From: Ard Biesheuvel <ardb+git@google.com>
To: linux-kernel@vger.kernel.org
Cc: x86@kernel.org, Ard Biesheuvel <ardb@kernel.org>,
Ingo Molnar <mingo@kernel.org>,
Linus Torvalds <torvalds@linux-foundation.org>
Subject: [RFC PATCH 3/3] x86/boot: Use alternatives based selector for 5-level paging constants
Date: Tue, 6 May 2025 17:45:36 +0200 [thread overview]
Message-ID: <20250506154532.1281909-8-ardb+git@google.com> (raw)
In-Reply-To: <20250506154532.1281909-5-ardb+git@google.com>
From: Ard Biesheuvel <ardb@kernel.org>
There are a couple of cases where pgtable_l5_enabled() is not used for
control flow, but for selecting the value of a global constant. There
are some other occurrences of such constants where the value is stored
in a global variable that needs to be assigned sufficiently early.
To make this more robust, base all of these on a new helper that uses
alternatives based code patching to select one of two immediate values,
based on whether 5 level paging is in use.
Base this on __pgtable_l5_enabled, which is guaranteed to be set to the
right value before C code ever observes it. This allows the helper to
have the 'const' attribute.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
---
arch/x86/include/asm/page_64_types.h | 2 +-
arch/x86/include/asm/pgtable_64_types.h | 28 +++++++++++++++++---
2 files changed, 26 insertions(+), 4 deletions(-)
diff --git a/arch/x86/include/asm/page_64_types.h b/arch/x86/include/asm/page_64_types.h
index 1faa8f88850a..c5631dc4ab16 100644
--- a/arch/x86/include/asm/page_64_types.h
+++ b/arch/x86/include/asm/page_64_types.h
@@ -54,7 +54,7 @@
#define __PHYSICAL_MASK_SHIFT 52
#ifdef CONFIG_X86_5LEVEL
-#define __VIRTUAL_MASK_SHIFT (pgtable_l5_enabled() ? 56 : 47)
+#define __VIRTUAL_MASK_SHIFT (choose_l5_enabled(56, 47))
/* See task_size_max() in <asm/page_64.h> */
#else
#define __VIRTUAL_MASK_SHIFT 47
diff --git a/arch/x86/include/asm/pgtable_64_types.h b/arch/x86/include/asm/pgtable_64_types.h
index 2c498d16609c..bb4f54ac2e62 100644
--- a/arch/x86/include/asm/pgtable_64_types.h
+++ b/arch/x86/include/asm/pgtable_64_types.h
@@ -43,6 +43,28 @@ static inline bool __attribute_const__ pgtable_l5_enabled(void)
t_no:
return false;
}
+
+static inline int __attribute_const__ choose_l5_enabled(int yes, int no)
+{
+ int ret = no;
+
+ asm_inline(ALTERNATIVE_TERNARY("jmp 6f; 8:", %c[feat], "movl %[yes], %[ret]", "")
+ " .pushsection .altinstr_aux,\"ax\" \n"
+ "6: pushfq \n"
+ " testb $1, %a[l5en] \n"
+ " jz 7f \n"
+ " movl %[yes], %[ret] \n"
+ "7: popfq \n"
+ " jmp 8b \n"
+ " .popsection \n"
+ : [ret] "+rm" (ret)
+ : [feat] "i" (X86_FEATURE_LA57),
+ [yes] "i" (yes),
+ [l5en] "i" (&__pgtable_l5_enabled));
+
+ return ret;
+}
+
#else
#define pgtable_l5_enabled() 0
#endif /* CONFIG_X86_5LEVEL */
@@ -59,7 +81,7 @@ extern unsigned int ptrs_per_p4d;
/*
* PGDIR_SHIFT determines what a top-level page table entry can map
*/
-#define PGDIR_SHIFT pgdir_shift
+#define PGDIR_SHIFT choose_l5_enabled(48, 39)
#define PTRS_PER_PGD 512
/*
@@ -67,7 +89,7 @@ extern unsigned int ptrs_per_p4d;
*/
#define P4D_SHIFT 39
#define MAX_PTRS_PER_P4D 512
-#define PTRS_PER_P4D ptrs_per_p4d
+#define PTRS_PER_P4D choose_l5_enabled(MAX_PTRS_PER_P4D, 1)
#define P4D_SIZE (_AC(1, UL) << P4D_SHIFT)
#define P4D_MASK (~(P4D_SIZE - 1))
@@ -138,7 +160,7 @@ extern unsigned int ptrs_per_p4d;
#ifdef CONFIG_DYNAMIC_MEMORY_LAYOUT
# define VMALLOC_START vmalloc_base
-# define VMALLOC_SIZE_TB (pgtable_l5_enabled() ? VMALLOC_SIZE_TB_L5 : VMALLOC_SIZE_TB_L4)
+# define VMALLOC_SIZE_TB ((unsigned long)choose_l5_enabled(VMALLOC_SIZE_TB_L5, VMALLOC_SIZE_TB_L4))
# define VMEMMAP_START vmemmap_base
#else
# define VMALLOC_START __VMALLOC_BASE_L4
--
2.49.0.987.g0cc8ee98dc-goog
next prev parent reply other threads:[~2025-05-06 15:49 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-05-06 15:45 [RFC PATCH 0/3] x86: Robustify pgtable_l5_enabled() Ard Biesheuvel
2025-05-06 15:45 ` [RFC PATCH 1/3] x86/boot: Use a single source of truth for pgtable_l5_enabled() Ard Biesheuvel
2025-05-06 15:45 ` [RFC PATCH 2/3] x86/boot: Set __pgtable_l5_enabled correctly before calling into C code Ard Biesheuvel
2025-05-10 13:54 ` David Laight
2025-05-10 14:51 ` Ard Biesheuvel
2025-05-06 15:45 ` Ard Biesheuvel [this message]
2025-05-06 16:23 ` [RFC PATCH 3/3] x86/boot: Use alternatives based selector for 5-level paging constants Linus Torvalds
2025-05-06 16:34 ` Ard Biesheuvel
2025-05-06 17:03 ` Linus Torvalds
2025-05-06 17:25 ` Ard Biesheuvel
2025-05-06 17:39 ` Ingo Molnar
2025-05-06 17:47 ` Ard Biesheuvel
2025-05-06 17:43 ` Linus Torvalds
2025-05-06 17:52 ` Ard Biesheuvel
2025-05-06 18:17 ` Linus Torvalds
2025-05-06 18:39 ` Linus Torvalds
2025-05-06 19:14 ` Ard Biesheuvel
2025-05-06 19:41 ` Linus Torvalds
2025-05-06 19:50 ` Ard Biesheuvel
2025-05-07 13:51 ` Ard Biesheuvel
2025-05-06 16:50 ` Ingo Molnar
2025-05-06 16:58 ` Ard Biesheuvel
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=20250506154532.1281909-8-ardb+git@google.com \
--to=ardb+git@google.com \
--cc=ardb@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=torvalds@linux-foundation.org \
--cc=x86@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
all inboxes | Powered by JetHome®