From: Andi Kleen <ak@suse.de>
To: jeremy@goop.org, ak@suse.de, zach@vmware.com,
chrisw@sous-sol.org, ebiederm@xmission.com,
torvalds@linux-foundation.org, patches@x86-64.org,
linux-kernel@vger.kernel.org
Subject: [PATCH] [29/40] i386: map enough initial memory to create lowmem mappings
Date: Mon, 30 Apr 2007 12:28:03 +0200 (CEST) [thread overview]
Message-ID: <20070430102803.C9463151D2@wotan.suse.de> (raw)
In-Reply-To: <200704301227.598020000@suse.de>
From: Jeremy Fitzhardinge <jeremy@goop.org>
head.S creates the very initial pagetable for the kernel. This just
maps enough space for the kernel itself, and an allocation bitmap.
The amount of mapped memory is rounded up to 4Mbytes, and so this
typically ends up mapping 8Mbytes of memory.
When booting, pagetable_init() needs to create mappings for all
lowmem, and the pagetables for these mappings are allocated from the
free pages around the kernel in low memory. If the number of
pagetable pages + kernel size exceeds head.S's initial mapping, it
will end up faulting on an unmapped page. This will only happen with
specific combinations of kernel size and memory size.
This patch makes sure that head.S also maps enough space to fit the
kernel pagetables as well as the kernel itself. It ends up using an
additional two pages of unreclaimable memory.
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Chris Wright <chrisw@sous-sol.org>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
---
arch/i386/kernel/asm-offsets.c | 6 ++++++
arch/i386/kernel/head.S | 23 +++++++++++++++++++----
2 files changed, 25 insertions(+), 4 deletions(-)
===================================================================
Index: linux/arch/i386/kernel/asm-offsets.c
===================================================================
--- linux.orig/arch/i386/kernel/asm-offsets.c
+++ linux/arch/i386/kernel/asm-offsets.c
@@ -11,6 +11,7 @@
#include <linux/suspend.h>
#include <asm/ucontext.h>
#include "sigframe.h"
+#include <asm/pgtable.h>
#include <asm/fixmap.h>
#include <asm/processor.h>
#include <asm/thread_info.h>
@@ -96,6 +97,11 @@ void foo(void)
sizeof(struct tss_struct));
DEFINE(PAGE_SIZE_asm, PAGE_SIZE);
+ DEFINE(PAGE_SHIFT_asm, PAGE_SHIFT);
+ DEFINE(PTRS_PER_PTE, PTRS_PER_PTE);
+ DEFINE(PTRS_PER_PMD, PTRS_PER_PMD);
+ DEFINE(PTRS_PER_PGD, PTRS_PER_PGD);
+
DEFINE(VDSO_PRELINK_asm, VDSO_PRELINK);
OFFSET(crypto_tfm_ctx_offset, crypto_tfm, __crt_ctx);
Index: linux/arch/i386/kernel/head.S
===================================================================
--- linux.orig/arch/i386/kernel/head.S
+++ linux/arch/i386/kernel/head.S
@@ -34,17 +34,32 @@
/*
* This is how much memory *in addition to the memory covered up to
- * and including _end* we need mapped initially. We need one bit for
- * each possible page, but only in low memory, which means
- * 2^32/4096/8 = 128K worst case (4G/4G split.)
+ * and including _end* we need mapped initially.
+ * We need:
+ * - one bit for each possible page, but only in low memory, which means
+ * 2^32/4096/8 = 128K worst case (4G/4G split.)
+ * - enough space to map all low memory, which means
+ * (2^32/4096) / 1024 pages (worst case, non PAE)
+ * (2^32/4096) / 512 + 4 pages (worst case for PAE)
+ * - a few pages for allocator use before the kernel pagetable has
+ * been set up
*
* Modulo rounding, each megabyte assigned here requires a kilobyte of
* memory, which is currently unreclaimed.
*
* This should be a multiple of a page.
*/
-#define INIT_MAP_BEYOND_END (128*1024)
+LOW_PAGES = 1<<(32-PAGE_SHIFT_asm)
+#if PTRS_PER_PMD > 1
+PAGE_TABLE_SIZE = (LOW_PAGES / PTRS_PER_PMD) + PTRS_PER_PGD
+#else
+PAGE_TABLE_SIZE = (LOW_PAGES / PTRS_PER_PGD)
+#endif
+BOOTBITMAP_SIZE = LOW_PAGES / 8
+ALLOCATOR_SLOP = 4
+
+INIT_MAP_BEYOND_END = BOOTBITMAP_SIZE + (PAGE_TABLE_SIZE + ALLOCATOR_SLOP)*PAGE_SIZE_asm
/*
* 32-bit kernel entrypoint; only used by the boot CPU. On entry,
next prev parent reply other threads:[~2007-04-30 10:30 UTC|newest]
Thread overview: 55+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-04-30 10:27 [PATCH] [0/40] x86 candidate patches for review V: paravirt patches Andi Kleen
2007-04-30 10:27 ` [PATCH] [1/40] x86_64: update MAINTAINERS Andi Kleen
2007-04-30 10:27 ` [PATCH] [2/40] i386: Remove CONFIG_DEBUG_PARAVIRT Andi Kleen
2007-04-30 10:27 ` [PATCH] [3/40] i386: use paravirt_nop to consistently mark no-op operations Andi Kleen
2007-04-30 10:27 ` [PATCH] [4/40] i386: Add pagetable accessors to pack and unpack pagetable entries Andi Kleen
2007-04-30 10:27 ` [PATCH] [5/40] i386: Hooks to set up initial pagetable Andi Kleen
2007-04-30 10:27 ` [PATCH] [6/40] i386: Allocate a fixmap slot Andi Kleen
2007-04-30 10:27 ` [PATCH] [7/40] i386: Allow paravirt backend to choose kernel PMD sharing Andi Kleen
2007-04-30 10:27 ` [PATCH] [8/40] x86: add hooks to intercept mm creation and destruction Andi Kleen
2007-04-30 10:27 ` [PATCH] [9/40] i386: rename struct paravirt_patch to paravirt_patch_site for clarity Andi Kleen
2007-04-30 10:27 ` [PATCH] [10/40] i386: Use patch site IDs computed from offset in paravirt_ops structure Andi Kleen
2007-04-30 10:27 ` [PATCH] [11/40] i386: Fix patch site clobbers to include return register Andi Kleen
2007-04-30 10:27 ` [PATCH] [12/40] i386: Consistently wrap paravirt ops callsites to make them patchable Andi Kleen
2007-04-30 10:27 ` [PATCH] [13/40] i386: Document asm-i386/paravirt.h Andi Kleen
2007-04-30 10:27 ` [PATCH] [14/40] i386: add common patching machinery Andi Kleen
2007-04-30 10:27 ` [PATCH] [15/40] i386: add flush_tlb_others paravirt_op Andi Kleen
2007-04-30 10:27 ` [PATCH] [16/40] i386: revert map_pt_hook Andi Kleen
2007-04-30 10:27 ` [PATCH] [17/40] i386: add kmap_atomic_pte for mapping highpte pages Andi Kleen
2007-04-30 10:27 ` [PATCH] [18/40] i386: flush lazy mmu updates on kunmap_atomic Andi Kleen
2007-04-30 10:27 ` [PATCH] [19/40] i386: fix paravirt-documentation Andi Kleen
2007-04-30 11:07 ` [patches] " Nigel Cunningham
2007-04-30 15:30 ` Jeremy Fitzhardinge
2007-04-30 15:37 ` Andi Kleen
2007-04-30 10:27 ` [PATCH] [20/40] i386: Clean up paravirt patchable wrappers Andi Kleen
2007-04-30 10:27 ` [PATCH] [21/40] i386: drop unused ptep_get_and_clear Andi Kleen
2007-04-30 10:27 ` [PATCH] [22/40] x86: deflate stack usage in lib/inflate.c Andi Kleen
2007-04-30 10:27 ` [PATCH] [23/40] x86_64: deflate inflate_dynamic too Andi Kleen
2007-04-30 10:27 ` [PATCH] [24/40] i386: Page-align the GDT Andi Kleen
2007-04-30 10:27 ` [PATCH] [25/40] i386: Convert PDA into the percpu section Andi Kleen
2007-04-30 10:28 ` [PATCH] [26/40] i386: cleanups to help using per-cpu variables from asm Andi Kleen
2007-04-30 10:28 ` [PATCH] [27/40] i386: Define per_cpu_offset Andi Kleen
2007-04-30 10:28 ` [PATCH] [28/40] i386: Fix UP gdt bugs Andi Kleen
2007-04-30 10:28 ` Andi Kleen [this message]
2007-04-30 10:28 ` [PATCH] [30/40] x86: update for i386 and x86-64 check_bugs Andi Kleen
2007-04-30 10:28 ` [PATCH] [31/40] i386: In compat mode, the return value here was uninitialized Andi Kleen
2007-04-30 10:28 ` [PATCH] [32/40] i386: kRemove a warning about unused variable in !CONFIG_ACPI compilation Andi Kleen
2007-04-30 10:28 ` [PATCH] [33/40] i386: Allow boot-time disable of paravirt_ops patching Andi Kleen
2007-04-30 10:28 ` [PATCH] [34/40] i386: Clean up arch/i386/kernel/cpu/mcheck/p4.c Andi Kleen
2007-04-30 10:28 ` [PATCH] [35/40] i386: Now that the VDSO can be relocated, we can support it in VMI configurations Andi Kleen
2007-04-30 10:28 ` [PATCH] [36/40] i386: Implement vmi_kmap_atomic_pte Andi Kleen
2007-04-30 10:28 ` [PATCH] [37/40] i386: Convert VMI timer to use clock events Andi Kleen
2007-04-30 10:28 ` [PATCH] [38/40] x86: Jeremy Fitzhardinge <jeremy@goop.org> Andi Kleen
2007-04-30 10:28 ` [PATCH] [39/40] i386: Export paravirt_ops for non GPL modules too Andi Kleen
2007-04-30 10:50 ` Christoph Hellwig
2007-04-30 11:00 ` Andi Kleen
2007-04-30 11:15 ` Jan Engelhardt
2007-04-30 11:19 ` Andi Kleen
2007-04-30 12:28 ` Peter Zijlstra
2007-04-30 13:40 ` Andi Kleen
2007-04-30 20:37 ` [PATCH] [39/40] i386: Export paravirt_ops for non GPL modulestoo David Schwartz
2007-04-30 11:04 ` [PATCH] [39/40] i386: Export paravirt_ops for non GPL modules too Jan Engelhardt
2007-04-30 14:55 ` Alan Cox
2007-04-30 16:30 ` Andi Kleen
2007-04-30 21:30 ` Jeremy Fitzhardinge
2007-04-30 10:28 ` [PATCH] [40/40] i386: Clean up ELF note generation Andi Kleen
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=20070430102803.C9463151D2@wotan.suse.de \
--to=ak@suse.de \
--cc=chrisw@sous-sol.org \
--cc=ebiederm@xmission.com \
--cc=jeremy@goop.org \
--cc=linux-kernel@vger.kernel.org \
--cc=patches@x86-64.org \
--cc=torvalds@linux-foundation.org \
--cc=zach@vmware.com \
/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
Powered by JetHome