From: Ethan Yang <ethan.yang.kernel@gmail.com>
To: chenhuacai@kernel.org
Cc: kernel@xen0n.name, loongarch@lists.linux.dev,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
akpm@linux-foundation.org
Subject: [PATCH v2] LoongArch: mm: fix FLATMEM ARCH_PFN_OFFSET for non-zero DRAM base
Date: Wed, 1 Apr 2026 04:17:00 +0000 [thread overview]
Message-ID: <20260401041700.3260-1-ethan.yang.kernel@gmail.com> (raw)
In-Reply-To: <CAKeWACY11_1JL_OagfxUhQckW4rkAjhwXxUnVHNPdHF60Uc81w@mail.gmail.com>
With CONFIG_FLATMEM, pfn_to_page() is computed as:
mem_map + (pfn - ARCH_PFN_OFFSET)
LoongArch currently derives ARCH_PFN_OFFSET from PFN_UP(PHYS_OFFSET),
while PHYS_OFFSET defaults to 0. On systems whose first DRAM range does
not start at physical address 0, ARCH_PFN_OFFSET becomes smaller than
the real RAM base PFN.
As a result, pfn_to_page() points past the intended mem_map base during
early memmap initialization. This causes struct page initialization to
touch wrong entries and can lead to early boot failures in the
memmap_init_range() path.
Fix this by deriving ARCH_PFN_OFFSET from min_low_pfn, and initialize/
recalculate min_low_pfn from memblock_start_of_DRAM() in LoongArch
memory init so PFN-to-page translation matches the actual DRAM base.
This change only fixes the FLATMEM PFN base and does not alter the
existing VA/PA linear mapping scheme.
Signed-off-by: Ethan Yang <ethan.yang.kernel@gmail.com>
---
v2:
- Fixed tab-to-space formatting.
---
arch/loongarch/include/asm/page.h | 3 ++-
arch/loongarch/kernel/mem.c | 1 +
arch/loongarch/kernel/setup.c | 1 +
3 files changed, 4 insertions(+), 1 deletion(-)
diff --git a/arch/loongarch/include/asm/page.h b/arch/loongarch/include/asm/page.h
index 327bf0bc9..80aea8b51 100644
--- a/arch/loongarch/include/asm/page.h
+++ b/arch/loongarch/include/asm/page.h
@@ -25,7 +25,8 @@
* used in our early mem init code for all memory models.
* So always define it.
*/
-#define ARCH_PFN_OFFSET PFN_UP(PHYS_OFFSET)
+extern unsigned long min_low_pfn;
+#define ARCH_PFN_OFFSET min_low_pfn
extern void clear_page(void *page);
extern void copy_page(void *to, void *from);
diff --git a/arch/loongarch/kernel/mem.c b/arch/loongarch/kernel/mem.c
index 8ab1ffedc..03835f8cb 100644
--- a/arch/loongarch/kernel/mem.c
+++ b/arch/loongarch/kernel/mem.c
@@ -46,6 +46,7 @@ void __init memblock_init(void)
}
}
+ min_low_pfn = PFN_UP(memblock_start_of_DRAM());
max_pfn = PFN_DOWN(memblock_end_of_DRAM());
max_low_pfn = min(PFN_DOWN(HIGHMEM_START), max_pfn);
memblock_set_current_limit(PFN_PHYS(max_low_pfn));
diff --git a/arch/loongarch/kernel/setup.c b/arch/loongarch/kernel/setup.c
index 839b23ede..625be2494 100644
--- a/arch/loongarch/kernel/setup.c
+++ b/arch/loongarch/kernel/setup.c
@@ -394,6 +394,7 @@ static void __init check_kernel_sections_mem(void)
static void __init arch_mem_init(char **cmdline_p)
{
/* Recalculate max_low_pfn for "mem=xxx" */
+ min_low_pfn = PFN_UP(memblock_start_of_DRAM());
max_pfn = PFN_DOWN(memblock_end_of_DRAM());
max_low_pfn = min(PFN_DOWN(HIGHMEM_START), max_pfn);
--
2.47.3
prev parent reply other threads:[~2026-04-01 4:17 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-01 2:33 [PATCH] LoongArch: mm: fix FLATMEM ARCH_PFN_OFFSET for non-zero DRAM base Ethan Yang
2026-04-01 4:17 ` Ethan Yang [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=20260401041700.3260-1-ethan.yang.kernel@gmail.com \
--to=ethan.yang.kernel@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=chenhuacai@kernel.org \
--cc=kernel@xen0n.name \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=loongarch@lists.linux.dev \
/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