.h

@@ -195,9 +195,13 @@ struct pt_regs;

 #elif defined(bpf_target_arm64)

-struct pt_regs;
+#if defined(__KERNEL__)
+#define PT_REGS_ARM64 const volatile struct pt_regs
+#else
 /* arm64 provides struct user_pt_regs instead of struct pt_regs to userspace */
 #define PT_REGS_ARM64 const volatile struct user_pt_regs
+#endif
+
 #define PT_REGS_PARM1(x) (((PT_REGS_ARM64 *)(x))->regs[0])
 #define PT_REGS_PARM2(x) (((PT_REGS_ARM64 *)(x))->regs[1])
 #define PT_REGS_PARM3(x) (((PT_REGS_ARM64 *)(x))->regs[2])

quoted

--- a/tools/testing/selftests/bpf/Makefile
+++ b/tools/testing/selftests/bpf/Makefile

@@ -294,7 +294,8 @@ MENDIAN=$(if

$(IS_LITTLE_ENDIAN),-mlittle-endian,-mbig-endian)
   CLANG_SYS_INCLUDES = $(call get_sys_includes,$(CLANG))
   BPF_CFLAGS = -g -D__TARGET_ARCH_$(SRCARCH) $(MENDIAN) \
              -I$(INCLUDE_DIR) -I$(CURDIR) -I$(APIDIR) \
-            -I$(abspath $(OUTPUT)/../usr/include)
+            -I$(abspath $(OUTPUT)/../usr/include) \
+            -I../../../../usr/include

quoted

quoted

Signed-off-by: Pu Lehui <pulehui@huawei.com>
---
   tools/testing/selftests/bpf/progs/loop1.c     |  8 ++------
   tools/testing/selftests/bpf/progs/loop2.c     |  8 ++------
   tools/testing/selftests/bpf/progs/loop3.c     |  8 ++------
   tools/testing/selftests/bpf/progs/loop6.c     | 20 ++++++-------------
   .../selftests/bpf/progs/test_overhead.c       |  8 ++------
   .../selftests/bpf/progs/test_probe_user.c     |  6 +-----
   6 files changed, 15 insertions(+), 43 deletions(-)

[...]
.

.

On 2021/12/20 22:02, Pu Lehui wrote:

Please ignore the last reply. User-space pt_regs of arm64/s390 is the first part of the kernel-space's, it should has covered both kernel and userspace.

quoted

quoted

--- a/tools/testing/selftests/bpf/Makefile
+++ b/tools/testing/selftests/bpf/Makefile

@@ -294,7 +294,8 @@ MENDIAN=$(if

$(IS_LITTLE_ENDIAN),-mlittle-endian,-mbig-endian)
   CLANG_SYS_INCLUDES = $(call get_sys_includes,$(CLANG))
   BPF_CFLAGS = -g -D__TARGET_ARCH_$(SRCARCH) $(MENDIAN) \
              -I$(INCLUDE_DIR) -I$(CURDIR) -I$(APIDIR) \
-            -I$(abspath $(OUTPUT)/../usr/include)
+            -I$(abspath $(OUTPUT)/../usr/include) \
+            -I../../../../usr/include

quoted

quoted

Signed-off-by: Pu Lehui <pulehui@huawei.com>
---
   tools/testing/selftests/bpf/progs/loop1.c     |  8 ++------
   tools/testing/selftests/bpf/progs/loop2.c     |  8 ++------
   tools/testing/selftests/bpf/progs/loop3.c     |  8 ++------
   tools/testing/selftests/bpf/progs/loop6.c     | 20 ++++++-------------
   .../selftests/bpf/progs/test_overhead.c       |  8 ++------
   .../selftests/bpf/progs/test_probe_user.c     |  6 +-----
   6 files changed, 15 insertions(+), 43 deletions(-)

[...]
.

.

.

On Mon, Dec 20, 2021 at 4:58 PM Pu Lehui [off-list ref] wrote:

On 2021/12/20 22:02, Pu Lehui wrote:

quoted

On 2021/12/18 0:45, Andrii Nakryiko wrote:

quoted

On Thu, Dec 16, 2021 at 6:25 PM Pu Lehui [off-list ref] wrote:

quoted

On 2021/12/16 12:06, Andrii Nakryiko wrote:

quoted

On Tue, Dec 14, 2021 at 5:54 AM Pu Lehui [off-list ref] wrote:

quoted

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

We could probably also extend bpf_tracing.h to work with
kernel-defined pt_regs, just like we do for x86 (see __KERNEL__ and
__VMLINUX_H__ checks). It's more work, but will benefit other end
users, not just selftests.

It might change a lot. We can use header file directory generated by
"make headers_install" to fix it.

We don't have dependency on "make headers_install" and I'd rather not
add it.

What do you mean by "change a lot"?

Maybe I misunderstood your advice. Your suggestion might be to extend
bpf_tracing.h to kernel-space pt_regs, while some archs, like arm64,

yes

quoted

only support user-space. So the patch might be like this:

diff --git a/tools/lib/bpf/bpf_tracing.h b/tools/lib/bpf/bpf_tracing.h
index db05a5937105..2c3cb8e9ae92 100644
--- a/tools/lib/bpf/bpf_tracing.h
+++ b/tools/lib/bpf/bpf_tracing.h

@@ -195,9 +195,13 @@ struct pt_regs;

  #elif defined(bpf_target_arm64)

-struct pt_regs;
+#if defined(__KERNEL__)
+#define PT_REGS_ARM64 const volatile struct pt_regs
+#else
  /* arm64 provides struct user_pt_regs instead of struct pt_regs to

userspace */
  #define PT_REGS_ARM64 const volatile struct user_pt_regs
+#endif
+
  #define PT_REGS_PARM1(x) (((PT_REGS_ARM64 *)(x))->regs[0])
  #define PT_REGS_PARM2(x) (((PT_REGS_ARM64 *)(x))->regs[1])
  #define PT_REGS_PARM3(x) (((PT_REGS_ARM64 *)(x))->regs[2])

Please ignore the last reply. User-space pt_regs of arm64/s390 is the
first part of the kernel-space's, it should has covered both kernel and
userspace.

Alright, so is there still a problem or not? Looking at the definition
of struct pt_regs for arm64, just casting struct pt_regs to struct
user_pt_regs will indeed just work. So in that case, what was your
original issue?

quoted

quoted

quoted

--- a/tools/testing/selftests/bpf/Makefile
+++ b/tools/testing/selftests/bpf/Makefile

@@ -294,7 +294,8 @@ MENDIAN=$(if

$(IS_LITTLE_ENDIAN),-mlittle-endian,-mbig-endian)
   CLANG_SYS_INCLUDES = $(call get_sys_includes,$(CLANG))
   BPF_CFLAGS = -g -D__TARGET_ARCH_$(SRCARCH) $(MENDIAN) \
              -I$(INCLUDE_DIR) -I$(CURDIR) -I$(APIDIR) \
-            -I$(abspath $(OUTPUT)/../usr/include)
+            -I$(abspath $(OUTPUT)/../usr/include) \
+            -I../../../../usr/include

quoted

quoted

Signed-off-by: Pu Lehui <pulehui@huawei.com>
---
   tools/testing/selftests/bpf/progs/loop1.c     |  8 ++------
   tools/testing/selftests/bpf/progs/loop2.c     |  8 ++------
   tools/testing/selftests/bpf/progs/loop3.c     |  8 ++------
   tools/testing/selftests/bpf/progs/loop6.c     | 20
++++++-------------
   .../selftests/bpf/progs/test_overhead.c       |  8 ++------
   .../selftests/bpf/progs/test_probe_user.c     |  6 +-----
   6 files changed, 15 insertions(+), 43 deletions(-)

[...]
.

.

.

On 2021/12/22 7:52, Andrii Nakryiko wrote:

On Mon, Dec 20, 2021 at 4:58 PM Pu Lehui [off-list ref] wrote:

quoted

On 2021/12/20 22:02, Pu Lehui wrote:

quoted

On 2021/12/18 0:45, Andrii Nakryiko wrote:

quoted

On Thu, Dec 16, 2021 at 6:25 PM Pu Lehui [off-list ref] wrote:

quoted

On 2021/12/16 12:06, Andrii Nakryiko wrote:

quoted

On Tue, Dec 14, 2021 at 5:54 AM Pu Lehui [off-list ref] wrote:

quoted

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

We could probably also extend bpf_tracing.h to work with
kernel-defined pt_regs, just like we do for x86 (see __KERNEL__ and
__VMLINUX_H__ checks). It's more work, but will benefit other end
users, not just selftests.

It might change a lot. We can use header file directory generated by
"make headers_install" to fix it.

We don't have dependency on "make headers_install" and I'd rather not
add it.

What do you mean by "change a lot"?

Maybe I misunderstood your advice. Your suggestion might be to extend
bpf_tracing.h to kernel-space pt_regs, while some archs, like arm64,

yes

quoted

quoted

only support user-space. So the patch might be like this:

diff --git a/tools/lib/bpf/bpf_tracing.h b/tools/lib/bpf/bpf_tracing.h
index db05a5937105..2c3cb8e9ae92 100644
--- a/tools/lib/bpf/bpf_tracing.h
+++ b/tools/lib/bpf/bpf_tracing.h

@@ -195,9 +195,13 @@ struct pt_regs;

   #elif defined(bpf_target_arm64)

-struct pt_regs;
+#if defined(__KERNEL__)
+#define PT_REGS_ARM64 const volatile struct pt_regs
+#else
   /* arm64 provides struct user_pt_regs instead of struct pt_regs to

userspace */
   #define PT_REGS_ARM64 const volatile struct user_pt_regs
+#endif
+
   #define PT_REGS_PARM1(x) (((PT_REGS_ARM64 *)(x))->regs[0])
   #define PT_REGS_PARM2(x) (((PT_REGS_ARM64 *)(x))->regs[1])
   #define PT_REGS_PARM3(x) (((PT_REGS_ARM64 *)(x))->regs[2])

Please ignore the last reply. User-space pt_regs of arm64/s390 is the
first part of the kernel-space's, it should has covered both kernel and
userspace.

Alright, so is there still a problem or not? Looking at the definition
of struct pt_regs for arm64, just casting struct pt_regs to struct
user_pt_regs will indeed just work. So in that case, what was your
original issue?

Thanks for your reply. The original issue is, when arm64 bpf selftests cross compiling in x86_64 host, clang cannot find the arch specific uapi ptrace.h, and then the above error occur. Of course it works when compiling in arm64 host for it owns the corresponding uapi ptrace.h. So my suggestion is to add arch specific use header file directory generated by "make headers_install" for the cross compiling issue.

quoted

quoted

quoted

quoted

--- a/tools/testing/selftests/bpf/Makefile
+++ b/tools/testing/selftests/bpf/Makefile

@@ -294,7 +294,8 @@ MENDIAN=$(if

$(IS_LITTLE_ENDIAN),-mlittle-endian,-mbig-endian)
    CLANG_SYS_INCLUDES = $(call get_sys_includes,$(CLANG))
    BPF_CFLAGS = -g -D__TARGET_ARCH_$(SRCARCH) $(MENDIAN) \
               -I$(INCLUDE_DIR) -I$(CURDIR) -I$(APIDIR) \
-            -I$(abspath $(OUTPUT)/../usr/include)
+            -I$(abspath $(OUTPUT)/../usr/include) \
+            -I../../../../usr/include

quoted

quoted

Signed-off-by: Pu Lehui <pulehui@huawei.com>
---
    tools/testing/selftests/bpf/progs/loop1.c     |  8 ++------
    tools/testing/selftests/bpf/progs/loop2.c     |  8 ++------
    tools/testing/selftests/bpf/progs/loop3.c     |  8 ++------
    tools/testing/selftests/bpf/progs/loop6.c     | 20
++++++-------------
    .../selftests/bpf/progs/test_overhead.c       |  8 ++------
    .../selftests/bpf/progs/test_probe_user.c     |  6 +-----
    6 files changed, 15 insertions(+), 43 deletions(-)

[...]
.

.

.

.

On Tue, Dec 21, 2021 at 5:33 PM Pu Lehui [off-list ref] wrote:

On 2021/12/22 7:52, Andrii Nakryiko wrote:

quoted

On Mon, Dec 20, 2021 at 4:58 PM Pu Lehui [off-list ref] wrote:

quoted

On 2021/12/20 22:02, Pu Lehui wrote:

quoted

On 2021/12/18 0:45, Andrii Nakryiko wrote:

quoted

On Thu, Dec 16, 2021 at 6:25 PM Pu Lehui [off-list ref] wrote:

quoted

On 2021/12/16 12:06, Andrii Nakryiko wrote:

quoted

On Tue, Dec 14, 2021 at 5:54 AM Pu Lehui [off-list ref] wrote:

quoted

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

We could probably also extend bpf_tracing.h to work with
kernel-defined pt_regs, just like we do for x86 (see __KERNEL__ and
__VMLINUX_H__ checks). It's more work, but will benefit other end
users, not just selftests.

It might change a lot. We can use header file directory generated by
"make headers_install" to fix it.

We don't have dependency on "make headers_install" and I'd rather not
add it.

What do you mean by "change a lot"?

Maybe I misunderstood your advice. Your suggestion might be to extend
bpf_tracing.h to kernel-space pt_regs, while some archs, like arm64,

yes

quoted

quoted

only support user-space. So the patch might be like this:

diff --git a/tools/lib/bpf/bpf_tracing.h b/tools/lib/bpf/bpf_tracing.h
index db05a5937105..2c3cb8e9ae92 100644
--- a/tools/lib/bpf/bpf_tracing.h
+++ b/tools/lib/bpf/bpf_tracing.h

@@ -195,9 +195,13 @@ struct pt_regs;

   #elif defined(bpf_target_arm64)

-struct pt_regs;
+#if defined(__KERNEL__)
+#define PT_REGS_ARM64 const volatile struct pt_regs
+#else
   /* arm64 provides struct user_pt_regs instead of struct pt_regs to

userspace */
   #define PT_REGS_ARM64 const volatile struct user_pt_regs
+#endif
+
   #define PT_REGS_PARM1(x) (((PT_REGS_ARM64 *)(x))->regs[0])
   #define PT_REGS_PARM2(x) (((PT_REGS_ARM64 *)(x))->regs[1])
   #define PT_REGS_PARM3(x) (((PT_REGS_ARM64 *)(x))->regs[2])

Please ignore the last reply. User-space pt_regs of arm64/s390 is the
first part of the kernel-space's, it should has covered both kernel and
userspace.

Alright, so is there still a problem or not? Looking at the definition
of struct pt_regs for arm64, just casting struct pt_regs to struct
user_pt_regs will indeed just work. So in that case, what was your
original issue?

Thanks for your reply. The original issue is, when arm64 bpf selftests
cross compiling in x86_64 host, clang cannot find the arch specific uapi
ptrace.h, and then the above error occur. Of course it works when
compiling in arm64 host for it owns the corresponding uapi ptrace.h. So
my suggestion is to add arch specific use header file directory
generated by "make headers_install" for the cross compiling issue.

I see. Can you try adding something like:

ARCH_APIDIR := $(abspath ../../../../arch/$(SRCARCH)/include/uapi)

and then add -I$(ARCH_APIDIR) to BPF_CFLAGS?

Please let me know if that works for your cross-compilation case.

quoted

quoted

quoted

quoted

quoted

--- a/tools/testing/selftests/bpf/Makefile
+++ b/tools/testing/selftests/bpf/Makefile

@@ -294,7 +294,8 @@ MENDIAN=$(if

$(IS_LITTLE_ENDIAN),-mlittle-endian,-mbig-endian)
    CLANG_SYS_INCLUDES = $(call get_sys_includes,$(CLANG))
    BPF_CFLAGS = -g -D__TARGET_ARCH_$(SRCARCH) $(MENDIAN) \
               -I$(INCLUDE_DIR) -I$(CURDIR) -I$(APIDIR) \
-            -I$(abspath $(OUTPUT)/../usr/include)
+            -I$(abspath $(OUTPUT)/../usr/include) \
+            -I../../../../usr/include

quoted

quoted

Signed-off-by: Pu Lehui <pulehui@huawei.com>
---
    tools/testing/selftests/bpf/progs/loop1.c     |  8 ++------
    tools/testing/selftests/bpf/progs/loop2.c     |  8 ++------
    tools/testing/selftests/bpf/progs/loop3.c     |  8 ++------
    tools/testing/selftests/bpf/progs/loop6.c     | 20
++++++-------------
    .../selftests/bpf/progs/test_overhead.c       |  8 ++------
    .../selftests/bpf/progs/test_probe_user.c     |  6 +-----
    6 files changed, 15 insertions(+), 43 deletions(-)

[...]
.

.

.

.

On 2021/12/23 7:17, Andrii Nakryiko wrote:

On Tue, Dec 21, 2021 at 5:33 PM Pu Lehui [off-list ref] wrote:

quoted

On 2021/12/22 7:52, Andrii Nakryiko wrote:

quoted

On Mon, Dec 20, 2021 at 4:58 PM Pu Lehui [off-list ref] wrote:

quoted

On 2021/12/20 22:02, Pu Lehui wrote:

quoted

On 2021/12/18 0:45, Andrii Nakryiko wrote:

quoted

On Thu, Dec 16, 2021 at 6:25 PM Pu Lehui [off-list ref] wrote:

quoted

On 2021/12/16 12:06, Andrii Nakryiko wrote:

quoted

On Tue, Dec 14, 2021 at 5:54 AM Pu Lehui [off-list ref] wrote:

quoted

When building bpf selftests on arm64, the following error will occur:

progs/loop2.c:20:7: error: incomplete definition of type 'struct
user_pt_regs'

Some archs, like arm64 and riscv, use userspace pt_regs in
bpf_tracing.h, which causes build failure when bpf prog use
macro in bpf_tracing.h. So let's use vmlinux.h directly.

We could probably also extend bpf_tracing.h to work with
kernel-defined pt_regs, just like we do for x86 (see __KERNEL__ and
__VMLINUX_H__ checks). It's more work, but will benefit other end
users, not just selftests.

It might change a lot. We can use header file directory generated by
"make headers_install" to fix it.

We don't have dependency on "make headers_install" and I'd rather not
add it.

What do you mean by "change a lot"?

Maybe I misunderstood your advice. Your suggestion might be to extend
bpf_tracing.h to kernel-space pt_regs, while some archs, like arm64,

yes

quoted

quoted

only support user-space. So the patch might be like this:

diff --git a/tools/lib/bpf/bpf_tracing.h b/tools/lib/bpf/bpf_tracing.h
index db05a5937105..2c3cb8e9ae92 100644
--- a/tools/lib/bpf/bpf_tracing.h
+++ b/tools/lib/bpf/bpf_tracing.h

@@ -195,9 +195,13 @@ struct pt_regs;

    #elif defined(bpf_target_arm64)

-struct pt_regs;
+#if defined(__KERNEL__)
+#define PT_REGS_ARM64 const volatile struct pt_regs
+#else
    /* arm64 provides struct user_pt_regs instead of struct pt_regs to

userspace */
    #define PT_REGS_ARM64 const volatile struct user_pt_regs
+#endif
+
    #define PT_REGS_PARM1(x) (((PT_REGS_ARM64 *)(x))->regs[0])
    #define PT_REGS_PARM2(x) (((PT_REGS_ARM64 *)(x))->regs[1])
    #define PT_REGS_PARM3(x) (((PT_REGS_ARM64 *)(x))->regs[2])

Please ignore the last reply. User-space pt_regs of arm64/s390 is the
first part of the kernel-space's, it should has covered both kernel and
userspace.

Alright, so is there still a problem or not? Looking at the definition
of struct pt_regs for arm64, just casting struct pt_regs to struct
user_pt_regs will indeed just work. So in that case, what was your
original issue?

Thanks for your reply. The original issue is, when arm64 bpf selftests
cross compiling in x86_64 host, clang cannot find the arch specific uapi
ptrace.h, and then the above error occur. Of course it works when
compiling in arm64 host for it owns the corresponding uapi ptrace.h. So
my suggestion is to add arch specific use header file directory
generated by "make headers_install" for the cross compiling issue.

I see. Can you try adding something like:

ARCH_APIDIR := $(abspath ../../../../arch/$(SRCARCH)/include/uapi)

and then add -I$(ARCH_APIDIR) to BPF_CFLAGS?

Please let me know if that works for your cross-compilation case.

It works, thanks. I will add it to v2.

quoted

quoted

quoted

quoted

quoted

quoted

--- a/tools/testing/selftests/bpf/Makefile
+++ b/tools/testing/selftests/bpf/Makefile

@@ -294,7 +294,8 @@ MENDIAN=$(if

$(IS_LITTLE_ENDIAN),-mlittle-endian,-mbig-endian)
     CLANG_SYS_INCLUDES = $(call get_sys_includes,$(CLANG))
     BPF_CFLAGS = -g -D__TARGET_ARCH_$(SRCARCH) $(MENDIAN) \
                -I$(INCLUDE_DIR) -I$(CURDIR) -I$(APIDIR) \
-            -I$(abspath $(OUTPUT)/../usr/include)
+            -I$(abspath $(OUTPUT)/../usr/include) \
+            -I../../../../usr/include

quoted

quoted

Signed-off-by: Pu Lehui <pulehui@huawei.com>
---
     tools/testing/selftests/bpf/progs/loop1.c     |  8 ++------
     tools/testing/selftests/bpf/progs/loop2.c     |  8 ++------
     tools/testing/selftests/bpf/progs/loop3.c     |  8 ++------
     tools/testing/selftests/bpf/progs/loop6.c     | 20
++++++-------------
     .../selftests/bpf/progs/test_overhead.c       |  8 ++------
     .../selftests/bpf/progs/test_probe_user.c     |  6 +-----
     6 files changed, 15 insertions(+), 43 deletions(-)

[...]
.

.

.

.

.