The specifications that talk about E820 map doesn't have an upper limit
on the number of E820 entries. But, today's kernel has a hard limit of 32.
With increase in memory size, we are seeing the number of E820 entries
reaching close to 32. Patch below bumps the number upto 128.
Patch does it for both i386 and x86-64.
Side-effect:
bss increases by ~ 2K and init.data increases by ~7.5K
on all systems, due to increase in size of static arrays.
Signed-off-by: Venkatesh Pallipadi <[email protected]>
diff -purN linux-2.6.11/Documentation/i386/zero-page.txt.orig linux-2.6.11/Documentation/i386/zero-page.txt
--- linux-2.6.11/Documentation/i386/zero-page.txt.orig 2005-04-21 17:09:08.059184904 -0700
+++ linux-2.6.11/Documentation/i386/zero-page.txt 2005-04-21 17:09:14.955136560 -0700
@@ -79,6 +79,6 @@ Offset Type Description
0x22c unsigned long ramdisk_max
0x230 16 bytes trampoline
0x290 - 0x2cf EDD_MBR_SIG_BUFFER (edd.S)
-0x2d0 - 0x600 E820MAP
-0x600 - 0x7ff EDDBUF (edd.S) for disk signature read sector
-0x600 - 0x7eb EDDBUF (edd.S) for edd data
+0x2d0 - 0xd00 E820MAP
+0xd00 - 0xeff EDDBUF (edd.S) for disk signature read sector
+0xd00 - 0xeeb EDDBUF (edd.S) for edd data
diff -purN linux-2.6.11/include/linux/edd.h.orig linux-2.6.11/include/linux/edd.h
--- linux-2.6.11/include/linux/edd.h.orig 2005-04-21 17:09:08.059184904 -0700
+++ linux-2.6.11/include/linux/edd.h 2005-04-21 17:09:14.956136408 -0700
@@ -32,7 +32,7 @@
#define EDDNR 0x1e9 /* addr of number of edd_info structs at EDDBUF
in boot_params - treat this as 1 byte */
-#define EDDBUF 0x600 /* addr of edd_info structs in boot_params */
+#define EDDBUF 0xd00 /* addr of edd_info structs in boot_params */
#define EDDMAXNR 6 /* number of edd_info structs starting at EDDBUF */
#define EDDEXTSIZE 8 /* change these if you muck with the structures */
#define EDDPARMSIZE 74
diff -purN linux-2.6.11/include/asm-i386/e820.h.orig linux-2.6.11/include/asm-i386/e820.h
--- linux-2.6.11/include/asm-i386/e820.h.orig 2005-04-21 17:09:08.060184752 -0700
+++ linux-2.6.11/include/asm-i386/e820.h 2005-04-21 17:09:14.956136408 -0700
@@ -13,7 +13,7 @@
#define __E820_HEADER
#define E820MAP 0x2d0 /* our map */
-#define E820MAX 32 /* number of entries in E820MAP */
+#define E820MAX 128 /* number of entries in E820MAP */
#define E820NR 0x1e8 /* # entries in E820MAP */
#define E820_RAM 1
diff -purN linux-2.6.11/include/asm-i386/setup.h.orig linux-2.6.11/include/asm-i386/setup.h
--- linux-2.6.11/include/asm-i386/setup.h.orig 2005-04-21 17:09:08.060184752 -0700
+++ linux-2.6.11/include/asm-i386/setup.h 2005-04-21 17:09:14.957136256 -0700
@@ -16,7 +16,7 @@
#define MAXMEM_PFN PFN_DOWN(MAXMEM)
#define MAX_NONPAE_PFN (1 << 20)
-#define PARAM_SIZE 2048
+#define PARAM_SIZE 4096
#define COMMAND_LINE_SIZE 256
#define OLD_CL_MAGIC_ADDR 0x90020
diff -purN linux-2.6.11/arch/i386/boot/setup.S.orig linux-2.6.11/arch/i386/boot/setup.S
--- linux-2.6.11/arch/i386/boot/setup.S.orig 2005-04-21 17:09:08.061184600 -0700
+++ linux-2.6.11/arch/i386/boot/setup.S 2005-04-21 17:09:14.957136256 -0700
@@ -164,7 +164,7 @@ ramdisk_max: .long (-__PAGE_OFFSET-(512
trampoline: call start_of_setup
.align 16
# The offset at this point is 0x240
- .space (0x7ff-0x240+1) # E820 & EDD space (ending at 0x7ff)
+ .space (0xeff-0x240+1) # E820 & EDD space (ending at 0xeff)
# End of setup header #####################################################
start_of_setup:
@@ -333,9 +333,9 @@ jmpe820:
# sizeof(e820rec).
#
good820:
- movb (E820NR), %al # up to 32 entries
+ movb (E820NR), %al # up to 128 entries
cmpb $E820MAX, %al
- jnl bail820
+ jae bail820
incb (E820NR)
movw %di, %ax
diff -purN linux-2.6.11/include/asm-x86_64/e820.h.orig linux-2.6.11/include/asm-x86_64/e820.h
--- linux-2.6.11/include/asm-x86_64/e820.h.orig 2005-04-21 17:09:08.061184600 -0700
+++ linux-2.6.11/include/asm-x86_64/e820.h 2005-04-21 17:09:14.958136104 -0700
@@ -14,7 +14,7 @@
#include <linux/mmzone.h>
#define E820MAP 0x2d0 /* our map */
-#define E820MAX 32 /* number of entries in E820MAP */
+#define E820MAX 128 /* number of entries in E820MAP */
#define E820NR 0x1e8 /* # entries in E820MAP */
#define E820_RAM 1
diff -purN linux-2.6.11/include/asm-x86_64/bootsetup.h.orig linux-2.6.11/include/asm-x86_64/bootsetup.h
--- linux-2.6.11/include/asm-x86_64/bootsetup.h.orig 2005-04-21 17:09:08.061184600 -0700
+++ linux-2.6.11/include/asm-x86_64/bootsetup.h 2005-04-22 14:38:39.793893360 -0700
@@ -2,7 +2,8 @@
#ifndef _X86_64_BOOTSETUP_H
#define _X86_64_BOOTSETUP_H 1
-extern char x86_boot_params[2048];
+#define BOOT_PARAM_SIZE 4096
+extern char x86_boot_params[BOOT_PARAM_SIZE];
/*
* This is set up by the setup-routine at boot-time
diff -purN linux-2.6.11/arch/x86_64/boot/setup.S.orig linux-2.6.11/arch/x86_64/boot/setup.S
--- linux-2.6.11/arch/x86_64/boot/setup.S.orig 2005-04-21 17:09:08.062184448 -0700
+++ linux-2.6.11/arch/x86_64/boot/setup.S 2005-04-22 15:32:57.074711408 -0700
@@ -160,7 +160,7 @@ ramdisk_max: .long 0xffffffff
trampoline: call start_of_setup
.align 16
# The offset at this point is 0x240
- .space (0x7ff-0x240+1) # E820 & EDD space (ending at 0x7ff)
+ .space (0xeff-0x240+1) # E820 & EDD space (ending at 0xeff)
# End of setup header #####################################################
start_of_setup:
@@ -412,9 +412,9 @@ jmpe820:
# sizeof(e820rec).
#
good820:
- movb (E820NR), %al # up to 32 entries
+ movb (E820NR), %al # up to 128 entries
cmpb $E820MAX, %al
- jnl bail820
+ jae bail820
incb (E820NR)
movw %di, %ax
diff -purN linux-2.6.11/arch/x86_64/kernel/setup64.c.orig linux-2.6.11/arch/x86_64/kernel/setup64.c
--- linux-2.6.11/arch/x86_64/kernel/setup64.c.orig 2005-04-22 15:29:38.304929008 -0700
+++ linux-2.6.11/arch/x86_64/kernel/setup64.c 2005-04-21 17:12:15.529685056 -0700
@@ -12,6 +12,7 @@
#include <linux/string.h>
#include <linux/bootmem.h>
#include <linux/bitops.h>
+#include <asm/bootsetup.h>
#include <asm/pda.h>
#include <asm/pgtable.h>
#include <asm/processor.h>
@@ -26,7 +27,7 @@
#include <asm/mman.h>
#include <asm/numa.h>
-char x86_boot_params[2048] __initdata = {0,};
+char x86_boot_params[BOOT_PARAM_SIZE] __initdata = {0,};
cpumask_t cpu_initialized __initdata = CPU_MASK_NONE;
diff -purN linux-2.6.11/arch/x86_64/kernel/head64.c.orig linux-2.6.11/arch/x86_64/kernel/head64.c
--- linux-2.6.11/arch/x86_64/kernel/head64.c.orig 2005-04-22 15:29:22.004407064 -0700
+++ linux-2.6.11/arch/x86_64/kernel/head64.c 2005-04-22 15:32:37.295718272 -0700
@@ -29,8 +29,6 @@ static void __init clear_bss(void)
(unsigned long) __bss_end - (unsigned long) __bss_start);
}
-extern char x86_boot_params[2048];
-
#define NEW_CL_POINTER 0x228 /* Relative to real mode data */
#define OLD_CL_MAGIC_ADDR 0x90020
#define OLD_CL_MAGIC 0xA33F
@@ -44,7 +42,7 @@ static void __init copy_bootdata(char *r
int new_data;
char * command_line;
- memcpy(x86_boot_params, real_mode_data, 2048);
+ memcpy(x86_boot_params, real_mode_data, BOOT_PARAM_SIZE);
new_data = *(int *) (x86_boot_params + NEW_CL_POINTER);
if (!new_data) {
if (OLD_CL_MAGIC != * (u16 *) OLD_CL_MAGIC_ADDR) {
This is a very good idea.
Jeff
Venkatesh Pallipadi wrote:
>The specifications that talk about E820 map doesn't have an upper limit
>on the number of E820 entries. But, today's kernel has a hard limit of 32.
>With increase in memory size, we are seeing the number of E820 entries
>reaching close to 32. Patch below bumps the number upto 128.
>
>Patch does it for both i386 and x86-64.
>
>Side-effect:
>bss increases by ~ 2K and init.data increases by ~7.5K
>on all systems, due to increase in size of static arrays.
>
>Signed-off-by: Venkatesh Pallipadi <[email protected]>
>
>diff -purN linux-2.6.11/Documentation/i386/zero-page.txt.orig linux-2.6.11/Documentation/i386/zero-page.txt
>--- linux-2.6.11/Documentation/i386/zero-page.txt.orig 2005-04-21 17:09:08.059184904 -0700
>+++ linux-2.6.11/Documentation/i386/zero-page.txt 2005-04-21 17:09:14.955136560 -0700
>@@ -79,6 +79,6 @@ Offset Type Description
> 0x22c unsigned long ramdisk_max
> 0x230 16 bytes trampoline
> 0x290 - 0x2cf EDD_MBR_SIG_BUFFER (edd.S)
>-0x2d0 - 0x600 E820MAP
>-0x600 - 0x7ff EDDBUF (edd.S) for disk signature read sector
>-0x600 - 0x7eb EDDBUF (edd.S) for edd data
>+0x2d0 - 0xd00 E820MAP
>+0xd00 - 0xeff EDDBUF (edd.S) for disk signature read sector
>+0xd00 - 0xeeb EDDBUF (edd.S) for edd data
>diff -purN linux-2.6.11/include/linux/edd.h.orig linux-2.6.11/include/linux/edd.h
>--- linux-2.6.11/include/linux/edd.h.orig 2005-04-21 17:09:08.059184904 -0700
>+++ linux-2.6.11/include/linux/edd.h 2005-04-21 17:09:14.956136408 -0700
>@@ -32,7 +32,7 @@
>
> #define EDDNR 0x1e9 /* addr of number of edd_info structs at EDDBUF
> in boot_params - treat this as 1 byte */
>-#define EDDBUF 0x600 /* addr of edd_info structs in boot_params */
>+#define EDDBUF 0xd00 /* addr of edd_info structs in boot_params */
> #define EDDMAXNR 6 /* number of edd_info structs starting at EDDBUF */
> #define EDDEXTSIZE 8 /* change these if you muck with the structures */
> #define EDDPARMSIZE 74
>diff -purN linux-2.6.11/include/asm-i386/e820.h.orig linux-2.6.11/include/asm-i386/e820.h
>--- linux-2.6.11/include/asm-i386/e820.h.orig 2005-04-21 17:09:08.060184752 -0700
>+++ linux-2.6.11/include/asm-i386/e820.h 2005-04-21 17:09:14.956136408 -0700
>@@ -13,7 +13,7 @@
> #define __E820_HEADER
>
> #define E820MAP 0x2d0 /* our map */
>-#define E820MAX 32 /* number of entries in E820MAP */
>+#define E820MAX 128 /* number of entries in E820MAP */
> #define E820NR 0x1e8 /* # entries in E820MAP */
>
> #define E820_RAM 1
>diff -purN linux-2.6.11/include/asm-i386/setup.h.orig linux-2.6.11/include/asm-i386/setup.h
>--- linux-2.6.11/include/asm-i386/setup.h.orig 2005-04-21 17:09:08.060184752 -0700
>+++ linux-2.6.11/include/asm-i386/setup.h 2005-04-21 17:09:14.957136256 -0700
>@@ -16,7 +16,7 @@
> #define MAXMEM_PFN PFN_DOWN(MAXMEM)
> #define MAX_NONPAE_PFN (1 << 20)
>
>-#define PARAM_SIZE 2048
>+#define PARAM_SIZE 4096
> #define COMMAND_LINE_SIZE 256
>
> #define OLD_CL_MAGIC_ADDR 0x90020
>diff -purN linux-2.6.11/arch/i386/boot/setup.S.orig linux-2.6.11/arch/i386/boot/setup.S
>--- linux-2.6.11/arch/i386/boot/setup.S.orig 2005-04-21 17:09:08.061184600 -0700
>+++ linux-2.6.11/arch/i386/boot/setup.S 2005-04-21 17:09:14.957136256 -0700
>@@ -164,7 +164,7 @@ ramdisk_max: .long (-__PAGE_OFFSET-(512
> trampoline: call start_of_setup
> .align 16
> # The offset at this point is 0x240
>- .space (0x7ff-0x240+1) # E820 & EDD space (ending at 0x7ff)
>+ .space (0xeff-0x240+1) # E820 & EDD space (ending at 0xeff)
> # End of setup header #####################################################
>
> start_of_setup:
>@@ -333,9 +333,9 @@ jmpe820:
> # sizeof(e820rec).
> #
> good820:
>- movb (E820NR), %al # up to 32 entries
>+ movb (E820NR), %al # up to 128 entries
> cmpb $E820MAX, %al
>- jnl bail820
>+ jae bail820
>
> incb (E820NR)
> movw %di, %ax
>diff -purN linux-2.6.11/include/asm-x86_64/e820.h.orig linux-2.6.11/include/asm-x86_64/e820.h
>--- linux-2.6.11/include/asm-x86_64/e820.h.orig 2005-04-21 17:09:08.061184600 -0700
>+++ linux-2.6.11/include/asm-x86_64/e820.h 2005-04-21 17:09:14.958136104 -0700
>@@ -14,7 +14,7 @@
> #include <linux/mmzone.h>
>
> #define E820MAP 0x2d0 /* our map */
>-#define E820MAX 32 /* number of entries in E820MAP */
>+#define E820MAX 128 /* number of entries in E820MAP */
> #define E820NR 0x1e8 /* # entries in E820MAP */
>
> #define E820_RAM 1
>diff -purN linux-2.6.11/include/asm-x86_64/bootsetup.h.orig linux-2.6.11/include/asm-x86_64/bootsetup.h
>--- linux-2.6.11/include/asm-x86_64/bootsetup.h.orig 2005-04-21 17:09:08.061184600 -0700
>+++ linux-2.6.11/include/asm-x86_64/bootsetup.h 2005-04-22 14:38:39.793893360 -0700
>@@ -2,7 +2,8 @@
> #ifndef _X86_64_BOOTSETUP_H
> #define _X86_64_BOOTSETUP_H 1
>
>-extern char x86_boot_params[2048];
>+#define BOOT_PARAM_SIZE 4096
>+extern char x86_boot_params[BOOT_PARAM_SIZE];
>
> /*
> * This is set up by the setup-routine at boot-time
>diff -purN linux-2.6.11/arch/x86_64/boot/setup.S.orig linux-2.6.11/arch/x86_64/boot/setup.S
>--- linux-2.6.11/arch/x86_64/boot/setup.S.orig 2005-04-21 17:09:08.062184448 -0700
>+++ linux-2.6.11/arch/x86_64/boot/setup.S 2005-04-22 15:32:57.074711408 -0700
>@@ -160,7 +160,7 @@ ramdisk_max: .long 0xffffffff
> trampoline: call start_of_setup
> .align 16
> # The offset at this point is 0x240
>- .space (0x7ff-0x240+1) # E820 & EDD space (ending at 0x7ff)
>+ .space (0xeff-0x240+1) # E820 & EDD space (ending at 0xeff)
> # End of setup header #####################################################
>
> start_of_setup:
>@@ -412,9 +412,9 @@ jmpe820:
> # sizeof(e820rec).
> #
> good820:
>- movb (E820NR), %al # up to 32 entries
>+ movb (E820NR), %al # up to 128 entries
> cmpb $E820MAX, %al
>- jnl bail820
>+ jae bail820
>
> incb (E820NR)
> movw %di, %ax
>diff -purN linux-2.6.11/arch/x86_64/kernel/setup64.c.orig linux-2.6.11/arch/x86_64/kernel/setup64.c
>--- linux-2.6.11/arch/x86_64/kernel/setup64.c.orig 2005-04-22 15:29:38.304929008 -0700
>+++ linux-2.6.11/arch/x86_64/kernel/setup64.c 2005-04-21 17:12:15.529685056 -0700
>@@ -12,6 +12,7 @@
> #include <linux/string.h>
> #include <linux/bootmem.h>
> #include <linux/bitops.h>
>+#include <asm/bootsetup.h>
> #include <asm/pda.h>
> #include <asm/pgtable.h>
> #include <asm/processor.h>
>@@ -26,7 +27,7 @@
> #include <asm/mman.h>
> #include <asm/numa.h>
>
>-char x86_boot_params[2048] __initdata = {0,};
>+char x86_boot_params[BOOT_PARAM_SIZE] __initdata = {0,};
>
> cpumask_t cpu_initialized __initdata = CPU_MASK_NONE;
>
>diff -purN linux-2.6.11/arch/x86_64/kernel/head64.c.orig linux-2.6.11/arch/x86_64/kernel/head64.c
>--- linux-2.6.11/arch/x86_64/kernel/head64.c.orig 2005-04-22 15:29:22.004407064 -0700
>+++ linux-2.6.11/arch/x86_64/kernel/head64.c 2005-04-22 15:32:37.295718272 -0700
>@@ -29,8 +29,6 @@ static void __init clear_bss(void)
> (unsigned long) __bss_end - (unsigned long) __bss_start);
> }
>
>-extern char x86_boot_params[2048];
>-
> #define NEW_CL_POINTER 0x228 /* Relative to real mode data */
> #define OLD_CL_MAGIC_ADDR 0x90020
> #define OLD_CL_MAGIC 0xA33F
>@@ -44,7 +42,7 @@ static void __init copy_bootdata(char *r
> int new_data;
> char * command_line;
>
>- memcpy(x86_boot_params, real_mode_data, 2048);
>+ memcpy(x86_boot_params, real_mode_data, BOOT_PARAM_SIZE);
> new_data = *(int *) (x86_boot_params + NEW_CL_POINTER);
> if (!new_data) {
> if (OLD_CL_MAGIC != * (u16 *) OLD_CL_MAGIC_ADDR) {
>-
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to [email protected]
>More majordomo info at http://vger.kernel.org/majordomo-info.html
>Please read the FAQ at http://www.tux.org/lkml/
>
>
>
On Fri, 22 Apr 2005, Venkatesh Pallipadi wrote:
>
> The specifications that talk about E820 map doesn't have an upper limit
> on the number of E820 entries. But, today's kernel has a hard limit of 32.
> With increase in memory size, we are seeing the number of E820 entries
> reaching close to 32. Patch below bumps the number upto 128.
Hmm. Anything that changes setup.S tends to have bootloader dependencies.
I worry whether this one does too..
Linus
On Fri, Apr 22, 2005 at 06:51:59PM -0700, Linus Torvalds wrote:
> On Fri, 22 Apr 2005, Venkatesh Pallipadi wrote:
> > The specifications that talk about E820 map doesn't have an upper limit
> > on the number of E820 entries. But, today's kernel has a hard limit of 32.
> > With increase in memory size, we are seeing the number of E820 entries
> > reaching close to 32. Patch below bumps the number upto 128.
>
> Hmm. Anything that changes setup.S tends to have bootloader dependencies.
> I worry whether this one does too..
>
The setup.S change in this patch should be OK. As it is adding to the
existing zero-page and keeping it within one page. I tested it on systems
with grub, adding some dummy E820 entries and it worked fine.
However, there is another place that needs to be changed. Boot loaders
also calls E820 while booting directly with vmlinux (instead of usual
bzImage - which is handled by this patch) and that needs to change to
incorporate more E820 entries. But, there we may need more changes, to
the boot protocol version and the like. On a side note, looking at the
grub source, it seems to have a limit of 50 entries today, which doesn't
agree with current 32 entry limit in the kernel. Not sure why grub has
this different limit though.
Thanks,
Venki
On Fri, Apr 22, 2005 at 07:32:50PM -0700, Venkatesh Pallipadi wrote:
> On Fri, Apr 22, 2005 at 06:51:59PM -0700, Linus Torvalds wrote:
> > On Fri, 22 Apr 2005, Venkatesh Pallipadi wrote:
> > > The specifications that talk about E820 map doesn't have an upper limit
> > > on the number of E820 entries. But, today's kernel has a hard limit of 32.
> > > With increase in memory size, we are seeing the number of E820 entries
> > > reaching close to 32. Patch below bumps the number upto 128.
> >
> > Hmm. Anything that changes setup.S tends to have bootloader dependencies.
> > I worry whether this one does too..
> >
>
> The setup.S change in this patch should be OK. As it is adding to the
> existing zero-page and keeping it within one page. I tested it on systems
> with grub, adding some dummy E820 entries and it worked fine.
The last time I tried to extend the zero page (with a longer command line)
it broke lilo on systems with EDID support and CONFIG_EDID enabled.
Make sure you test that case.
-Andi
On Sat, Apr 23, 2005 at 05:10:48PM +0200, Andi Kleen wrote:
> On Fri, Apr 22, 2005 at 07:32:50PM -0700, Venkatesh Pallipadi wrote:
> > On Fri, Apr 22, 2005 at 06:51:59PM -0700, Linus Torvalds wrote:
> > > On Fri, 22 Apr 2005, Venkatesh Pallipadi wrote:
> > > > The specifications that talk about E820 map doesn't have an upper limit
> > > > on the number of E820 entries. But, today's kernel has a hard limit of 32.
> > > > With increase in memory size, we are seeing the number of E820 entries
> > > > reaching close to 32. Patch below bumps the number upto 128.
> > >
> > > Hmm. Anything that changes setup.S tends to have bootloader dependencies.
> > > I worry whether this one does too..
> > >
> >
> > The setup.S change in this patch should be OK. As it is adding to the
> > existing zero-page and keeping it within one page. I tested it on systems
> > with grub, adding some dummy E820 entries and it worked fine.
>
> The last time I tried to extend the zero page (with a longer command line)
> it broke lilo on systems with EDID support and CONFIG_EDID enabled.
> Make sure you test that case.
>
Tested this patch with some more configuration and I did not see any breakage.
- LILO with EDID enabled
- pxeboot
And in the current zero-page, EDID info is at a lower address (before E820MAP).
So, there should not be any issues with EDID info. Only field (other than E820)
that is changing in zero page is EDDBUF (that comes after E820MAP).
The patch changes the reference to EDDBUF inside kernel to new position in
zero page. And I don't see EDDBUF being used by boot loader anywhere. So, we
should be OK with that change.
Thanks,
Venki