Skip to contentSkip to conversation

Fix a few selftest failure due to 64K page

PATCH Applied Yonghong Song 4 participants 7 messages Jan 13

Subsystems (3)
lore
Share & tools
lore

Yonghong Song

author reviewer
#0 lore raw

cover Fix a few selftest failure due to 64K page

Fix a few arm64 selftest failures due to 64K page. Please see each indvidual patch for why the test failed and how the test gets fixed.

Yonghong Song (3):
selftests/bpf: Fix dmabuf_iter/lots_of_buffers failure with 64K page selftests/bpf: Fix sk_bypass_prot_mem failure with 64K page selftests/bpf: Fix verifier_arena_globals1 failure with 64K page

 tools/testing/selftests/bpf/prog_tests/dmabuf_iter.c       | 2 +-
 .../testing/selftests/bpf/prog_tests/sk_bypass_prot_mem.c  | 7 ++++++-
 .../testing/selftests/bpf/progs/verifier_arena_globals1.c  | 2 +-
 3 files changed, 8 insertions(+), 3 deletions(-)
Signature · 1 line
2.47.3
author reviewer
#1 lore raw

PATCH 1/3 selftests/bpf: Fix dmabuf_iter/lots_of_buffers failure with 64K page

On arm64 with 64K page , I observed the following test failure:
...
subtest_dmabuf_iter_check_lots_of_buffers:FAIL:total_bytes_read unexpected total_bytes_read:

      actual 4696 <= expected 65536

#97/3 dmabuf_iter/lots_of_buffers:FAIL

With 4K page on x86, the total_bytes_read is 4593. With 64K page on arm64, the total_byte_read is 4696.

In progs/dmabuf_iter.c, for each iteration, the output is
BPF_SEQ_PRINTF(seq, "%lu\n%llu\n%s\n%s\n", inode, size, name, exporter);

The only difference between 4K and 64K page is 'size' in
the above BPF_SEQ_PRINTF. The 4K page will output '4096' and the 64K page will output '65536'. So the total_bytes_read with 64K page is slighter greater than 4K page.

Adjusting the total_bytes_read from 65536 to 4096 fixed the issue.

---
 tools/testing/selftests/bpf/prog_tests/dmabuf_iter.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/bpf/prog_tests/dmabuf_iter.c +1 -1
index e442be9dde7e..fb2cea710db3 100644
@@ -233,7 +233,7 @@ static void subtest_dmabuf_iter_check_lots_of_buffers(struct dmabuf_iter *skel)
 	while ((bytes_read = read(iter_fd, buf, sizeof(buf))) > 0)
 		total_bytes_read += bytes_read;
 
-	ASSERT_GT(total_bytes_read, getpagesize(), "total_bytes_read");
+	ASSERT_GT(total_bytes_read, 4096, "total_bytes_read");
 
 	close(iter_fd);
 }
Signature · 1 line
2.47.3
author reviewer
#2 lore raw

PATCH 2/3 selftests/bpf: Fix sk_bypass_prot_mem failure with 64K page

In reply to Yonghong Song

The current selftest sk_bypass_prot_mem only supports 4K page. When running with 64K page on arm64, the following failure happens:
...
check_bypass:FAIL:no bypass unexpected no bypass: actual 3 <= expected 32
...
#385/1 sk_bypass_prot_mem/TCP :FAIL
...
check_bypass:FAIL:no bypass unexpected no bypass: actual 4 <= expected 32
...
#385/2 sk_bypass_prot_mem/UDP :FAIL
...

Adding support to 64K page as well fixed the failure.

Cc Kuniyuki Iwashima <[email protected]>
---
 .../testing/selftests/bpf/prog_tests/sk_bypass_prot_mem.c  | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
tools/testing/selftests/bpf/prog_tests/sk_bypass_prot_mem.c +6 -1
index e4940583924b..e2c867fd5244 100644
@@ -5,9 +5,14 @@
 #include "sk_bypass_prot_mem.skel.h"
 #include "network_helpers.h"
 
+#ifndef PAGE_SIZE
+#include <unistd.h>
+#define PAGE_SIZE getpagesize()
+#endif
+
 #define NR_PAGES	32
 #define NR_SOCKETS	2
-#define BUF_TOTAL	(NR_PAGES * 4096 / NR_SOCKETS)
+#define BUF_TOTAL	(NR_PAGES * PAGE_SIZE / NR_SOCKETS)
 #define BUF_SINGLE	1024
 #define NR_SEND		(BUF_TOTAL / BUF_SINGLE)
 
Signature · 1 line
2.47.3
author reviewer
#3 lore raw

PATCH 3/3 selftests/bpf: Fix verifier_arena_globals1 failure with 64K page

In reply to Yonghong Song

With 64K page on arm64, verifier_arena_globals1 failed like below:
...
libbpf: map 'arena': failed to create: -E2BIG
...
#509/1 verifier_arena_globals1/check_reserve1:FAIL
...

For 64K page, if the number of arena pages is (1UL << 20), the total memory will exceed 4G and this will cause map creation failure. Adjusting ARENA_PAGES based on the actual page size fixed the problem.

Cc Emil Tsalapatis <[email protected]>
---
 tools/testing/selftests/bpf/progs/verifier_arena_globals1.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/bpf/progs/verifier_arena_globals1.c +1 -1
index 14afef3d6442..83182ddbfb95 100644
@@ -9,7 +9,7 @@
 #include "bpf_arena_common.h"
 #include "bpf_misc.h"
 
-#define ARENA_PAGES (1UL<< (32 - 12))
+#define ARENA_PAGES (1UL<< (32 - __builtin_ffs(__PAGE_SIZE) + 1))
 #define GLOBAL_PAGES (16)
 
 struct {
Signature · 1 line
2.47.3
reviewer Reviewed-by on v1 3/3
#4 lore raw

Re: [PATCH bpf-next 3/3] selftests/bpf: Fix verifier_arena_globals1 failure with 64K page

Quoted 7 lines · Yonghong Song

With 64K page on arm64, verifier_arena_globals1 failed like below:
...
libbpf: map 'arena': failed to create: -E2BIG
...
#509/1 verifier_arena_globals1/check_reserve1:FAIL
...

Quoted 7 lines

For 64K page, if the number of arena pages is (1UL << 20), the total memory will exceed 4G and this will cause map creation failure. Adjusting ARENA_PAGES based on the actual page size fixed the problem.

Cc Emil Tsalapatis <[email protected]>

---

Quoted 16 lines
 tools/testing/selftests/bpf/progs/verifier_arena_globals1.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/bpf/progs/verifier_arena_globals1.c +1 -1
index 14afef3d6442..83182ddbfb95 100644
@@ -9,7 +9,7 @@
 #include "bpf_arena_common.h"
 #include "bpf_misc.h"
 
-#define ARENA_PAGES (1UL<< (32 - 12))
+#define ARENA_PAGES (1UL<< (32 - __builtin_ffs(__PAGE_SIZE) + 1))
 #define GLOBAL_PAGES (16)
 
 struct {
Tested-by on the v1 series
#5 lore raw
In reply to Yonghong Song
Quoted 8 lines · Yonghong Song

Fix a few arm64 selftest failures due to 64K page. Please see each indvidual patch for why the test failed and how the test gets fixed.

Yonghong Song (3):
selftests/bpf: Fix dmabuf_iter/lots_of_buffers failure with 64K page selftests/bpf: Fix sk_bypass_prot_mem failure with 64K page selftests/bpf: Fix verifier_arena_globals1 failure with 64K page

For the series,

#6 lore raw
In reply to Yonghong Song
This series was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:

Hello:

This series was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:

Quoted 9 lines · On Mon, 12 Jan 2026 22:10:18 -0800 you wrote:

Fix a few arm64 selftest failures due to 64K page. Please see each indvidual patch for why the test failed and how the test gets fixed.

Yonghong Song (3):
selftests/bpf: Fix dmabuf_iter/lots_of_buffers failure with 64K page selftests/bpf: Fix sk_bypass_prot_mem failure with 64K page selftests/bpf: Fix verifier_arena_globals1 failure with 64K page

[...]

Here is the summary with links:
- [bpf-next,1/3] selftests/bpf: Fix dmabuf_iter/lots_of_buffers failure with 64K page

    https://git.kernel.org/bpf/bpf-next/c/2465a08d433d
  - [bpf-next,2/3] selftests/bpf: Fix sk_bypass_prot_mem failure with 64K page
    https://git.kernel.org/bpf/bpf-next/c/d2f7cd20a7c7
  - [bpf-next,3/3] selftests/bpf: Fix verifier_arena_globals1 failure with 64K page
    https://git.kernel.org/bpf/bpf-next/c/951d79017e8a

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html