linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: k.khlebnikov@samsung.com (Konstantin Khlebnikov)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH RFC 3/3] ARM: reduce size of page table directory for short mode
Date: Tue, 18 Nov 2014 19:53:25 +0400	[thread overview]
Message-ID: <20141118165325.9958.84370.stgit@buzz> (raw)
In-Reply-To: <20141118165313.9958.1242.stgit@buzz>

Virtual memory above PAGE_OFFSET is covered by TTBR1 which always points
to swapper_pg_dir, thus size of the rest pgd tables might be reduced.
Of course idmap_pgd must be full 16k because it's used without split.
It seems nobody accesses pgd entries above PAGE_OFFSET via pgd_offset(),
all access is done via pgd_offset_k() which refers directly to init_mm.

This patch saves 8k per process for 2G/2G split (12k for 1G/3G split)
and kills source of frequent order-2 allocations. Unfortunately most
commonly used 3G/1G mode isn't supported, because first part must be
power of two. LPAE supports this mode, but it's pgd is already tiny.

Signed-off-by: Konstantin Khlebnikov <k.khlebnikov@samsung.com>
---
 arch/arm/mm/idmap.c |    6 ++++++
 arch/arm/mm/pgd.c   |   14 +++++++++-----
 2 files changed, 15 insertions(+), 5 deletions(-)

diff --git a/arch/arm/mm/idmap.c b/arch/arm/mm/idmap.c
index cc51b40..cea7ee1 100644
--- a/arch/arm/mm/idmap.c
+++ b/arch/arm/mm/idmap.c
@@ -100,7 +100,13 @@ extern char  __idmap_text_start[], __idmap_text_end[];
 
 static int __init init_static_idmap(void)
 {
+#ifdef CONFIG_ARM_LPAE
 	idmap_pgd = pgd_alloc(&init_mm);
+#else
+	idmap_pgd = (pgd_t *)__get_free_pages(GFP_KERNEL, 2);
+	if (idmap_pgd)
+		memcpy(idmap_pgd, swapper_pg_dir, PTRS_PER_PGD * sizeof(pgd_t));
+#endif
 	if (!idmap_pgd)
 		return -ENOMEM;
 
diff --git a/arch/arm/mm/pgd.c b/arch/arm/mm/pgd.c
index 2493795..3fbcb5a 100644
--- a/arch/arm/mm/pgd.c
+++ b/arch/arm/mm/pgd.c
@@ -20,15 +20,19 @@
 #include "mm.h"
 
 #ifdef CONFIG_ARM_LPAE
+#define TTBR0_PTRS_PER_PGD	PTRS_PER_PGD
 #define __pgd_alloc()	kmalloc(PTRS_PER_PGD * sizeof(pgd_t), GFP_KERNEL)
 #define __pgd_free(pgd)	kfree(pgd)
 #else
-#define __pgd_alloc()	(pgd_t *)__get_free_pages(GFP_KERNEL | __GFP_REPEAT, 2)
-#define __pgd_free(pgd)	free_pages((unsigned long)pgd, 2)
+#define TTBR0_PTRS_PER_PGD	(PTRS_PER_PGD >> TTBR1_SIZE)
+#define __pgd_alloc()	(pgd_t *)__get_free_pages(GFP_KERNEL | __GFP_REPEAT, \
+							2 - TTBR1_SIZE)
+#define __pgd_free(pgd)	free_pages((unsigned long)pgd, 2 - TTBR1_SIZE)
 #endif
 
 /*
- * need to get a 16k page for level 1
+ * We need 4k/8k/16k for pgd in short mode for CONFIG_VMSPLIT_1G/2G/3G
+ * or only 64 bytes if LPAE is enabled.
  */
 pgd_t *pgd_alloc(struct mm_struct *mm)
 {
@@ -48,9 +52,9 @@ pgd_t *pgd_alloc(struct mm_struct *mm)
 	 */
 	init_pgd = pgd_offset_k(0);
 	memcpy(new_pgd + USER_PTRS_PER_PGD, init_pgd + USER_PTRS_PER_PGD,
-		       (PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t));
+		(TTBR0_PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t));
 
-	clean_dcache_area(new_pgd, PTRS_PER_PGD * sizeof(pgd_t));
+	clean_dcache_area(new_pgd, TTBR0_PTRS_PER_PGD * sizeof(pgd_t));
 
 #ifdef CONFIG_ARM_LPAE
 	/*

  parent reply	other threads:[~2014-11-18 15:53 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-11-18 15:53 [PATCH RFC 1/3] ARM: LPAE: rework TTBR0/TTBR1 split Konstantin Khlebnikov
2014-11-18 15:53 ` [PATCH RFC 2/3] ARM: enable TTBR0/TTBR1 split in short mode Konstantin Khlebnikov
2014-11-18 18:04   ` Catalin Marinas
2014-11-18 15:53 ` Konstantin Khlebnikov [this message]
2014-11-18 18:14   ` [PATCH RFC 3/3] ARM: reduce size of page table directory for " Catalin Marinas
2014-11-18 19:11     ` Russell King - ARM Linux
2014-11-18 23:02       ` Catalin Marinas

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=20141118165325.9958.84370.stgit@buzz \
    --to=k.khlebnikov@samsung.com \
    --cc=linux-arm-kernel@lists.infradead.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;
as well as URLs for NNTP newsgroup(s).