The quilt patch titled
Subject: mm: vmalloc: don't account for number of nodes for HUGE_VMAP allocations
has been removed from the -mm tree. Its filename was
mm-vmalloc-dont-account-for-number-of-nodes-for-huge_vmap-allocations.patch
This patch was dropped because it was merged into the mm-stable branch
of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
------------------------------------------------------
From: "Mike Rapoport (Microsoft)" <rppt@xxxxxxxxxx>
Subject: mm: vmalloc: don't account for number of nodes for HUGE_VMAP allocations
Date: Wed, 23 Oct 2024 19:27:05 +0300
vmalloc allocations with VM_ALLOW_HUGE_VMAP that do not explicitly specify
node ID will use huge pages only if size_per_node is larger than a huge
page.
Still the actual allocated memory is not distributed between nodes and
there is no advantage in such approach. On the contrary, BPF allocates
SZ_2M * num_possible_nodes() for each new bpf_prog_pack, while it could do
with a single huge page per pack.
Don't account for number of nodes for VM_ALLOW_HUGE_VMAP with NUMA_NO_NODE
and use huge pages whenever the requested allocation size is larger than a
huge page.
Link: https://lkml.kernel.org/r/20241023162711.2579610-3-rppt@xxxxxxxxxx
Signed-off-by: Mike Rapoport (Microsoft) <rppt@xxxxxxxxxx>
Reviewed-by: Christoph Hellwig <hch@xxxxxx>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@xxxxxxxxx>
Reviewed-by: Luis Chamberlain <mcgrof@xxxxxxxxxx>
Tested-by: kdevops <kdevops@xxxxxxxxxxxxxxx>
Cc: Andreas Larsson <andreas@xxxxxxxxxxx>
Cc: Andy Lutomirski <luto@xxxxxxxxxx>
Cc: Ard Biesheuvel <ardb@xxxxxxxxxx>
Cc: Arnd Bergmann <arnd@xxxxxxxx>
Cc: Borislav Petkov (AMD) <bp@xxxxxxxxx>
Cc: Brian Cain <bcain@xxxxxxxxxxx>
Cc: Catalin Marinas <catalin.marinas@xxxxxxx>
Cc: Christophe Leroy <christophe.leroy@xxxxxxxxxx>
Cc: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>
Cc: Dinh Nguyen <dinguyen@xxxxxxxxxx>
Cc: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>
Cc: Guo Ren <guoren@xxxxxxxxxx>
Cc: Helge Deller <deller@xxxxxx>
Cc: Huacai Chen <chenhuacai@xxxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: Johannes Berg <johannes@xxxxxxxxxxxxxxxx>
Cc: John Paul Adrian Glaubitz <glaubitz@xxxxxxxxxxxxxxxxxxx>
Cc: Kent Overstreet <kent.overstreet@xxxxxxxxx>
Cc: Liam R. Howlett <Liam.Howlett@xxxxxxxxxx>
Cc: Mark Rutland <mark.rutland@xxxxxxx>
Cc: Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx>
Cc: Matt Turner <mattst88@xxxxxxxxx>
Cc: Max Filippov <jcmvbkbc@xxxxxxxxx>
Cc: Michael Ellerman <mpe@xxxxxxxxxxxxxx>
Cc: Michal Simek <monstr@xxxxxxxxx>
Cc: Oleg Nesterov <oleg@xxxxxxxxxx>
Cc: Palmer Dabbelt <palmer@xxxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Richard Weinberger <richard@xxxxxx>
Cc: Russell King <linux@xxxxxxxxxxxxxxx>
Cc: Song Liu <song@xxxxxxxxxx>
Cc: Stafford Horne <shorne@xxxxxxxxx>
Cc: Steven Rostedt (Google) <rostedt@xxxxxxxxxxx>
Cc: Suren Baghdasaryan <surenb@xxxxxxxxxx>
Cc: Thomas Bogendoerfer <tsbogend@xxxxxxxxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Vineet Gupta <vgupta@xxxxxxxxxx>
Cc: Will Deacon <will@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/vmalloc.c | 9 ++-------
1 file changed, 2 insertions(+), 7 deletions(-)
--- a/mm/vmalloc.c~mm-vmalloc-dont-account-for-number-of-nodes-for-huge_vmap-allocations
+++ a/mm/vmalloc.c
@@ -3779,8 +3779,6 @@ void *__vmalloc_node_range_noprof(unsign
}
if (vmap_allow_huge && (vm_flags & VM_ALLOW_HUGE_VMAP)) {
- unsigned long size_per_node;
-
/*
* Try huge pages. Only try for PAGE_KERNEL allocations,
* others like modules don't yet expect huge pages in
@@ -3788,13 +3786,10 @@ void *__vmalloc_node_range_noprof(unsign
* supporting them.
*/
- size_per_node = size;
- if (node == NUMA_NO_NODE)
- size_per_node /= num_online_nodes();
- if (arch_vmap_pmd_supported(prot) && size_per_node >= PMD_SIZE)
+ if (arch_vmap_pmd_supported(prot) && size >= PMD_SIZE)
shift = PMD_SHIFT;
else
- shift = arch_vmap_pte_supported_shift(size_per_node);
+ shift = arch_vmap_pte_supported_shift(size);
align = max(real_align, 1UL << shift);
size = ALIGN(real_size, 1UL << shift);
_
Patches currently in -mm which might be from rppt@xxxxxxxxxx are
