From: Pratyush Yadav <pratyush@kernel.org>
To: Mike Rapoport <rppt@kernel.org>,
Pasha Tatashin <pasha.tatashin@soleen.com>,
Pratyush Yadav <pratyush@kernel.org>,
Alexander Graf <graf@amazon.com>,
Muchun Song <muchun.song@linux.dev>,
Oscar Salvador <osalvador@suse.de>,
David Hildenbrand <david@kernel.org>,
Andrew Morton <akpm@linux-foundation.org>,
Jason Miu <jasonmiu@google.com>,
Jork Loeser <jloeser@linux.microsoft.com>
Cc: kexec@lists.infradead.org, linux-mm@kvack.org,
linux-kernel@vger.kernel.org
Subject: [PATCH v2 18/18] kho: exclude hugetlb memory from scratch size calculation
Date: Fri, 5 Jun 2026 20:34:51 +0200 [thread overview]
Message-ID: <20260605183501.3884950-19-pratyush@kernel.org> (raw)
In-Reply-To: <20260605183501.3884950-1-pratyush@kernel.org>
From: "Pratyush Yadav (Google)" <pratyush@kernel.org>
HugeTLB pages can be preserved memory. So they are never allocated from
scratch. Instead, they are allocated from the memory blocks with no
preserved memory. These areas are detected at runtime on each boot.
But since they are allocated via memblock, they show up as RSRV_KERN,
and blow up the scratch size when scratch scale is in use.
All hugetlb pages are marked RSRV_HUGETLB. Subtract their size from
RSRV_KERN when calculating scratch sizes.
Signed-off-by: Pratyush Yadav (Google) <pratyush@kernel.org>
---
kernel/liveupdate/kexec_handover.c | 13 ++++++++++++-
1 file changed, 12 insertions(+), 1 deletion(-)
diff --git a/kernel/liveupdate/kexec_handover.c b/kernel/liveupdate/kexec_handover.c
index b3c33f150e85..0d106c9197d9 100644
--- a/kernel/liveupdate/kexec_handover.c
+++ b/kernel/liveupdate/kexec_handover.c
@@ -744,7 +744,8 @@ static void __init scratch_size_update(void)
{
/*
* If fixed sizes are not provided via command line, calculate them
- * now.
+ * now. Use RSRV_KERN to count allocated memory, but remove HugeTLB
+ * allocations from it because they never get allocated from scratch.
*/
if (scratch_scale) {
phys_addr_t size;
@@ -752,12 +753,19 @@ static void __init scratch_size_update(void)
size = memblock_reserved_size_flags(ARCH_LOW_ADDRESS_LIMIT,
NUMA_NO_NODE,
MEMBLOCK_RSRV_KERN);
+ size -= memblock_reserved_size_flags(ARCH_LOW_ADDRESS_LIMIT,
+ NUMA_NO_NODE,
+ MEMBLOCK_RSRV_HUGETLB);
+
size = size * scratch_scale / 100;
scratch_size_lowmem = size;
size = memblock_reserved_size_flags(MEMBLOCK_ALLOC_ANYWHERE,
NUMA_NO_NODE,
MEMBLOCK_RSRV_KERN);
+ size -= memblock_reserved_size_flags(MEMBLOCK_ALLOC_ANYWHERE,
+ NUMA_NO_NODE,
+ MEMBLOCK_RSRV_HUGETLB);
size = size * scratch_scale / 100 - scratch_size_lowmem;
scratch_size_global = size;
}
@@ -777,6 +785,9 @@ static phys_addr_t __init scratch_size_node(int nid)
if (scratch_scale) {
size = memblock_reserved_size_flags(MEMBLOCK_ALLOC_ANYWHERE,
nid, MEMBLOCK_RSRV_KERN);
+ /* Do not count HugeTLB pages. */
+ size -= memblock_reserved_size_flags(MEMBLOCK_ALLOC_ANYWHERE,
+ nid, MEMBLOCK_RSRV_HUGETLB);
size = size * scratch_scale / 100;
} else {
size = scratch_size_pernode;
--
2.54.0.1032.g2f8565e1d1-goog
prev parent reply other threads:[~2026-06-05 18:36 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-05 18:34 [PATCH v2 00/18] kho: make boot time huge page allocation work nicely with KHO Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 01/18] kho: generalize radix tree APIs Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 02/18] kho: disallow wide keys in radix tree Pratyush Yadav
2026-06-05 22:06 ` Jork Loeser
2026-06-08 9:10 ` Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 03/18] kho: return virtual address of mem_map Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 04/18] kho: store incoming radix tree in kho_in Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 05/18] kho: move all memory retrieval logic to kho_mem_retrieve() Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 06/18] kho: add a struct for radix callbacks Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 07/18] kho: add callback for table pages Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 08/18] kho: add data argument to radix walk callback Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 09/18] kho: allow early-boot usage of the KHO radix tree Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 10/18] kho: allow destroying " Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 11/18] kho: add kho_radix_init_tree() Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 12/18] kho: export kho_scratch_overlap() Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 13/18] kho: initialize kho_scratch pointer earlier in boot Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 14/18] memblock: use kho_scratch_overlap() to decide migratetype Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 15/18] kho: extend scratch Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 16/18] memblock: make HugeTLB bootmem allocation work with KHO Pratyush Yadav
2026-06-05 18:34 ` [PATCH v2 17/18] memblock: allow calculating reserved size by flags Pratyush Yadav
2026-06-05 18:34 ` Pratyush Yadav [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260605183501.3884950-19-pratyush@kernel.org \
--to=pratyush@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=david@kernel.org \
--cc=graf@amazon.com \
--cc=jasonmiu@google.com \
--cc=jloeser@linux.microsoft.com \
--cc=kexec@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=muchun.song@linux.dev \
--cc=osalvador@suse.de \
--cc=pasha.tatashin@soleen.com \
--cc=rppt@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox