2024-03-11 12:27:40

by Puranjay Mohan

Subject: [PATCH bpf-next v2] bpf: hardcode BPF_PROG_PACK_SIZE to 2MB * num_possible_nodes()

On some architectures like ARM64, PMD_SIZE can be really large in some
configurations. Like with CONFIG_ARM64_64K_PAGES=y the PMD_SIZE is
512MB.

Use 2MB * num_possible_nodes() as the size for allocations done through
the prog pack allocator. On most architectures, PMD_SIZE will be equal
to 2MB in case of 4KB pages and will be greater than 2MB for bigger page
sizes.

Fixes: ea2babac63d4 ("bpf: Simplify bpf_prog_pack_[size|mask]")
Reported-by: "kernelci.org bot" <[email protected]>
Closes: https://lore.kernel.org/all/[email protected]/
Reported-by: kernel test robot <[email protected]>
Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/
Suggested-by: Song Liu <[email protected]>
Signed-off-by: Puranjay Mohan <[email protected]>
---
kernel/bpf/core.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index 134b7979f537..a8007f7f706d 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -893,7 +893,12 @@ static LIST_HEAD(pack_list);
* CONFIG_MMU=n. Use PAGE_SIZE in these cases.
*/
#ifdef PMD_SIZE
-#define BPF_PROG_PACK_SIZE (PMD_SIZE * num_possible_nodes())
+/* PMD_SIZE is really big for some archs. It doesn't make sense to
+ * reserve too much memory in one allocation. Hardcode BPF_PROG_PACK_SIZE to
+ * 2MiB * num_possible_nodes(). On most architectures PMD_SIZE will be
+ * greater than or equal to 2MB.
+ */
+#define BPF_PROG_PACK_SIZE (SZ_2M * num_possible_nodes())
#else
#define BPF_PROG_PACK_SIZE PAGE_SIZE
#endif
--
2.40.1

2024-03-11 23:40:40

by patchwork-bot+netdevbpf

Subject: Re: [PATCH bpf-next v2] bpf: hardcode BPF_PROG_PACK_SIZE to 2MB * num_possible_nodes()

Hello:

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

On Mon, 11 Mar 2024 12:27:22 +0000 you wrote:
> On some architectures like ARM64, PMD_SIZE can be really large in some
> configurations. Like with CONFIG_ARM64_64K_PAGES=y the PMD_SIZE is
> 512MB.
>
> Use 2MB * num_possible_nodes() as the size for allocations done through
> the prog pack allocator. On most architectures, PMD_SIZE will be equal
> to 2MB in case of 4KB pages and will be greater than 2MB for bigger page
> sizes.
>
> [...]

Here is the summary with links:
- [bpf-next,v2] bpf: hardcode BPF_PROG_PACK_SIZE to 2MB * num_possible_nodes()
https://git.kernel.org/bpf/bpf-next/c/d6170e4aaf86

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html