[PATCH 0/8] x86, boot: clean up setup_data handling

[PATCH 0/8] x86, boot: clean up setup_data handling

[Yinghai Lu]

Now we setup_data is reserved via memblock and e820 and different
handlers have different ways, and it is confusing.
1. SETUP_E820_EXT: is consumed early and will not copy or access again.
	have memory wasted.
2. SETUP_EFI: is accessed via ioremap every time at early stage.
	have memory wasted.
3. SETUP_DTB: is copied locally.
	have memory wasted.
4. SETUP_PCI: is accessed via ioremap for every pci devices, even run-time.
5. SETUP_KASLR: is accessed early, will not copy or access again.
	have memory wasted.

Also setup_data is exported to debugfs for debug purpose.

Here will convert to let every handler to decide how to handle it.
and will not reserve the setup_data generally, so will not
waste memory and also make memblock/e820 keep page aligned.
1. not touch E820 anymore.
2. copy SETUP_EFI to __initdata variable and access it without ioremap.
3. SETUP_DTB: reserver and copy to local and free.
4. SETUP_PCI: reverve localy and convert to list, to avoid keeping ioremap.
5. SETUP_KASLR: fix accessing kaslr_enabled accessing...
6. export SETUP_PCI via sysfs.

Those patches could be applied on top of tip/x86/urgent with SETUP_KASLR
support.

Yinghai Lu (8):
  x86, kaslr: get kaslr_enabled back correctly
  x86: Kill E820_RESERVED_KERN
  x86, efi: copy SETUP_EFI data and access directly
  x86, of: let add_dtb reserve by itself
  x86, boot: Add add_pci handler for SETUP_PCI
  x86: kill not used setup_data handling code
  x86, pci: convert SETUP_PCI data to list
  x86, pci: export SETUP_PCI data via sysfs

 arch/x86/include/asm/efi.h       |   2 +-
 arch/x86/include/asm/pci.h       |   2 +
 arch/x86/include/asm/prom.h      |   9 +-
 arch/x86/include/uapi/asm/e820.h |   9 --
 arch/x86/kernel/devicetree.c     |  39 +++---
 arch/x86/kernel/e820.c           |   6 +-
 arch/x86/kernel/kdebugfs.c       | 142 --------------------
 arch/x86/kernel/setup.c          |  57 ++------
 arch/x86/kernel/tboot.c          |   3 +-
 arch/x86/mm/init_64.c            |  11 +-
 arch/x86/pci/common.c            | 281 ++++++++++++++++++++++++++++++++++++---
 arch/x86/platform/efi/efi.c      |  13 +-
 arch/x86/platform/efi/efi_64.c   |  13 +-
 arch/x86/platform/efi/quirks.c   |  23 +---
 14 files changed, 336 insertions(+), 274 deletions(-)

-- 
1.8.4.5

[PATCH 0/8] x86, boot: clean up setup_data handling

[Yinghai Lu]

Now we setup_data is reserved via memblock and e820 and different
handlers have different ways, and it is confusing.
1. SETUP_E820_EXT: is consumed early and will not copy or access again.
	have memory wasted.
2. SETUP_EFI: is accessed via ioremap every time at early stage.
	have memory wasted.
3. SETUP_DTB: is copied locally.
	have memory wasted.
4. SETUP_PCI: is accessed via ioremap for every pci devices, even run-time.
5. SETUP_KASLR: is accessed early, will not copy or access again.
	have memory wasted.

Also setup_data is exported to debugfs for debug purpose.

Here will convert to let every handler to decide how to handle it.
and will not reserve the setup_data generally, so will not
waste memory and also make memblock/e820 keep page aligned.
1. not touch E820 anymore.
2. copy SETUP_EFI to __initdata variable and access it without ioremap.
3. SETUP_DTB: reserver and copy to local and free.
4. SETUP_PCI: reverve localy and convert to list, to avoid keeping ioremap.
5. SETUP_KASLR: fix accessing kaslr_enabled accessing...
6. export SETUP_PCI via sysfs.

Those patches could be applied on top of tip/x86/urgent with SETUP_KASLR
support.

Yinghai Lu (8):
  x86, kaslr: get kaslr_enabled back correctly
  x86: Kill E820_RESERVED_KERN
  x86, efi: copy SETUP_EFI data and access directly
  x86, of: let add_dtb reserve by itself
  x86, boot: Add add_pci handler for SETUP_PCI
  x86: kill not used setup_data handling code
  x86, pci: convert SETUP_PCI data to list
  x86, pci: export SETUP_PCI data via sysfs

 arch/x86/include/asm/efi.h       |   2 +-
 arch/x86/include/asm/pci.h       |   2 +
 arch/x86/include/asm/prom.h      |   9 +-
 arch/x86/include/uapi/asm/e820.h |   9 --
 arch/x86/kernel/devicetree.c     |  39 +++---
 arch/x86/kernel/e820.c           |   6 +-
 arch/x86/kernel/kdebugfs.c       | 142 --------------------
 arch/x86/kernel/setup.c          |  57 ++------
 arch/x86/kernel/tboot.c          |   3 +-
 arch/x86/mm/init_64.c            |  11 +-
 arch/x86/pci/common.c            | 281 ++++++++++++++++++++++++++++++++++++---
 arch/x86/platform/efi/efi.c      |  13 +-
 arch/x86/platform/efi/efi_64.c   |  13 +-
 arch/x86/platform/efi/quirks.c   |  23 +---
 14 files changed, 336 insertions(+), 274 deletions(-)

-- 
1.8.4.5

[PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

We should access variable with referrence instead of using physical
address as value.

Cc: Matt Fleming <matt.fleming@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
---
 arch/x86/kernel/setup.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 98dc931..05d444f 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -429,7 +429,13 @@ static void __init reserve_initrd(void)
 
 static void __init parse_kaslr_setup(u64 pa_data, u32 data_len)
 {
-	kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));
+	/* kaslr_setup_data is defined in aslr.c */
+	unsigned char *data;
+	unsigned long offset = sizeof(struct setup_data);
+
+	data = early_memremap(pa_data, offset + 1);
+	kaslr_enabled = *(data + offset);
+	early_memunmap(data, offset + 1);
 }
 
 static void __init parse_setup_data(void)
-- 
1.8.4.5

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

On Sat, Feb 28, 2015 at 6:17 PM, Yinghai Lu <yinghai@kernel.org> wrote:
> We should access variable with referrence instead of using physical
> address as value.
>
> Cc: Matt Fleming <matt.fleming@intel.com>
> Cc: Borislav Petkov <bp@suse.de>
> Signed-off-by: Yinghai Lu <yinghai@kernel.org>
> ---
>  arch/x86/kernel/setup.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index 98dc931..05d444f 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -429,7 +429,13 @@ static void __init reserve_initrd(void)
>
>  static void __init parse_kaslr_setup(u64 pa_data, u32 data_len)
>  {
> -       kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));
> +       /* kaslr_setup_data is defined in aslr.c */
> +       unsigned char *data;
> +       unsigned long offset = sizeof(struct setup_data);
> +
> +       data = early_memremap(pa_data, offset + 1);
> +       kaslr_enabled = *(data + offset);
> +       early_memunmap(data, offset + 1);
>  }
>
>  static void __init parse_setup_data(void)
> --

oh, no. the offending commit already got into linus tree.

commit f47233c2d34f243ecdaac179c3408a39ff9216a7
Author: Jiri Kosina <jkosina@suse.cz>
Date:   Fri Feb 13 16:04:55 2015 +0100

    x86/mm/ASLR: Propagate base load address calculation

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

On Sat, Feb 28, 2015 at 6:17 PM, Yinghai Lu <yinghai-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote:
> We should access variable with referrence instead of using physical
> address as value.
>
> Cc: Matt Fleming <matt.fleming-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
> Cc: Borislav Petkov <bp-l3A5Bk7waGM@public.gmane.org>
> Signed-off-by: Yinghai Lu <yinghai-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> ---
>  arch/x86/kernel/setup.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index 98dc931..05d444f 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -429,7 +429,13 @@ static void __init reserve_initrd(void)
>
>  static void __init parse_kaslr_setup(u64 pa_data, u32 data_len)
>  {
> -       kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));
> +       /* kaslr_setup_data is defined in aslr.c */
> +       unsigned char *data;
> +       unsigned long offset = sizeof(struct setup_data);
> +
> +       data = early_memremap(pa_data, offset + 1);
> +       kaslr_enabled = *(data + offset);
> +       early_memunmap(data, offset + 1);
>  }
>
>  static void __init parse_setup_data(void)
> --

oh, no. the offending commit already got into linus tree.

commit f47233c2d34f243ecdaac179c3408a39ff9216a7
Author: Jiri Kosina <jkosina-AlSwsSmVLrQ@public.gmane.org>
Date:   Fri Feb 13 16:04:55 2015 +0100

    x86/mm/ASLR: Propagate base load address calculation

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Borislav Petkov]

On Sat, Feb 28, 2015 at 06:40:39PM -0800, Yinghai Lu wrote:
> oh, no. the offending commit already got into linus tree.

We're working on it, follow this thread:

http://lkml.kernel.org/r/1424929021.10337.24.camel@intel.com

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Ingo Molnar]

(Cc:-ed Jiri and Kees as well.)

* Borislav Petkov <bp@suse.de> wrote:

> On Sat, Feb 28, 2015 at 06:40:39PM -0800, Yinghai Lu wrote:
> >
> > oh, no. the offending commit already got into linus tree.
> 
> We're working on it, follow this thread:
> 
> http://lkml.kernel.org/r/1424929021.10337.24.camel@intel.com

I think that's a different bug.

parse_kaslr_setup() is simply bogus, it does:

        kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));

which makes no sense whatsoever: it randomly enables (or disables, 
depending on the physical address of the setup page) KASLR when it 
meets a SETUP_KASLR record.

So I'm inclined to apply Yinghai's fix, with a better changelog that 
explains what happened ...

Agreed?

Thanks,

	Ingo

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Ingo Molnar]

(Cc:-ed Jiri and Kees as well.)

* Borislav Petkov <bp@suse.de> wrote:

> On Sat, Feb 28, 2015 at 06:40:39PM -0800, Yinghai Lu wrote:
> >
> > oh, no. the offending commit already got into linus tree.
> 
> We're working on it, follow this thread:
> 
> http://lkml.kernel.org/r/1424929021.10337.24.camel@intel.com

I think that's a different bug.

parse_kaslr_setup() is simply bogus, it does:

        kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));

which makes no sense whatsoever: it randomly enables (or disables, 
depending on the physical address of the setup page) KASLR when it 
meets a SETUP_KASLR record.

So I'm inclined to apply Yinghai's fix, with a better changelog that 
explains what happened ...

Agreed?

Thanks,

	Ingo

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

On Sun, Mar 1, 2015 at 7:23 AM, Ingo Molnar <mingo@kernel.org> wrote:
>
> I think that's a different bug.
>
> parse_kaslr_setup() is simply bogus, it does:
>
>         kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));
>
> which makes no sense whatsoever: it randomly enables (or disables,
> depending on the physical address of the setup page) KASLR when it
> meets a SETUP_KASLR record.
>

Yes, you are right.

there are 8 patches in this series.
  x86, kaslr: get kaslr_enabled back correctly
  x86: Kill E820_RESERVED_KERN
  x86, efi: copy SETUP_EFI data and access directly
  x86, of: let add_dtb reserve by itself
  x86, boot: Add add_pci handler for SETUP_PCI
  x86: kill not used setup_data handling code
  x86, pci: convert SETUP_PCI data to list
  x86, pci: export SETUP_PCI data via sysfs

other 7 should also address the problem in
   http://lkml.kernel.org/r/1424929021.10337.24.camel@intel.com

the root cause is E820_RESERVED_KERN and setup_data handling
will make E820 are not page aligned anymore.

We should change setup_data handling.

Please check updated version at

git://git.kernel.org/pub/scm/linux/kernel/git/yinghai/linux-yinghai.git
for-x86-4.0

(with a fix for config without CONFIG_PCI).

Thanks

Yinghai

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Borislav Petkov]

On Sun, Mar 01, 2015 at 11:27:48AM -0800, Yinghai Lu wrote:
> other 7 should also address the problem in
>    http://lkml.kernel.org/r/1424929021.10337.24.camel@intel.com

No, they don't:

[    0.000000] parse_setup_data: data: 0x2206e50 (va: ffffffffff200e50) { next: 0x0, type: 0x0, len: 16, data[0]: 0x0 }
[    0.000000] Unknown setup_data type: 0 ignored!

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

On Sun, Mar 1, 2015 at 11:49 AM, Borislav Petkov <bp@suse.de> wrote:
> On Sun, Mar 01, 2015 at 11:27:48AM -0800, Yinghai Lu wrote:
>> other 7 should also address the problem in
>>    http://lkml.kernel.org/r/1424929021.10337.24.camel@intel.com
>
> No, they don't:
>
> [    0.000000] parse_setup_data: data: 0x2206e50 (va: ffffffffff200e50) { next: 0x0, type: 0x0, len: 16, data[0]: 0x0 }
> [    0.000000] Unknown setup_data type: 0 ignored!
>
> --

ok, i get your point.

static allocation in misc.c can not be used to kernel/head_64.S stage safely.

Thanks

Yinghai

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

On Sun, Mar 1, 2015 at 11:49 AM, Borislav Petkov <bp-l3A5Bk7waGM@public.gmane.org> wrote:
> On Sun, Mar 01, 2015 at 11:27:48AM -0800, Yinghai Lu wrote:
>> other 7 should also address the problem in
>>    http://lkml.kernel.org/r/1424929021.10337.24.camel-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org
>
> No, they don't:
>
> [    0.000000] parse_setup_data: data: 0x2206e50 (va: ffffffffff200e50) { next: 0x0, type: 0x0, len: 16, data[0]: 0x0 }
> [    0.000000] Unknown setup_data type: 0 ignored!
>
> --

ok, i get your point.

static allocation in misc.c can not be used to kernel/head_64.S stage safely.

Thanks

Yinghai

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Borislav Petkov]

On Sun, Mar 01, 2015 at 12:24:08PM -0800, Yinghai Lu wrote:
> static allocation in misc.c can not be used to kernel/head_64.S stage safely.

Correct. One possibility that works is sticking it right below
LOAD_PHYSICAL_ADDR:

+static void add_kaslr_setup_data(struct boot_params *params,
+				 u8 *output, __u8 enabled)
 {
+	/*
+	 * Stick it right under LOAD_PHYSICAL_ADDR
+	 */
+	ksd = (struct setup_data *)(output - 2 * sizeof(struct setup_data));

output is LOAD_PHYSICAL_ADDR AFAICT and is the minimum kASLR considers
so right below it should work:

https://lkml.kernel.org/r/20150228105049.GA11038@pd.tnic

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Borislav Petkov]

On Sun, Mar 01, 2015 at 12:24:08PM -0800, Yinghai Lu wrote:
> static allocation in misc.c can not be used to kernel/head_64.S stage safely.

Correct. One possibility that works is sticking it right below
LOAD_PHYSICAL_ADDR:

+static void add_kaslr_setup_data(struct boot_params *params,
+				 u8 *output, __u8 enabled)
 {
+	/*
+	 * Stick it right under LOAD_PHYSICAL_ADDR
+	 */
+	ksd = (struct setup_data *)(output - 2 * sizeof(struct setup_data));

output is LOAD_PHYSICAL_ADDR AFAICT and is the minimum kASLR considers
so right below it should work:

https://lkml.kernel.org/r/20150228105049.GA11038-fF5Pk5pvG8Y@public.gmane.org

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

On Sun, Mar 1, 2015 at 12:29 PM, Borislav Petkov <bp@suse.de> wrote:
> On Sun, Mar 01, 2015 at 12:24:08PM -0800, Yinghai Lu wrote:
>> static allocation in misc.c can not be used to kernel/head_64.S stage safely.
>
> Correct. One possibility that works is sticking it right below
> LOAD_PHYSICAL_ADDR:
>
> +static void add_kaslr_setup_data(struct boot_params *params,
> +                                u8 *output, __u8 enabled)
>  {
> +       /*
> +        * Stick it right under LOAD_PHYSICAL_ADDR
> +        */
> +       ksd = (struct setup_data *)(output - 2 * sizeof(struct setup_data));
>
> output is LOAD_PHYSICAL_ADDR AFAICT and is the minimum kASLR considers
> so right below it should work:

Does not look safe yet...

Let me think about it more.

Thanks

Yinghai

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Borislav Petkov]

On Sun, Mar 01, 2015 at 12:41:10PM -0800, Yinghai Lu wrote:
> Does not look safe yet...

Why?

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Borislav Petkov]

On Sun, Mar 01, 2015 at 12:41:10PM -0800, Yinghai Lu wrote:
> Does not look safe yet...

Why?

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

[-- Attachment #1: Type: text/plain, Size: 293 bytes --]

On Mon, Mar 2, 2015 at 12:56 AM, Borislav Petkov <bp@suse.de> wrote:
> On Sun, Mar 01, 2015 at 12:41:10PM -0800, Yinghai Lu wrote:
>> Does not look safe yet...
>
> Why?

We can not assume that range is safe to use.

Please check attach one that should fix the problem really.

Thanks

Yinghai

[-- Attachment #2: fix_decompress_offset.patch --]
[-- Type: text/x-patch, Size: 5440 bytes --]

Subject: [PATCH] x86, boot: keep data from boot stage to kernel stage.

Need to
1. move compressed kernel close the end of buffer instead of middle of it.
2. use BRK to extend init_size so no one from kernel bss and brk will
   touch the data region from boot/compressed/misc.c

Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 arch/x86/boot/compressed/head_32.S |   13 +++++++++++--
 arch/x86/boot/compressed/head_64.S |   10 ++++++++--
 arch/x86/boot/compressed/misc.c    |    4 ++++
 arch/x86/boot/compressed/mkpiggy.c |    3 ---
 arch/x86/include/asm/boot.h        |    2 ++
 arch/x86/kernel/asm-offsets.c      |    1 +
 arch/x86/kernel/setup.c            |    3 +++
 7 files changed, 29 insertions(+), 7 deletions(-)

Index: linux-2.6/arch/x86/boot/compressed/head_64.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_64.S
+++ linux-2.6/arch/x86/boot/compressed/head_64.S
@@ -102,7 +102,10 @@ ENTRY(startup_32)
 1:
 
 	/* Target address to relocate to for decompression */
-	addl	$z_extract_offset, %ebx
+	movl	BP_init_size(%esi), %eax
+	subl	$_end, %eax
+	andl	$(~(4096-1)), %eax
+	addl	%eax, %ebx
 
 /*
  * Prepare for entering 64 bit mode
@@ -330,7 +333,10 @@ preferred_addr:
 1:
 
 	/* Target address to relocate to for decompression */
-	leaq	z_extract_offset(%rbp), %rbx
+	movl	BP_init_size(%rsi), %ebx
+	subl	$_end, %ebx
+	andl	$(~(4096-1)), %ebx
+	addq	%rbp, %rbx
 
 	/* Set up the stack */
 	leaq	boot_stack_end(%rbx), %rsp
Index: linux-2.6/arch/x86/kernel/asm-offsets.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/asm-offsets.c
+++ linux-2.6/arch/x86/kernel/asm-offsets.c
@@ -66,6 +66,7 @@ void common(void) {
 	OFFSET(BP_hardware_subarch, boot_params, hdr.hardware_subarch);
 	OFFSET(BP_version, boot_params, hdr.version);
 	OFFSET(BP_kernel_alignment, boot_params, hdr.kernel_alignment);
+	OFFSET(BP_init_size, boot_params, hdr.init_size);
 	OFFSET(BP_pref_address, boot_params, hdr.pref_address);
 	OFFSET(BP_code32_start, boot_params, hdr.code32_start);
 
Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ linux-2.6/arch/x86/kernel/setup.c
@@ -111,6 +111,7 @@
 #include <asm/mce.h>
 #include <asm/alternative.h>
 #include <asm/prom.h>
+#include <asm/boot.h>
 
 /*
  * max_low_pfn_mapped: highest direct mapped pfn under 4GB
@@ -144,6 +145,8 @@ int default_check_phys_apicid_present(in
 }
 #endif
 
+RESERVE_BRK(addon_zo, BOOT_ADDON_ZO_SIZE);
+
 struct boot_params boot_params;
 
 /*
Index: linux-2.6/arch/x86/boot/compressed/head_32.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_32.S
+++ linux-2.6/arch/x86/boot/compressed/head_32.S
@@ -148,7 +148,10 @@ preferred_addr:
 1:
 
 	/* Target address to relocate to for decompression */
-	addl	$z_extract_offset, %ebx
+	movl    BP_init_size(%esi), %eax
+	subl    $_end, %eax
+	andl    $(~(4096-1)), %eax
+	addl    %eax, %ebx
 
 	/* Set up the stack */
 	leal	boot_stack_end(%ebx), %esp
@@ -210,8 +213,14 @@ relocated:
 				/* push arguments for decompress_kernel: */
 	pushl	$z_run_size	/* size of kernel with .bss and .brk */
 	pushl	$z_output_len	/* decompressed length, end of relocs */
-	leal	z_extract_offset_negative(%ebx), %ebp
+
+	movl    BP_init_size(%esi), %eax
+	subl    $_end, %eax
+	andl    $(~(4096-1)), %eax
+	movl	%ebx, %ebp
+	subl	%eax, %ebp
 	pushl	%ebp		/* output address */
+
 	pushl	$z_input_len	/* input_len */
 	leal	input_data(%ebx), %eax
 	pushl	%eax		/* input_data */
Index: linux-2.6/arch/x86/boot/compressed/mkpiggy.c
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/mkpiggy.c
+++ linux-2.6/arch/x86/boot/compressed/mkpiggy.c
@@ -85,9 +85,6 @@ int main(int argc, char *argv[])
 	printf("z_output_len = %lu\n", (unsigned long)olen);
 	printf(".globl z_extract_offset\n");
 	printf("z_extract_offset = 0x%lx\n", offs);
-	/* z_extract_offset_negative allows simplification of head_32.S */
-	printf(".globl z_extract_offset_negative\n");
-	printf("z_extract_offset_negative = -0x%lx\n", offs);
 	printf(".globl z_run_size\n");
 	printf("z_run_size = %lu\n", run_size);
 
Index: linux-2.6/arch/x86/boot/compressed/misc.c
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/misc.c
+++ linux-2.6/arch/x86/boot/compressed/misc.c
@@ -366,6 +366,8 @@ static void parse_elf(void *output)
 	free(phdrs);
 }
 
+extern char _rodata[], _end[];
+
 asmlinkage __visible void *decompress_kernel(void *rmode, memptr heap,
 				  unsigned char *input_data,
 				  unsigned long input_len,
@@ -392,6 +394,8 @@ asmlinkage __visible void *decompress_ke
 
 	console_init();
 	debug_putstr("early console in decompress_kernel\n");
+	if (_end - _rodata > BOOT_ADDON_ZO_SIZE)
+		debug_putstr("BOOT_ADDON_ZO_SIZE too small\n");
 
 	free_mem_ptr     = heap;	/* Heap */
 	free_mem_end_ptr = heap + BOOT_HEAP_SIZE;
Index: linux-2.6/arch/x86/include/asm/boot.h
===================================================================
--- linux-2.6.orig/arch/x86/include/asm/boot.h
+++ linux-2.6/arch/x86/include/asm/boot.h
@@ -37,4 +37,6 @@
 #define BOOT_STACK_SIZE	0x1000
 #endif
 
+#define BOOT_ADDON_ZO_SIZE 0x40000
+
 #endif /* _ASM_X86_BOOT_H */

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

[-- Attachment #1: Type: text/plain, Size: 314 bytes --]

On Mon, Mar 2, 2015 at 12:56 AM, Borislav Petkov <bp-l3A5Bk7waGM@public.gmane.org> wrote:
> On Sun, Mar 01, 2015 at 12:41:10PM -0800, Yinghai Lu wrote:
>> Does not look safe yet...
>
> Why?

We can not assume that range is safe to use.

Please check attach one that should fix the problem really.

Thanks

Yinghai

[-- Attachment #2: fix_decompress_offset.patch --]
[-- Type: text/x-patch, Size: 5440 bytes --]

Subject: [PATCH] x86, boot: keep data from boot stage to kernel stage.

Need to
1. move compressed kernel close the end of buffer instead of middle of it.
2. use BRK to extend init_size so no one from kernel bss and brk will
   touch the data region from boot/compressed/misc.c

Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 arch/x86/boot/compressed/head_32.S |   13 +++++++++++--
 arch/x86/boot/compressed/head_64.S |   10 ++++++++--
 arch/x86/boot/compressed/misc.c    |    4 ++++
 arch/x86/boot/compressed/mkpiggy.c |    3 ---
 arch/x86/include/asm/boot.h        |    2 ++
 arch/x86/kernel/asm-offsets.c      |    1 +
 arch/x86/kernel/setup.c            |    3 +++
 7 files changed, 29 insertions(+), 7 deletions(-)

Index: linux-2.6/arch/x86/boot/compressed/head_64.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_64.S
+++ linux-2.6/arch/x86/boot/compressed/head_64.S
@@ -102,7 +102,10 @@ ENTRY(startup_32)
 1:
 
 	/* Target address to relocate to for decompression */
-	addl	$z_extract_offset, %ebx
+	movl	BP_init_size(%esi), %eax
+	subl	$_end, %eax
+	andl	$(~(4096-1)), %eax
+	addl	%eax, %ebx
 
 /*
  * Prepare for entering 64 bit mode
@@ -330,7 +333,10 @@ preferred_addr:
 1:
 
 	/* Target address to relocate to for decompression */
-	leaq	z_extract_offset(%rbp), %rbx
+	movl	BP_init_size(%rsi), %ebx
+	subl	$_end, %ebx
+	andl	$(~(4096-1)), %ebx
+	addq	%rbp, %rbx
 
 	/* Set up the stack */
 	leaq	boot_stack_end(%rbx), %rsp
Index: linux-2.6/arch/x86/kernel/asm-offsets.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/asm-offsets.c
+++ linux-2.6/arch/x86/kernel/asm-offsets.c
@@ -66,6 +66,7 @@ void common(void) {
 	OFFSET(BP_hardware_subarch, boot_params, hdr.hardware_subarch);
 	OFFSET(BP_version, boot_params, hdr.version);
 	OFFSET(BP_kernel_alignment, boot_params, hdr.kernel_alignment);
+	OFFSET(BP_init_size, boot_params, hdr.init_size);
 	OFFSET(BP_pref_address, boot_params, hdr.pref_address);
 	OFFSET(BP_code32_start, boot_params, hdr.code32_start);
 
Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ linux-2.6/arch/x86/kernel/setup.c
@@ -111,6 +111,7 @@
 #include <asm/mce.h>
 #include <asm/alternative.h>
 #include <asm/prom.h>
+#include <asm/boot.h>
 
 /*
  * max_low_pfn_mapped: highest direct mapped pfn under 4GB
@@ -144,6 +145,8 @@ int default_check_phys_apicid_present(in
 }
 #endif
 
+RESERVE_BRK(addon_zo, BOOT_ADDON_ZO_SIZE);
+
 struct boot_params boot_params;
 
 /*
Index: linux-2.6/arch/x86/boot/compressed/head_32.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_32.S
+++ linux-2.6/arch/x86/boot/compressed/head_32.S
@@ -148,7 +148,10 @@ preferred_addr:
 1:
 
 	/* Target address to relocate to for decompression */
-	addl	$z_extract_offset, %ebx
+	movl    BP_init_size(%esi), %eax
+	subl    $_end, %eax
+	andl    $(~(4096-1)), %eax
+	addl    %eax, %ebx
 
 	/* Set up the stack */
 	leal	boot_stack_end(%ebx), %esp
@@ -210,8 +213,14 @@ relocated:
 				/* push arguments for decompress_kernel: */
 	pushl	$z_run_size	/* size of kernel with .bss and .brk */
 	pushl	$z_output_len	/* decompressed length, end of relocs */
-	leal	z_extract_offset_negative(%ebx), %ebp
+
+	movl    BP_init_size(%esi), %eax
+	subl    $_end, %eax
+	andl    $(~(4096-1)), %eax
+	movl	%ebx, %ebp
+	subl	%eax, %ebp
 	pushl	%ebp		/* output address */
+
 	pushl	$z_input_len	/* input_len */
 	leal	input_data(%ebx), %eax
 	pushl	%eax		/* input_data */
Index: linux-2.6/arch/x86/boot/compressed/mkpiggy.c
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/mkpiggy.c
+++ linux-2.6/arch/x86/boot/compressed/mkpiggy.c
@@ -85,9 +85,6 @@ int main(int argc, char *argv[])
 	printf("z_output_len = %lu\n", (unsigned long)olen);
 	printf(".globl z_extract_offset\n");
 	printf("z_extract_offset = 0x%lx\n", offs);
-	/* z_extract_offset_negative allows simplification of head_32.S */
-	printf(".globl z_extract_offset_negative\n");
-	printf("z_extract_offset_negative = -0x%lx\n", offs);
 	printf(".globl z_run_size\n");
 	printf("z_run_size = %lu\n", run_size);
 
Index: linux-2.6/arch/x86/boot/compressed/misc.c
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/misc.c
+++ linux-2.6/arch/x86/boot/compressed/misc.c
@@ -366,6 +366,8 @@ static void parse_elf(void *output)
 	free(phdrs);
 }
 
+extern char _rodata[], _end[];
+
 asmlinkage __visible void *decompress_kernel(void *rmode, memptr heap,
 				  unsigned char *input_data,
 				  unsigned long input_len,
@@ -392,6 +394,8 @@ asmlinkage __visible void *decompress_ke
 
 	console_init();
 	debug_putstr("early console in decompress_kernel\n");
+	if (_end - _rodata > BOOT_ADDON_ZO_SIZE)
+		debug_putstr("BOOT_ADDON_ZO_SIZE too small\n");
 
 	free_mem_ptr     = heap;	/* Heap */
 	free_mem_end_ptr = heap + BOOT_HEAP_SIZE;
Index: linux-2.6/arch/x86/include/asm/boot.h
===================================================================
--- linux-2.6.orig/arch/x86/include/asm/boot.h
+++ linux-2.6/arch/x86/include/asm/boot.h
@@ -37,4 +37,6 @@
 #define BOOT_STACK_SIZE	0x1000
 #endif
 
+#define BOOT_ADDON_ZO_SIZE 0x40000
+
 #endif /* _ASM_X86_BOOT_H */

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Borislav Petkov]

On Mon, Mar 02, 2015 at 03:04:30AM -0800, Yinghai Lu wrote:
> We can not assume that range is safe to use.
> 
> Please check attach one that should fix the problem really.

Well, it seems to work here but it still doesn't look reliable enough to
me. And this addon_zo thing of arbitrary 256K is strange.

What we should do instead and IMHO is have a special setup_data section
located right above .text where struct setup_data things are put in so
that they never get overwritten. This should be the cleanest IMHO.

It is hpa's call in the end of the day anyway.

Thanks.

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

On Mon, Mar 2, 2015 at 6:53 AM, Borislav Petkov <bp@suse.de> wrote:
> Well, it seems to work here but it still doesn't look reliable enough to
> me. And this addon_zo thing of arbitrary 256K is strange.

Thanks for check that out.

That is not arbitrary number. Need to make that bigger than _end - _rodata

> objdump -t arch/x86/boot/compressed/vmlinux | grep \ _end
0000000000996000 g       .pgtable    0000000000000000 _end
> objdump -t arch/x86/boot/compressed/vmlinux | grep \ _rodata
0000000000981780 g       .rodata    0000000000000000 _rodata

We only get that size after arch/x86/boot/compressed/vmlinux

but need number during building kernel vmlinux.

other way would be just increase init_size in arch/x86/boot/header.S
instead of put it BRK area.

Thanks

Yinghai

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Borislav Petkov]

On Mon, Mar 02, 2015 at 10:58:23AM -0800, Yinghai Lu wrote:
> On Mon, Mar 2, 2015 at 6:53 AM, Borislav Petkov <bp@suse.de> wrote:
> > Well, it seems to work here but it still doesn't look reliable enough to
> > me. And this addon_zo thing of arbitrary 256K is strange.
> 
> Thanks for check that out.
> 
> That is not arbitrary number. Need to make that bigger than _end - _rodata
> 
> > objdump -t arch/x86/boot/compressed/vmlinux | grep \ _end
> 0000000000996000 g       .pgtable    0000000000000000 _end
> > objdump -t arch/x86/boot/compressed/vmlinux | grep \ _rodata
> 0000000000981780 g       .rodata    0000000000000000 _rodata
> 
> We only get that size after arch/x86/boot/compressed/vmlinux
> 
> but need number during building kernel vmlinux.
> 
> other way would be just increase init_size in arch/x86/boot/header.S
> instead of put it BRK area.

Ok, I'll take a look at this more tomorrow, on a clear head, but why
can't we make a special, explicit section which is not touched by
the early decompression and relocation code and which we can use for
setup_data only?

IOW, something like that although that doesn't work yet:

---
diff --git a/arch/x86/boot/compressed/aslr.c b/arch/x86/boot/compressed/aslr.c
index 7083c16cccba..01d5ddf1d22f 100644
--- a/arch/x86/boot/compressed/aslr.c
+++ b/arch/x86/boot/compressed/aslr.c
@@ -14,12 +14,12 @@
 static const char build_str[] = UTS_RELEASE " (" LINUX_COMPILE_BY "@"
 		LINUX_COMPILE_HOST ") (" LINUX_COMPILER ") " UTS_VERSION;
 
-struct kaslr_setup_data {
-	__u64 next;
-	__u32 type;
-	__u32 len;
-	__u8 data[1];
-} kaslr_setup_data;
+extern char _setup_data[];
+struct setup_data __attribute__((section (".setup_data"))) ksd = {
+	.type	= SETUP_KASLR,
+	.len	= 1,
+	.next	= 0,
+};
 
 #define I8254_PORT_CONTROL	0x43
 #define I8254_PORT_COUNTER0	0x40
@@ -306,10 +306,7 @@ static void add_kaslr_setup_data(struct boot_params *params, __u8 enabled)
 {
 	struct setup_data *data;
 
-	kaslr_setup_data.type = SETUP_KASLR;
-	kaslr_setup_data.len = 1;
-	kaslr_setup_data.next = 0;
-	kaslr_setup_data.data[0] = enabled;
+	ksd.data[0] = enabled;
 
 	data = (struct setup_data *)(unsigned long)params->hdr.setup_data;
 
@@ -317,10 +314,9 @@ static void add_kaslr_setup_data(struct boot_params *params, __u8 enabled)
 		data = (struct setup_data *)(unsigned long)data->next;
 
 	if (data)
-		data->next = (unsigned long)&kaslr_setup_data;
+		data->next = (unsigned long)&ksd;
 	else
-		params->hdr.setup_data = (unsigned long)&kaslr_setup_data;
-
+		params->hdr.setup_data = (unsigned long)&ksd;
 }
 
 unsigned char *choose_kernel_location(struct boot_params *params,
diff --git a/arch/x86/boot/compressed/vmlinux.lds.S b/arch/x86/boot/compressed/vmlinux.lds.S
index 34d047c98284..26d62f4b27b9 100644
--- a/arch/x86/boot/compressed/vmlinux.lds.S
+++ b/arch/x86/boot/compressed/vmlinux.lds.S
@@ -29,6 +29,10 @@ SECTIONS
 	.rodata..compressed : {
 		*(.rodata..compressed)
 	}
+	.setup_data : {
+		_setup_data = . ;
+		*(.setup_data)
+	}
 	.text :	{
 		_text = .; 	/* Text */
 		*(.text)
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 98dc9317286e..0978c61f84bd 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -429,7 +429,11 @@ static void __init reserve_initrd(void)
 
 static void __init parse_kaslr_setup(u64 pa_data, u32 data_len)
 {
-	kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));
+	struct setup_data *kdata;
+
+	kdata = early_memremap(pa_data, data_len);
+	kaslr_enabled = kdata->data[0];
+	early_iounmap(kdata, data_len);
 }
 
 static void __init parse_setup_data(void)

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

[-- Attachment #1: Type: text/plain, Size: 1230 bytes --]

On Mon, Mar 2, 2015 at 12:25 PM, Borislav Petkov <bp@suse.de> wrote:

>  unsigned char *choose_kernel_location(struct boot_params *params,
> diff --git a/arch/x86/boot/compressed/vmlinux.lds.S b/arch/x86/boot/compressed/vmlinux.lds.S
> index 34d047c98284..26d62f4b27b9 100644
> --- a/arch/x86/boot/compressed/vmlinux.lds.S
> +++ b/arch/x86/boot/compressed/vmlinux.lds.S
> @@ -29,6 +29,10 @@ SECTIONS
>         .rodata..compressed : {
>                 *(.rodata..compressed)
>         }
> +       .setup_data : {
> +               _setup_data = . ;
> +               *(.setup_data)
> +       }
>         .text : {
>                 _text = .;      /* Text */
>                 *(.text)

That does not help, we will still have overlap between copied Zo
vmlinux and decompressed Vo vmlinux

because the Zo vmlinux is copied to middle of buffer instead of end of
the buffer.

definition:
Zo vmliunx: arch/x86/boot/compressed/vmlinux, from
arch/x86/boot/compressed/vmlinux.lds.S
Vo vmlinux: vmlinux from arch/x86/kernel/vmlinux.lds.S

BTW, found more problem about run_size for kaslr, we should use
init_size instead.

Here the three patches that should go into v4.0. Help Ingo will be
happy with updated change log.

Thanks

Yinghai

[-- Attachment #2: use_init_size_instead_of_run_size.patch --]
[-- Type: text/x-patch, Size: 7639 bytes --]

Subject: [PATCH] x86, boot, aslr: Use init_size instead of run_size

commit e6023367d779 ("x86, kaslr: Prevent .bss from overlaping initrd")

introduced one run_size for kaslr.

We do not need to have home grown run_size.

We should use real runtime size (include copy/decompress) aka init_size

Please check arch/x86/boot/header.S about init_size for detail.

Fixes: e6023367d779 ("x86, kaslr: Prevent .bss from overlaping initrd")
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Matt Fleming <matt.fleming@intel.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Junjie Mao <eternal.n08@gmail.com>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 arch/x86/boot/compressed/Makefile  |    4 ---
 arch/x86/boot/compressed/head_32.S |    5 +---
 arch/x86/boot/compressed/head_64.S |    5 ----
 arch/x86/boot/compressed/misc.c    |   15 ++++++-------
 arch/x86/boot/compressed/mkpiggy.c |    9 +------
 arch/x86/tools/calc_run_size.sh    |   42 -------------------------------------
 6 files changed, 13 insertions(+), 67 deletions(-)

Index: linux-2.6/arch/x86/boot/compressed/Makefile
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/Makefile
+++ linux-2.6/arch/x86/boot/compressed/Makefile
@@ -92,10 +92,8 @@ suffix-$(CONFIG_KERNEL_XZ)	:= xz
 suffix-$(CONFIG_KERNEL_LZO) 	:= lzo
 suffix-$(CONFIG_KERNEL_LZ4) 	:= lz4
 
-RUN_SIZE = $(shell $(OBJDUMP) -h vmlinux | \
-	     $(CONFIG_SHELL) $(srctree)/arch/x86/tools/calc_run_size.sh)
 quiet_cmd_mkpiggy = MKPIGGY $@
-      cmd_mkpiggy = $(obj)/mkpiggy $< $(RUN_SIZE) > $@ || ( rm -f $@ ; false )
+      cmd_mkpiggy = $(obj)/mkpiggy $< > $@ || ( rm -f $@ ; false )
 
 targets += piggy.S
 $(obj)/piggy.S: $(obj)/vmlinux.bin.$(suffix-y) $(obj)/mkpiggy FORCE
Index: linux-2.6/arch/x86/boot/compressed/head_32.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_32.S
+++ linux-2.6/arch/x86/boot/compressed/head_32.S
@@ -208,8 +208,7 @@ relocated:
  * Do the decompression, and jump to the new kernel..
  */
 				/* push arguments for decompress_kernel: */
-	pushl	$z_run_size	/* size of kernel with .bss and .brk */
-	pushl	$z_output_len	/* decompressed length, end of relocs */
+	pushl	$z_output_len	/* decompressed length */
 	leal	z_extract_offset_negative(%ebx), %ebp
 	pushl	%ebp		/* output address */
 	pushl	$z_input_len	/* input_len */
@@ -219,7 +218,7 @@ relocated:
 	pushl	%eax		/* heap area */
 	pushl	%esi		/* real mode pointer */
 	call	decompress_kernel /* returns kernel location in %eax */
-	addl	$28, %esp
+	addl	$24, %esp
 
 /*
  * Jump to the decompressed kernel.
Index: linux-2.6/arch/x86/boot/compressed/head_64.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_64.S
+++ linux-2.6/arch/x86/boot/compressed/head_64.S
@@ -403,16 +403,13 @@ relocated:
  * Do the decompression, and jump to the new kernel..
  */
 	pushq	%rsi			/* Save the real mode argument */
-	movq	$z_run_size, %r9	/* size of kernel with .bss and .brk */
-	pushq	%r9
 	movq	%rsi, %rdi		/* real mode address */
 	leaq	boot_heap(%rip), %rsi	/* malloc area for uncompression */
 	leaq	input_data(%rip), %rdx  /* input_data */
 	movl	$z_input_len, %ecx	/* input_len */
 	movq	%rbp, %r8		/* output target address */
-	movq	$z_output_len, %r9	/* decompressed length, end of relocs */
+	movq	$z_output_len, %r9	/* decompressed length */
 	call	decompress_kernel	/* returns kernel location in %rax */
-	popq	%r9
 	popq	%rsi
 
 /*
Index: linux-2.6/arch/x86/boot/compressed/misc.c
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/misc.c
+++ linux-2.6/arch/x86/boot/compressed/misc.c
@@ -370,10 +370,10 @@ asmlinkage __visible void *decompress_ke
 				  unsigned char *input_data,
 				  unsigned long input_len,
 				  unsigned char *output,
-				  unsigned long output_len,
-				  unsigned long run_size)
+				  unsigned long output_len)
 {
 	unsigned char *output_orig = output;
+	unsigned long init_size;
 
 	real_mode = rmode;
 
@@ -396,15 +396,14 @@ asmlinkage __visible void *decompress_ke
 	free_mem_ptr     = heap;	/* Heap */
 	free_mem_end_ptr = heap + BOOT_HEAP_SIZE;
 
+	init_size = real_mode->hdr.init_size;
+
 	/*
-	 * The memory hole needed for the kernel is the larger of either
-	 * the entire decompressed kernel plus relocation table, or the
-	 * entire decompressed kernel plus .bss and .brk sections.
+	 * The memory hole needed for the kernel is init_size for running
+	 * and init_size is bigger than output_len always.
 	 */
 	output = choose_kernel_location(real_mode, input_data, input_len,
-					output,
-					output_len > run_size ? output_len
-							      : run_size);
+					output, init_size);
 
 	/* Validate memory location choices. */
 	if ((unsigned long)output & (MIN_KERNEL_ALIGN - 1))
Index: linux-2.6/arch/x86/boot/compressed/mkpiggy.c
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/mkpiggy.c
+++ linux-2.6/arch/x86/boot/compressed/mkpiggy.c
@@ -36,13 +36,11 @@ int main(int argc, char *argv[])
 	uint32_t olen;
 	long ilen;
 	unsigned long offs;
-	unsigned long run_size;
 	FILE *f = NULL;
 	int retval = 1;
 
-	if (argc < 3) {
-		fprintf(stderr, "Usage: %s compressed_file run_size\n",
-				argv[0]);
+	if (argc < 2) {
+		fprintf(stderr, "Usage: %s compressed_file\n", argv[0]);
 		goto bail;
 	}
 
@@ -76,7 +74,6 @@ int main(int argc, char *argv[])
 	offs += olen >> 12;	/* Add 8 bytes for each 32K block */
 	offs += 64*1024 + 128;	/* Add 64K + 128 bytes slack */
 	offs = (offs+4095) & ~4095; /* Round to a 4K boundary */
-	run_size = atoi(argv[2]);
 
 	printf(".section \".rodata..compressed\",\"a\",@progbits\n");
 	printf(".globl z_input_len\n");
@@ -88,8 +85,6 @@ int main(int argc, char *argv[])
 	/* z_extract_offset_negative allows simplification of head_32.S */
 	printf(".globl z_extract_offset_negative\n");
 	printf("z_extract_offset_negative = -0x%lx\n", offs);
-	printf(".globl z_run_size\n");
-	printf("z_run_size = %lu\n", run_size);
 
 	printf(".globl input_data, input_data_end\n");
 	printf("input_data:\n");
Index: linux-2.6/arch/x86/tools/calc_run_size.sh
===================================================================
--- linux-2.6.orig/arch/x86/tools/calc_run_size.sh
+++ /dev/null
@@ -1,42 +0,0 @@
-#!/bin/sh
-#
-# Calculate the amount of space needed to run the kernel, including room for
-# the .bss and .brk sections.
-#
-# Usage:
-# objdump -h a.out | sh calc_run_size.sh
-
-NUM='\([0-9a-fA-F]*[ \t]*\)'
-OUT=$(sed -n 's/^[ \t0-9]*.b[sr][sk][ \t]*'"$NUM$NUM$NUM$NUM"'.*/\1\4/p')
-if [ -z "$OUT" ] ; then
-	echo "Never found .bss or .brk file offset" >&2
-	exit 1
-fi
-
-OUT=$(echo ${OUT# })
-sizeA=$(printf "%d" 0x${OUT%% *})
-OUT=${OUT#* }
-offsetA=$(printf "%d" 0x${OUT%% *})
-OUT=${OUT#* }
-sizeB=$(printf "%d" 0x${OUT%% *})
-OUT=${OUT#* }
-offsetB=$(printf "%d" 0x${OUT%% *})
-
-run_size=$(( $offsetA + $sizeA + $sizeB ))
-
-# BFD linker shows the same file offset in ELF.
-if [ "$offsetA" -ne "$offsetB" ] ; then
-	# Gold linker shows them as consecutive.
-	endB=$(( $offsetB + $sizeB ))
-	if [ "$endB" != "$run_size" ] ; then
-		printf "sizeA: 0x%x\n" $sizeA >&2
-		printf "offsetA: 0x%x\n" $offsetA >&2
-		printf "sizeB: 0x%x\n" $sizeB >&2
-		printf "offsetB: 0x%x\n" $offsetB >&2
-		echo ".bss and .brk are non-contiguous" >&2
-		exit 1
-	fi
-fi
-
-printf "%d\n" $run_size
-exit 0

[-- Attachment #3: fix_decompress_offset.patch --]
[-- Type: text/x-patch, Size: 6418 bytes --]

Subject: [PATCH v2] x86, boot: keep data from boot stage to kernel stage.

bp found data from boot stage can not be used kernel stage.

Acctually those data area is overlapped with kernel bss stage, and clear_bss()
clear them before code in arch/x86/kernel/setup.c access them.

To make the data survive that later, we should avoid the overlapping.

Need to:
1. move compressed kernel close the end of buffer instead of middle of it.
2. extend init_size so no one from kernel bss and brk will touch the data
   region from boot/compressed/misc.c

-v2: add init_size in arch/x86/boot/header.S instead of BRK.

Fixes: f47233c2d34f ("x86/mm/ASLR: Propagate base load address calculation")
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Matt Fleming <matt.fleming@intel.com> 
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 arch/x86/boot/Makefile                 |    2 +-
 arch/x86/boot/compressed/head_32.S     |   11 +++++++++--
 arch/x86/boot/compressed/head_64.S     |    8 ++++++--
 arch/x86/boot/compressed/mkpiggy.c     |    3 ---
 arch/x86/boot/compressed/vmlinux.lds.S |    2 ++
 arch/x86/boot/header.S                 |    7 +++++--
 arch/x86/kernel/asm-offsets.c          |    1 +
 7 files changed, 24 insertions(+), 10 deletions(-)

Index: linux-2.6/arch/x86/boot/compressed/head_64.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_64.S
+++ linux-2.6/arch/x86/boot/compressed/head_64.S
@@ -102,7 +102,9 @@ ENTRY(startup_32)
 1:
 
 	/* Target address to relocate to for decompression */
-	addl	$z_extract_offset, %ebx
+	movl	BP_init_size(%esi), %eax
+	subl	$_end, %eax
+	addl	%eax, %ebx
 
 /*
  * Prepare for entering 64 bit mode
@@ -330,7 +332,9 @@ preferred_addr:
 1:
 
 	/* Target address to relocate to for decompression */
-	leaq	z_extract_offset(%rbp), %rbx
+	movl	BP_init_size(%rsi), %ebx
+	subl	$_end, %ebx
+	addq	%rbp, %rbx
 
 	/* Set up the stack */
 	leaq	boot_stack_end(%rbx), %rsp
Index: linux-2.6/arch/x86/kernel/asm-offsets.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/asm-offsets.c
+++ linux-2.6/arch/x86/kernel/asm-offsets.c
@@ -66,6 +66,7 @@ void common(void) {
 	OFFSET(BP_hardware_subarch, boot_params, hdr.hardware_subarch);
 	OFFSET(BP_version, boot_params, hdr.version);
 	OFFSET(BP_kernel_alignment, boot_params, hdr.kernel_alignment);
+	OFFSET(BP_init_size, boot_params, hdr.init_size);
 	OFFSET(BP_pref_address, boot_params, hdr.pref_address);
 	OFFSET(BP_code32_start, boot_params, hdr.code32_start);
 
Index: linux-2.6/arch/x86/boot/compressed/head_32.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_32.S
+++ linux-2.6/arch/x86/boot/compressed/head_32.S
@@ -148,7 +148,9 @@ preferred_addr:
 1:
 
 	/* Target address to relocate to for decompression */
-	addl	$z_extract_offset, %ebx
+	movl    BP_init_size(%esi), %eax
+	subl    $_end, %eax
+	addl    %eax, %ebx
 
 	/* Set up the stack */
 	leal	boot_stack_end(%ebx), %esp
@@ -209,8 +211,13 @@ relocated:
  */
 				/* push arguments for decompress_kernel: */
 	pushl	$z_output_len	/* decompressed length */
-	leal	z_extract_offset_negative(%ebx), %ebp
+
+	movl    BP_init_size(%esi), %eax
+	subl    $_end, %eax
+	movl    %ebx, %ebp
+	subl    %eax, %ebp
 	pushl	%ebp		/* output address */
+
 	pushl	$z_input_len	/* input_len */
 	leal	input_data(%ebx), %eax
 	pushl	%eax		/* input_data */
Index: linux-2.6/arch/x86/boot/compressed/mkpiggy.c
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/mkpiggy.c
+++ linux-2.6/arch/x86/boot/compressed/mkpiggy.c
@@ -82,9 +82,6 @@ int main(int argc, char *argv[])
 	printf("z_output_len = %lu\n", (unsigned long)olen);
 	printf(".globl z_extract_offset\n");
 	printf("z_extract_offset = 0x%lx\n", offs);
-	/* z_extract_offset_negative allows simplification of head_32.S */
-	printf(".globl z_extract_offset_negative\n");
-	printf("z_extract_offset_negative = -0x%lx\n", offs);
 
 	printf(".globl input_data, input_data_end\n");
 	printf("input_data:\n");
Index: linux-2.6/arch/x86/boot/Makefile
===================================================================
--- linux-2.6.orig/arch/x86/boot/Makefile
+++ linux-2.6/arch/x86/boot/Makefile
@@ -86,7 +86,7 @@ targets += voffset.h
 $(obj)/voffset.h: vmlinux FORCE
 	$(call if_changed,voffset)
 
-sed-zoffset := -e 's/^\([0-9a-fA-F]*\) [ABCDGRSTVW] \(startup_32\|startup_64\|efi32_stub_entry\|efi64_stub_entry\|efi_pe_entry\|input_data\|_end\|z_.*\)$$/\#define ZO_\2 0x\1/p'
+sed-zoffset := -e 's/^\([0-9a-fA-F]*\) [ABCDGRSTVW] \(startup_32\|startup_64\|efi32_stub_entry\|efi64_stub_entry\|efi_pe_entry\|input_data\|_end\|_rodata\|z_.*\)$$/\#define ZO_\2 0x\1/p'
 
 quiet_cmd_zoffset = ZOFFSET $@
       cmd_zoffset = $(NM) $< | sed -n $(sed-zoffset) > $@
Index: linux-2.6/arch/x86/boot/compressed/vmlinux.lds.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/vmlinux.lds.S
+++ linux-2.6/arch/x86/boot/compressed/vmlinux.lds.S
@@ -35,6 +35,7 @@ SECTIONS
 		*(.text.*)
 		_etext = . ;
 	}
+        . = ALIGN(PAGE_SIZE); /* keep ADDON_ZO_SIZE page aligned */
 	.rodata : {
 		_rodata = . ;
 		*(.rodata)	 /* read-only data */
@@ -70,5 +71,6 @@ SECTIONS
 		_epgtable = . ;
 	}
 #endif
+        . = ALIGN(PAGE_SIZE);  /* keep ADDON_ZO_SIZE page aligned */
 	_end = .;
 }
Index: linux-2.6/arch/x86/boot/header.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/header.S
+++ linux-2.6/arch/x86/boot/header.S
@@ -440,12 +440,15 @@ setup_data:		.quad 0			# 64-bit physical
 
 pref_address:		.quad LOAD_PHYSICAL_ADDR	# preferred load addr
 
+# don't overlap data area of ZO with VO
+#define ADDON_ZO_SIZE (ZO__end - ZO__rodata)
+
 #define ZO_INIT_SIZE	(ZO__end - ZO_startup_32 + ZO_z_extract_offset)
 #define VO_INIT_SIZE	(VO__end - VO__text)
 #if ZO_INIT_SIZE > VO_INIT_SIZE
-#define INIT_SIZE ZO_INIT_SIZE
+#define INIT_SIZE (ZO_INIT_SIZE + ADDON_ZO_SIZE)
 #else
-#define INIT_SIZE VO_INIT_SIZE
+#define INIT_SIZE (VO_INIT_SIZE + ADDON_ZO_SIZE)
 #endif
 init_size:		.long INIT_SIZE		# kernel initialization size
 handover_offset:	.long 0			# Filled in by build.c

[-- Attachment #4: fix_kaslr.patch --]
[-- Type: text/x-patch, Size: 1405 bytes --]

Subject: [PATCH v2] x86, kaslr: get kaslr_enabled back correctly

commit f47233c2d34f ("x86/mm/ASLR: Propagate base load address calculation")
is using address as vaule for kaslr_enabled.

That will random kaslr_enabled get that set or cleared.
Will have problem for system really have kaslr enabled.

-v2: update changelog.

Fixes: f47233c2d34f ("x86/mm/ASLR: Propagate base load address calculation")
Cc: Matt Fleming <matt.fleming@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Kees Cook <keescook@chromium.org>
Cc: Jiri Kosina <jkosina@suse.cz>
Acked-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 arch/x86/kernel/setup.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ linux-2.6/arch/x86/kernel/setup.c
@@ -429,7 +429,13 @@ static void __init reserve_initrd(void)
 
 static void __init parse_kaslr_setup(u64 pa_data, u32 data_len)
 {
-	kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));
+	/* kaslr_setup_data is defined in aslr.c */
+	unsigned char *data;
+	unsigned long offset = sizeof(struct setup_data);
+
+	data = early_memremap(pa_data, offset + 1);
+	kaslr_enabled = *(data + offset);
+	early_memunmap(data, offset + 1);
 }
 
 static void __init parse_setup_data(void)

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Yinghai Lu]

[-- Attachment #1: Type: text/plain, Size: 1251 bytes --]

On Mon, Mar 2, 2015 at 12:25 PM, Borislav Petkov <bp-l3A5Bk7waGM@public.gmane.org> wrote:

>  unsigned char *choose_kernel_location(struct boot_params *params,
> diff --git a/arch/x86/boot/compressed/vmlinux.lds.S b/arch/x86/boot/compressed/vmlinux.lds.S
> index 34d047c98284..26d62f4b27b9 100644
> --- a/arch/x86/boot/compressed/vmlinux.lds.S
> +++ b/arch/x86/boot/compressed/vmlinux.lds.S
> @@ -29,6 +29,10 @@ SECTIONS
>         .rodata..compressed : {
>                 *(.rodata..compressed)
>         }
> +       .setup_data : {
> +               _setup_data = . ;
> +               *(.setup_data)
> +       }
>         .text : {
>                 _text = .;      /* Text */
>                 *(.text)

That does not help, we will still have overlap between copied Zo
vmlinux and decompressed Vo vmlinux

because the Zo vmlinux is copied to middle of buffer instead of end of
the buffer.

definition:
Zo vmliunx: arch/x86/boot/compressed/vmlinux, from
arch/x86/boot/compressed/vmlinux.lds.S
Vo vmlinux: vmlinux from arch/x86/kernel/vmlinux.lds.S

BTW, found more problem about run_size for kaslr, we should use
init_size instead.

Here the three patches that should go into v4.0. Help Ingo will be
happy with updated change log.

Thanks

Yinghai

[-- Attachment #2: use_init_size_instead_of_run_size.patch --]
[-- Type: text/x-patch, Size: 7639 bytes --]

Subject: [PATCH] x86, boot, aslr: Use init_size instead of run_size

commit e6023367d779 ("x86, kaslr: Prevent .bss from overlaping initrd")

introduced one run_size for kaslr.

We do not need to have home grown run_size.

We should use real runtime size (include copy/decompress) aka init_size

Please check arch/x86/boot/header.S about init_size for detail.

Fixes: e6023367d779 ("x86, kaslr: Prevent .bss from overlaping initrd")
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Matt Fleming <matt.fleming@intel.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Junjie Mao <eternal.n08@gmail.com>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 arch/x86/boot/compressed/Makefile  |    4 ---
 arch/x86/boot/compressed/head_32.S |    5 +---
 arch/x86/boot/compressed/head_64.S |    5 ----
 arch/x86/boot/compressed/misc.c    |   15 ++++++-------
 arch/x86/boot/compressed/mkpiggy.c |    9 +------
 arch/x86/tools/calc_run_size.sh    |   42 -------------------------------------
 6 files changed, 13 insertions(+), 67 deletions(-)

Index: linux-2.6/arch/x86/boot/compressed/Makefile
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/Makefile
+++ linux-2.6/arch/x86/boot/compressed/Makefile
@@ -92,10 +92,8 @@ suffix-$(CONFIG_KERNEL_XZ)	:= xz
 suffix-$(CONFIG_KERNEL_LZO) 	:= lzo
 suffix-$(CONFIG_KERNEL_LZ4) 	:= lz4
 
-RUN_SIZE = $(shell $(OBJDUMP) -h vmlinux | \
-	     $(CONFIG_SHELL) $(srctree)/arch/x86/tools/calc_run_size.sh)
 quiet_cmd_mkpiggy = MKPIGGY $@
-      cmd_mkpiggy = $(obj)/mkpiggy $< $(RUN_SIZE) > $@ || ( rm -f $@ ; false )
+      cmd_mkpiggy = $(obj)/mkpiggy $< > $@ || ( rm -f $@ ; false )
 
 targets += piggy.S
 $(obj)/piggy.S: $(obj)/vmlinux.bin.$(suffix-y) $(obj)/mkpiggy FORCE
Index: linux-2.6/arch/x86/boot/compressed/head_32.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_32.S
+++ linux-2.6/arch/x86/boot/compressed/head_32.S
@@ -208,8 +208,7 @@ relocated:
  * Do the decompression, and jump to the new kernel..
  */
 				/* push arguments for decompress_kernel: */
-	pushl	$z_run_size	/* size of kernel with .bss and .brk */
-	pushl	$z_output_len	/* decompressed length, end of relocs */
+	pushl	$z_output_len	/* decompressed length */
 	leal	z_extract_offset_negative(%ebx), %ebp
 	pushl	%ebp		/* output address */
 	pushl	$z_input_len	/* input_len */
@@ -219,7 +218,7 @@ relocated:
 	pushl	%eax		/* heap area */
 	pushl	%esi		/* real mode pointer */
 	call	decompress_kernel /* returns kernel location in %eax */
-	addl	$28, %esp
+	addl	$24, %esp
 
 /*
  * Jump to the decompressed kernel.
Index: linux-2.6/arch/x86/boot/compressed/head_64.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_64.S
+++ linux-2.6/arch/x86/boot/compressed/head_64.S
@@ -403,16 +403,13 @@ relocated:
  * Do the decompression, and jump to the new kernel..
  */
 	pushq	%rsi			/* Save the real mode argument */
-	movq	$z_run_size, %r9	/* size of kernel with .bss and .brk */
-	pushq	%r9
 	movq	%rsi, %rdi		/* real mode address */
 	leaq	boot_heap(%rip), %rsi	/* malloc area for uncompression */
 	leaq	input_data(%rip), %rdx  /* input_data */
 	movl	$z_input_len, %ecx	/* input_len */
 	movq	%rbp, %r8		/* output target address */
-	movq	$z_output_len, %r9	/* decompressed length, end of relocs */
+	movq	$z_output_len, %r9	/* decompressed length */
 	call	decompress_kernel	/* returns kernel location in %rax */
-	popq	%r9
 	popq	%rsi
 
 /*
Index: linux-2.6/arch/x86/boot/compressed/misc.c
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/misc.c
+++ linux-2.6/arch/x86/boot/compressed/misc.c
@@ -370,10 +370,10 @@ asmlinkage __visible void *decompress_ke
 				  unsigned char *input_data,
 				  unsigned long input_len,
 				  unsigned char *output,
-				  unsigned long output_len,
-				  unsigned long run_size)
+				  unsigned long output_len)
 {
 	unsigned char *output_orig = output;
+	unsigned long init_size;
 
 	real_mode = rmode;
 
@@ -396,15 +396,14 @@ asmlinkage __visible void *decompress_ke
 	free_mem_ptr     = heap;	/* Heap */
 	free_mem_end_ptr = heap + BOOT_HEAP_SIZE;
 
+	init_size = real_mode->hdr.init_size;
+
 	/*
-	 * The memory hole needed for the kernel is the larger of either
-	 * the entire decompressed kernel plus relocation table, or the
-	 * entire decompressed kernel plus .bss and .brk sections.
+	 * The memory hole needed for the kernel is init_size for running
+	 * and init_size is bigger than output_len always.
 	 */
 	output = choose_kernel_location(real_mode, input_data, input_len,
-					output,
-					output_len > run_size ? output_len
-							      : run_size);
+					output, init_size);
 
 	/* Validate memory location choices. */
 	if ((unsigned long)output & (MIN_KERNEL_ALIGN - 1))
Index: linux-2.6/arch/x86/boot/compressed/mkpiggy.c
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/mkpiggy.c
+++ linux-2.6/arch/x86/boot/compressed/mkpiggy.c
@@ -36,13 +36,11 @@ int main(int argc, char *argv[])
 	uint32_t olen;
 	long ilen;
 	unsigned long offs;
-	unsigned long run_size;
 	FILE *f = NULL;
 	int retval = 1;
 
-	if (argc < 3) {
-		fprintf(stderr, "Usage: %s compressed_file run_size\n",
-				argv[0]);
+	if (argc < 2) {
+		fprintf(stderr, "Usage: %s compressed_file\n", argv[0]);
 		goto bail;
 	}
 
@@ -76,7 +74,6 @@ int main(int argc, char *argv[])
 	offs += olen >> 12;	/* Add 8 bytes for each 32K block */
 	offs += 64*1024 + 128;	/* Add 64K + 128 bytes slack */
 	offs = (offs+4095) & ~4095; /* Round to a 4K boundary */
-	run_size = atoi(argv[2]);
 
 	printf(".section \".rodata..compressed\",\"a\",@progbits\n");
 	printf(".globl z_input_len\n");
@@ -88,8 +85,6 @@ int main(int argc, char *argv[])
 	/* z_extract_offset_negative allows simplification of head_32.S */
 	printf(".globl z_extract_offset_negative\n");
 	printf("z_extract_offset_negative = -0x%lx\n", offs);
-	printf(".globl z_run_size\n");
-	printf("z_run_size = %lu\n", run_size);
 
 	printf(".globl input_data, input_data_end\n");
 	printf("input_data:\n");
Index: linux-2.6/arch/x86/tools/calc_run_size.sh
===================================================================
--- linux-2.6.orig/arch/x86/tools/calc_run_size.sh
+++ /dev/null
@@ -1,42 +0,0 @@
-#!/bin/sh
-#
-# Calculate the amount of space needed to run the kernel, including room for
-# the .bss and .brk sections.
-#
-# Usage:
-# objdump -h a.out | sh calc_run_size.sh
-
-NUM='\([0-9a-fA-F]*[ \t]*\)'
-OUT=$(sed -n 's/^[ \t0-9]*.b[sr][sk][ \t]*'"$NUM$NUM$NUM$NUM"'.*/\1\4/p')
-if [ -z "$OUT" ] ; then
-	echo "Never found .bss or .brk file offset" >&2
-	exit 1
-fi
-
-OUT=$(echo ${OUT# })
-sizeA=$(printf "%d" 0x${OUT%% *})
-OUT=${OUT#* }
-offsetA=$(printf "%d" 0x${OUT%% *})
-OUT=${OUT#* }
-sizeB=$(printf "%d" 0x${OUT%% *})
-OUT=${OUT#* }
-offsetB=$(printf "%d" 0x${OUT%% *})
-
-run_size=$(( $offsetA + $sizeA + $sizeB ))
-
-# BFD linker shows the same file offset in ELF.
-if [ "$offsetA" -ne "$offsetB" ] ; then
-	# Gold linker shows them as consecutive.
-	endB=$(( $offsetB + $sizeB ))
-	if [ "$endB" != "$run_size" ] ; then
-		printf "sizeA: 0x%x\n" $sizeA >&2
-		printf "offsetA: 0x%x\n" $offsetA >&2
-		printf "sizeB: 0x%x\n" $sizeB >&2
-		printf "offsetB: 0x%x\n" $offsetB >&2
-		echo ".bss and .brk are non-contiguous" >&2
-		exit 1
-	fi
-fi
-
-printf "%d\n" $run_size
-exit 0

[-- Attachment #3: fix_decompress_offset.patch --]
[-- Type: text/x-patch, Size: 6418 bytes --]

Subject: [PATCH v2] x86, boot: keep data from boot stage to kernel stage.

bp found data from boot stage can not be used kernel stage.

Acctually those data area is overlapped with kernel bss stage, and clear_bss()
clear them before code in arch/x86/kernel/setup.c access them.

To make the data survive that later, we should avoid the overlapping.

Need to:
1. move compressed kernel close the end of buffer instead of middle of it.
2. extend init_size so no one from kernel bss and brk will touch the data
   region from boot/compressed/misc.c

-v2: add init_size in arch/x86/boot/header.S instead of BRK.

Fixes: f47233c2d34f ("x86/mm/ASLR: Propagate base load address calculation")
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Matt Fleming <matt.fleming@intel.com> 
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 arch/x86/boot/Makefile                 |    2 +-
 arch/x86/boot/compressed/head_32.S     |   11 +++++++++--
 arch/x86/boot/compressed/head_64.S     |    8 ++++++--
 arch/x86/boot/compressed/mkpiggy.c     |    3 ---
 arch/x86/boot/compressed/vmlinux.lds.S |    2 ++
 arch/x86/boot/header.S                 |    7 +++++--
 arch/x86/kernel/asm-offsets.c          |    1 +
 7 files changed, 24 insertions(+), 10 deletions(-)

Index: linux-2.6/arch/x86/boot/compressed/head_64.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_64.S
+++ linux-2.6/arch/x86/boot/compressed/head_64.S
@@ -102,7 +102,9 @@ ENTRY(startup_32)
 1:
 
 	/* Target address to relocate to for decompression */
-	addl	$z_extract_offset, %ebx
+	movl	BP_init_size(%esi), %eax
+	subl	$_end, %eax
+	addl	%eax, %ebx
 
 /*
  * Prepare for entering 64 bit mode
@@ -330,7 +332,9 @@ preferred_addr:
 1:
 
 	/* Target address to relocate to for decompression */
-	leaq	z_extract_offset(%rbp), %rbx
+	movl	BP_init_size(%rsi), %ebx
+	subl	$_end, %ebx
+	addq	%rbp, %rbx
 
 	/* Set up the stack */
 	leaq	boot_stack_end(%rbx), %rsp
Index: linux-2.6/arch/x86/kernel/asm-offsets.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/asm-offsets.c
+++ linux-2.6/arch/x86/kernel/asm-offsets.c
@@ -66,6 +66,7 @@ void common(void) {
 	OFFSET(BP_hardware_subarch, boot_params, hdr.hardware_subarch);
 	OFFSET(BP_version, boot_params, hdr.version);
 	OFFSET(BP_kernel_alignment, boot_params, hdr.kernel_alignment);
+	OFFSET(BP_init_size, boot_params, hdr.init_size);
 	OFFSET(BP_pref_address, boot_params, hdr.pref_address);
 	OFFSET(BP_code32_start, boot_params, hdr.code32_start);
 
Index: linux-2.6/arch/x86/boot/compressed/head_32.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/head_32.S
+++ linux-2.6/arch/x86/boot/compressed/head_32.S
@@ -148,7 +148,9 @@ preferred_addr:
 1:
 
 	/* Target address to relocate to for decompression */
-	addl	$z_extract_offset, %ebx
+	movl    BP_init_size(%esi), %eax
+	subl    $_end, %eax
+	addl    %eax, %ebx
 
 	/* Set up the stack */
 	leal	boot_stack_end(%ebx), %esp
@@ -209,8 +211,13 @@ relocated:
  */
 				/* push arguments for decompress_kernel: */
 	pushl	$z_output_len	/* decompressed length */
-	leal	z_extract_offset_negative(%ebx), %ebp
+
+	movl    BP_init_size(%esi), %eax
+	subl    $_end, %eax
+	movl    %ebx, %ebp
+	subl    %eax, %ebp
 	pushl	%ebp		/* output address */
+
 	pushl	$z_input_len	/* input_len */
 	leal	input_data(%ebx), %eax
 	pushl	%eax		/* input_data */
Index: linux-2.6/arch/x86/boot/compressed/mkpiggy.c
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/mkpiggy.c
+++ linux-2.6/arch/x86/boot/compressed/mkpiggy.c
@@ -82,9 +82,6 @@ int main(int argc, char *argv[])
 	printf("z_output_len = %lu\n", (unsigned long)olen);
 	printf(".globl z_extract_offset\n");
 	printf("z_extract_offset = 0x%lx\n", offs);
-	/* z_extract_offset_negative allows simplification of head_32.S */
-	printf(".globl z_extract_offset_negative\n");
-	printf("z_extract_offset_negative = -0x%lx\n", offs);
 
 	printf(".globl input_data, input_data_end\n");
 	printf("input_data:\n");
Index: linux-2.6/arch/x86/boot/Makefile
===================================================================
--- linux-2.6.orig/arch/x86/boot/Makefile
+++ linux-2.6/arch/x86/boot/Makefile
@@ -86,7 +86,7 @@ targets += voffset.h
 $(obj)/voffset.h: vmlinux FORCE
 	$(call if_changed,voffset)
 
-sed-zoffset := -e 's/^\([0-9a-fA-F]*\) [ABCDGRSTVW] \(startup_32\|startup_64\|efi32_stub_entry\|efi64_stub_entry\|efi_pe_entry\|input_data\|_end\|z_.*\)$$/\#define ZO_\2 0x\1/p'
+sed-zoffset := -e 's/^\([0-9a-fA-F]*\) [ABCDGRSTVW] \(startup_32\|startup_64\|efi32_stub_entry\|efi64_stub_entry\|efi_pe_entry\|input_data\|_end\|_rodata\|z_.*\)$$/\#define ZO_\2 0x\1/p'
 
 quiet_cmd_zoffset = ZOFFSET $@
       cmd_zoffset = $(NM) $< | sed -n $(sed-zoffset) > $@
Index: linux-2.6/arch/x86/boot/compressed/vmlinux.lds.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/compressed/vmlinux.lds.S
+++ linux-2.6/arch/x86/boot/compressed/vmlinux.lds.S
@@ -35,6 +35,7 @@ SECTIONS
 		*(.text.*)
 		_etext = . ;
 	}
+        . = ALIGN(PAGE_SIZE); /* keep ADDON_ZO_SIZE page aligned */
 	.rodata : {
 		_rodata = . ;
 		*(.rodata)	 /* read-only data */
@@ -70,5 +71,6 @@ SECTIONS
 		_epgtable = . ;
 	}
 #endif
+        . = ALIGN(PAGE_SIZE);  /* keep ADDON_ZO_SIZE page aligned */
 	_end = .;
 }
Index: linux-2.6/arch/x86/boot/header.S
===================================================================
--- linux-2.6.orig/arch/x86/boot/header.S
+++ linux-2.6/arch/x86/boot/header.S
@@ -440,12 +440,15 @@ setup_data:		.quad 0			# 64-bit physical
 
 pref_address:		.quad LOAD_PHYSICAL_ADDR	# preferred load addr
 
+# don't overlap data area of ZO with VO
+#define ADDON_ZO_SIZE (ZO__end - ZO__rodata)
+
 #define ZO_INIT_SIZE	(ZO__end - ZO_startup_32 + ZO_z_extract_offset)
 #define VO_INIT_SIZE	(VO__end - VO__text)
 #if ZO_INIT_SIZE > VO_INIT_SIZE
-#define INIT_SIZE ZO_INIT_SIZE
+#define INIT_SIZE (ZO_INIT_SIZE + ADDON_ZO_SIZE)
 #else
-#define INIT_SIZE VO_INIT_SIZE
+#define INIT_SIZE (VO_INIT_SIZE + ADDON_ZO_SIZE)
 #endif
 init_size:		.long INIT_SIZE		# kernel initialization size
 handover_offset:	.long 0			# Filled in by build.c

[-- Attachment #4: fix_kaslr.patch --]
[-- Type: text/x-patch, Size: 1405 bytes --]

Subject: [PATCH v2] x86, kaslr: get kaslr_enabled back correctly

commit f47233c2d34f ("x86/mm/ASLR: Propagate base load address calculation")
is using address as vaule for kaslr_enabled.

That will random kaslr_enabled get that set or cleared.
Will have problem for system really have kaslr enabled.

-v2: update changelog.

Fixes: f47233c2d34f ("x86/mm/ASLR: Propagate base load address calculation")
Cc: Matt Fleming <matt.fleming@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Kees Cook <keescook@chromium.org>
Cc: Jiri Kosina <jkosina@suse.cz>
Acked-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 arch/x86/kernel/setup.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

Index: linux-2.6/arch/x86/kernel/setup.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/setup.c
+++ linux-2.6/arch/x86/kernel/setup.c
@@ -429,7 +429,13 @@ static void __init reserve_initrd(void)
 
 static void __init parse_kaslr_setup(u64 pa_data, u32 data_len)
 {
-	kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));
+	/* kaslr_setup_data is defined in aslr.c */
+	unsigned char *data;
+	unsigned long offset = sizeof(struct setup_data);
+
+	data = early_memremap(pa_data, offset + 1);
+	kaslr_enabled = *(data + offset);
+	early_memunmap(data, offset + 1);
 }
 
 static void __init parse_setup_data(void)

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Borislav Petkov]

On Sun, Mar 01, 2015 at 04:23:51PM +0100, Ingo Molnar wrote:
> I think that's a different bug.
> 
> parse_kaslr_setup() is simply bogus, it does:
> 
>         kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));

Well, we found that while debugging the other issue too:

https://lkml.kernel.org/r/20150228105049.GA11038@pd.tnic

Scroll to the end.

> So I'm inclined to apply Yinghai's fix, with a better changelog that 
> explains what happened ...
> 
> Agreed?

Yes, this fix needs a commit message which actually explains what's it
fixing before it gets applied.

Thanks.

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Jiri Kosina]

On Sat, 28 Feb 2015, Yinghai Lu wrote:

> We should access variable with referrence instead of using physical
> address as value.
> 
> Cc: Matt Fleming <matt.fleming@intel.com>
> Cc: Borislav Petkov <bp@suse.de>
> Signed-off-by: Yinghai Lu <yinghai@kernel.org>

	Acked-by: Jiri Kosina <jkosina@suse.cz>


Thanks for fixing my brainfart. This is not fixing the problem with the 
later variable overwrite, but definitely needs to go in as well.

Thanks.

> ---
>  arch/x86/kernel/setup.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index 98dc931..05d444f 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -429,7 +429,13 @@ static void __init reserve_initrd(void)
>  
>  static void __init parse_kaslr_setup(u64 pa_data, u32 data_len)
>  {
> -	kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));
> +	/* kaslr_setup_data is defined in aslr.c */
> +	unsigned char *data;
> +	unsigned long offset = sizeof(struct setup_data);
> +
> +	data = early_memremap(pa_data, offset + 1);
> +	kaslr_enabled = *(data + offset);
> +	early_memunmap(data, offset + 1);
>  }
>  
>  static void __init parse_setup_data(void)

-- 
Jiri Kosina
SUSE Labs

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Matt Fleming]

On Sat, 28 Feb, at 06:17:32PM, Yinghai Lu wrote:
> We should access variable with referrence instead of using physical
> address as value.
> 
> Cc: Matt Fleming <matt.fleming@intel.com>
> Cc: Borislav Petkov <bp@suse.de>
> Signed-off-by: Yinghai Lu <yinghai@kernel.org>
> ---
>  arch/x86/kernel/setup.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index 98dc931..05d444f 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -429,7 +429,13 @@ static void __init reserve_initrd(void)
>  
>  static void __init parse_kaslr_setup(u64 pa_data, u32 data_len)
>  {
> -	kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));
> +	/* kaslr_setup_data is defined in aslr.c */
> +	unsigned char *data;
> +	unsigned long offset = sizeof(struct setup_data);
> +
> +	data = early_memremap(pa_data, offset + 1);
> +	kaslr_enabled = *(data + offset);
> +	early_memunmap(data, offset + 1);
>  }

ACK in principle, but could we change 'offset' to 'size' or 'len' along
with the commit log updates people have suggested?

-- 
Matt Fleming, Intel Open Source Technology Center

Re: [PATCH 1/8] x86, kaslr: get kaslr_enabled back correctly

[Matt Fleming]

On Sat, 28 Feb, at 06:17:32PM, Yinghai Lu wrote:
> We should access variable with referrence instead of using physical
> address as value.
> 
> Cc: Matt Fleming <matt.fleming-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
> Cc: Borislav Petkov <bp-l3A5Bk7waGM@public.gmane.org>
> Signed-off-by: Yinghai Lu <yinghai-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> ---
>  arch/x86/kernel/setup.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index 98dc931..05d444f 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -429,7 +429,13 @@ static void __init reserve_initrd(void)
>  
>  static void __init parse_kaslr_setup(u64 pa_data, u32 data_len)
>  {
> -	kaslr_enabled = (bool)(pa_data + sizeof(struct setup_data));
> +	/* kaslr_setup_data is defined in aslr.c */
> +	unsigned char *data;
> +	unsigned long offset = sizeof(struct setup_data);
> +
> +	data = early_memremap(pa_data, offset + 1);
> +	kaslr_enabled = *(data + offset);
> +	early_memunmap(data, offset + 1);
>  }

ACK in principle, but could we change 'offset' to 'size' or 'len' along
with the commit log updates people have suggested?

-- 
Matt Fleming, Intel Open Source Technology Center

[PATCH 2/8] x86: Kill E820_RESERVED_KERN

[Yinghai Lu]

Now we are using memblock to do early resource reserver/allocation
instead of using e820 map directly, and setup_data is reserved in
memblock early already.
Also kexec will generate setup_data and pass pointer to second kernel,
so second kernel will reserve setup_data by their own.

We can kill E820_RESERVED_KERN and not touch e820 map at all.

That will fix bug in mark_nonsave_region that can not handle that
case: E820_RAM and E820_RESERVED_KERN ranges are continuous and
boundary is not page aligned.

Bugzilla: https://bugzilla.opensuse.org/show_bug.cgi?id=913885
Reported-by: "Lee, Chun-Yi" <jlee@suse.com>
Tested-by: "Lee, Chun-Yi" <jlee@suse.com>
Cc: "Lee, Chun-Yi" <jlee@suse.com>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: stable@vger.kernel.org
---
 arch/x86/include/uapi/asm/e820.h |  9 ---------
 arch/x86/kernel/e820.c           |  6 ++----
 arch/x86/kernel/setup.c          | 26 --------------------------
 arch/x86/kernel/tboot.c          |  3 +--
 arch/x86/mm/init_64.c            | 11 ++++-------
 5 files changed, 7 insertions(+), 48 deletions(-)

diff --git a/arch/x86/include/uapi/asm/e820.h b/arch/x86/include/uapi/asm/e820.h
index d993e33..edc8a71 100644
--- a/arch/x86/include/uapi/asm/e820.h
+++ b/arch/x86/include/uapi/asm/e820.h
@@ -33,15 +33,6 @@
 #define E820_NVS	4
 #define E820_UNUSABLE	5
 
-
-/*
- * reserved RAM used by kernel itself
- * if CONFIG_INTEL_TXT is enabled, memory of this type will be
- * included in the S3 integrity calculation and so should not include
- * any memory that BIOS might alter over the S3 transition
- */
-#define E820_RESERVED_KERN        128
-
 #ifndef __ASSEMBLY__
 #include <linux/types.h>
 struct e820entry {
diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c
index 46201de..2a6bed9 100644
--- a/arch/x86/kernel/e820.c
+++ b/arch/x86/kernel/e820.c
@@ -134,7 +134,6 @@ static void __init e820_print_type(u32 type)
 {
 	switch (type) {
 	case E820_RAM:
-	case E820_RESERVED_KERN:
 		printk(KERN_CONT "usable");
 		break;
 	case E820_RESERVED:
@@ -688,7 +687,7 @@ void __init e820_mark_nosave_regions(unsigned long limit_pfn)
 			register_nosave_region(pfn, PFN_UP(ei->addr));
 
 		pfn = PFN_DOWN(ei->addr + ei->size);
-		if (ei->type != E820_RAM && ei->type != E820_RESERVED_KERN)
+		if (ei->type != E820_RAM)
 			register_nosave_region(PFN_UP(ei->addr), pfn);
 
 		if (pfn >= limit_pfn)
@@ -902,7 +901,6 @@ void __init finish_e820_parsing(void)
 static inline const char *e820_type_to_string(int e820_type)
 {
 	switch (e820_type) {
-	case E820_RESERVED_KERN:
 	case E820_RAM:	return "System RAM";
 	case E820_ACPI:	return "ACPI Tables";
 	case E820_NVS:	return "ACPI Non-volatile Storage";
@@ -1077,7 +1075,7 @@ void __init memblock_x86_fill(void)
 		if (end != (resource_size_t)end)
 			continue;
 
-		if (ei->type != E820_RAM && ei->type != E820_RESERVED_KERN)
+		if (ei->type != E820_RAM)
 			continue;
 
 		memblock_add(ei->addr, ei->size);
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 05d444f..c9b3e2f 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -473,30 +473,6 @@ static void __init parse_setup_data(void)
 	}
 }
 
-static void __init e820_reserve_setup_data(void)
-{
-	struct setup_data *data;
-	u64 pa_data;
-	int found = 0;
-
-	pa_data = boot_params.hdr.setup_data;
-	while (pa_data) {
-		data = early_memremap(pa_data, sizeof(*data));
-		e820_update_range(pa_data, sizeof(*data)+data->len,
-			 E820_RAM, E820_RESERVED_KERN);
-		found = 1;
-		pa_data = data->next;
-		early_iounmap(data, sizeof(*data));
-	}
-	if (!found)
-		return;
-
-	sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map);
-	memcpy(&e820_saved, &e820, sizeof(struct e820map));
-	printk(KERN_INFO "extended physical RAM map:\n");
-	e820_print_map("reserve setup_data");
-}
-
 static void __init memblock_x86_reserve_range_setup_data(void)
 {
 	struct setup_data *data;
@@ -1032,8 +1008,6 @@ void __init setup_arch(char **cmdline_p)
 		early_dump_pci_devices();
 #endif
 
-	/* update the e820_saved too */
-	e820_reserve_setup_data();
 	finish_e820_parsing();
 
 	if (efi_enabled(EFI_BOOT))
diff --git a/arch/x86/kernel/tboot.c b/arch/x86/kernel/tboot.c
index 91a4496..3c2752a 100644
--- a/arch/x86/kernel/tboot.c
+++ b/arch/x86/kernel/tboot.c
@@ -195,8 +195,7 @@ static int tboot_setup_sleep(void)
 	tboot->num_mac_regions = 0;
 
 	for (i = 0; i < e820.nr_map; i++) {
-		if ((e820.map[i].type != E820_RAM)
-		 && (e820.map[i].type != E820_RESERVED_KERN))
+		if (e820.map[i].type != E820_RAM)
 			continue;
 
 		add_mac_region(e820.map[i].addr, e820.map[i].size);
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index 30eb05a..19430d5 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -426,8 +426,7 @@ phys_pte_init(pte_t *pte_page, unsigned long addr, unsigned long end,
 		next = (addr & PAGE_MASK) + PAGE_SIZE;
 		if (addr >= end) {
 			if (!after_bootmem &&
-			    !e820_any_mapped(addr & PAGE_MASK, next, E820_RAM) &&
-			    !e820_any_mapped(addr & PAGE_MASK, next, E820_RESERVED_KERN))
+			    !e820_any_mapped(addr & PAGE_MASK, next, E820_RAM))
 				set_pte(pte, __pte(0));
 			continue;
 		}
@@ -473,9 +472,8 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long address, unsigned long end,
 
 		next = (address & PMD_MASK) + PMD_SIZE;
 		if (address >= end) {
-			if (!after_bootmem &&
-			    !e820_any_mapped(address & PMD_MASK, next, E820_RAM) &&
-			    !e820_any_mapped(address & PMD_MASK, next, E820_RESERVED_KERN))
+			if (!after_bootmem && !e820_any_mapped(
+					address & PMD_MASK, next, E820_RAM))
 				set_pmd(pmd, __pmd(0));
 			continue;
 		}
@@ -548,8 +546,7 @@ phys_pud_init(pud_t *pud_page, unsigned long addr, unsigned long end,
 		next = (addr & PUD_MASK) + PUD_SIZE;
 		if (addr >= end) {
 			if (!after_bootmem &&
-			    !e820_any_mapped(addr & PUD_MASK, next, E820_RAM) &&
-			    !e820_any_mapped(addr & PUD_MASK, next, E820_RESERVED_KERN))
+			    !e820_any_mapped(addr & PUD_MASK, next, E820_RAM))
 				set_pud(pud, __pud(0));
 			continue;
 		}
-- 
1.8.4.5

[PATCH 3/8] x86, efi: copy SETUP_EFI data and access directly

[Yinghai Lu]

the copy will be in __initdata, and it is small.

We can use pointer to access the setup_data instead of keeping on
early_memmap and early_memunmap everywhere.

Cc: Matt Fleming <matt.fleming@intel.com>
Cc: linux-efi@vger.kernel.org
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
---
 arch/x86/include/asm/efi.h     |  2 +-
 arch/x86/platform/efi/efi.c    | 13 ++-----------
 arch/x86/platform/efi/efi_64.c | 13 ++++++++++++-
 arch/x86/platform/efi/quirks.c | 23 ++++++-----------------
 4 files changed, 21 insertions(+), 30 deletions(-)

diff --git a/arch/x86/include/asm/efi.h b/arch/x86/include/asm/efi.h
index 25bce45..edbecd6 100644
--- a/arch/x86/include/asm/efi.h
+++ b/arch/x86/include/asm/efi.h
@@ -114,7 +114,7 @@ struct efi_setup_data {
 	u64 reserved[8];
 };
 
-extern u64 efi_setup;
+extern struct efi_setup_data *efi_setup;
 
 #ifdef CONFIG_EFI
 
diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c
index dbc8627..1cd38e8 100644
--- a/arch/x86/platform/efi/efi.c
+++ b/arch/x86/platform/efi/efi.c
@@ -68,7 +68,7 @@ static efi_config_table_type_t arch_tables[] __initdata = {
 	{NULL_GUID, NULL, NULL},
 };
 
-u64 efi_setup;		/* efi setup_data physical address */
+struct efi_setup_data *efi_setup __initdata; /* cached efi setup_data pointer */
 
 static int add_efi_memmap __initdata;
 static int __init setup_add_efi_memmap(char *arg)
@@ -225,20 +225,13 @@ static int __init efi_systab_init(void *phys)
 {
 	if (efi_enabled(EFI_64BIT)) {
 		efi_system_table_64_t *systab64;
-		struct efi_setup_data *data = NULL;
+		struct efi_setup_data *data = efi_setup;
 		u64 tmp = 0;
 
-		if (efi_setup) {
-			data = early_memremap(efi_setup, sizeof(*data));
-			if (!data)
-				return -ENOMEM;
-		}
 		systab64 = early_memremap((unsigned long)phys,
 					 sizeof(*systab64));
 		if (systab64 == NULL) {
 			pr_err("Couldn't map the system table!\n");
-			if (data)
-				early_memunmap(data, sizeof(*data));
 			return -ENOMEM;
 		}
 
@@ -271,8 +264,6 @@ static int __init efi_systab_init(void *phys)
 		tmp |= data ? data->tables : systab64->tables;
 
 		early_memunmap(systab64, sizeof(*systab64));
-		if (data)
-			early_memunmap(data, sizeof(*data));
 #ifdef CONFIG_X86_32
 		if (tmp >> 32) {
 			pr_err("EFI data located above 4GB, disabling EFI.\n");
diff --git a/arch/x86/platform/efi/efi_64.c b/arch/x86/platform/efi/efi_64.c
index 17e80d8..a541c6c 100644
--- a/arch/x86/platform/efi/efi_64.c
+++ b/arch/x86/platform/efi/efi_64.c
@@ -292,9 +292,20 @@ void __iomem *__init efi_ioremap(unsigned long phys_addr, unsigned long size,
 	return (void __iomem *)__va(phys_addr);
 }
 
+static struct efi_setup_data efi_setup_data __initdata;
+
 void __init parse_efi_setup(u64 phys_addr, u32 data_len)
 {
-	efi_setup = phys_addr + sizeof(struct setup_data);
+	struct efi_setup_data *data;
+
+	data = early_memremap(phys_addr + sizeof(struct setup_data),
+			      sizeof(*data));
+	if (!data)
+		return;
+
+	efi_setup_data = *data;
+	early_memunmap(data, sizeof(*data));
+	efi_setup = &efi_setup_data;
 }
 
 void __init efi_runtime_mkexec(void)
diff --git a/arch/x86/platform/efi/quirks.c b/arch/x86/platform/efi/quirks.c
index 1c7380d..45fec7d 100644
--- a/arch/x86/platform/efi/quirks.c
+++ b/arch/x86/platform/efi/quirks.c
@@ -203,9 +203,8 @@ void __init efi_free_boot_services(void)
  */
 int __init efi_reuse_config(u64 tables, int nr_tables)
 {
-	int i, sz, ret = 0;
+	int i, sz;
 	void *p, *tablep;
-	struct efi_setup_data *data;
 
 	if (!efi_setup)
 		return 0;
@@ -213,22 +212,15 @@ int __init efi_reuse_config(u64 tables, int nr_tables)
 	if (!efi_enabled(EFI_64BIT))
 		return 0;
 
-	data = early_memremap(efi_setup, sizeof(*data));
-	if (!data) {
-		ret = -ENOMEM;
-		goto out;
-	}
-
-	if (!data->smbios)
-		goto out_memremap;
+	if (!efi_setup->smbios)
+		return 0;
 
 	sz = sizeof(efi_config_table_64_t);
 
 	p = tablep = early_memremap(tables, nr_tables * sz);
 	if (!p) {
 		pr_err("Could not map Configuration table!\n");
-		ret = -ENOMEM;
-		goto out_memremap;
+		return -ENOMEM;
 	}
 
 	for (i = 0; i < efi.systab->nr_tables; i++) {
@@ -237,15 +229,12 @@ int __init efi_reuse_config(u64 tables, int nr_tables)
 		guid = ((efi_config_table_64_t *)p)->guid;
 
 		if (!efi_guidcmp(guid, SMBIOS_TABLE_GUID))
-			((efi_config_table_64_t *)p)->table = data->smbios;
+			((efi_config_table_64_t *)p)->table = efi_setup->smbios;
 		p += sz;
 	}
 	early_memunmap(tablep, nr_tables * sz);
 
-out_memremap:
-	early_memunmap(data, sizeof(*data));
-out:
-	return ret;
+	return 0;
 }
 
 void __init efi_apply_memmap_quirks(void)
-- 
1.8.4.5

[PATCH 4/8] x86, of: let add_dtb reserve by itself

[Yinghai Lu]

We will not reserve setup_data in general code. Every handler
need to reserve and copy.

Current dtd handling already have code copying, just add reserve code ...

also simplify code a bit with storing real dtb size.

Cc: Rob Herring <robh@kernel.org>
Cc: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
---
 arch/x86/include/asm/prom.h  |  9 ++++++---
 arch/x86/kernel/devicetree.c | 39 +++++++++++++++++++++------------------
 2 files changed, 27 insertions(+), 21 deletions(-)

diff --git a/arch/x86/include/asm/prom.h b/arch/x86/include/asm/prom.h
index 1d081ac..fb716eddc 100644
--- a/arch/x86/include/asm/prom.h
+++ b/arch/x86/include/asm/prom.h
@@ -24,17 +24,20 @@
 
 #ifdef CONFIG_OF
 extern int of_ioapic;
-extern u64 initial_dtb;
-extern void add_dtb(u64 data);
 void x86_of_pci_init(void);
 void x86_dtb_init(void);
 #else
-static inline void add_dtb(u64 data) { }
 static inline void x86_of_pci_init(void) { }
 static inline void x86_dtb_init(void) { }
 #define of_ioapic 0
 #endif
 
+#ifdef CONFIG_OF_FLATTREE
+extern void add_dtb(u64 data);
+#else
+static inline void add_dtb(u64 data) { }
+#endif
+
 extern char cmd_line[COMMAND_LINE_SIZE];
 
 #endif /* __ASSEMBLY__ */
diff --git a/arch/x86/kernel/devicetree.c b/arch/x86/kernel/devicetree.c
index 3d35033..cc2fb61 100644
--- a/arch/x86/kernel/devicetree.c
+++ b/arch/x86/kernel/devicetree.c
@@ -2,6 +2,7 @@
  * Architecture specific OF callbacks.
  */
 #include <linux/bootmem.h>
+#include <linux/memblock.h>
 #include <linux/export.h>
 #include <linux/io.h>
 #include <linux/irqdomain.h>
@@ -23,7 +24,6 @@
 #include <asm/setup.h>
 #include <asm/i8259.h>
 
-__initdata u64 initial_dtb;
 char __initdata cmd_line[COMMAND_LINE_SIZE];
 
 int __initdata of_ioapic;
@@ -43,11 +43,23 @@ void * __init early_init_dt_alloc_memory_arch(u64 size, u64 align)
 	return __alloc_bootmem(size, align, __pa(MAX_DMA_ADDRESS));
 }
 
+#ifdef CONFIG_OF_FLATTREE
+static u64 initial_dtb __initdata;
+static u32 initial_dtb_size __initdata;
 void __init add_dtb(u64 data)
 {
+	u32 map_len;
+
 	initial_dtb = data + offsetof(struct setup_data, data);
-}
 
+	map_len = max(PAGE_SIZE - (initial_dtb & ~PAGE_MASK), (u64)128);
+	initial_boot_params = early_memremap(initial_dtb, map_len);
+	initial_dtb_size = of_get_flat_dt_size();
+	early_memunmap(initial_boot_params, map_len);
+	initial_boot_params = NULL;
+	memblock_reserve(initial_dtb, initial_dtb_size);
+}
+#endif
 /*
  * CE4100 ids. Will be moved to machine_device_initcall() once we have it.
  */
@@ -272,31 +284,22 @@ static void __init dtb_apic_setup(void)
 	dtb_ioapic_setup();
 }
 
-#ifdef CONFIG_OF_FLATTREE
 static void __init x86_flattree_get_config(void)
 {
-	u32 size, map_len;
+#ifdef CONFIG_OF_FLATTREE
 	void *dt;
 
 	if (!initial_dtb)
 		return;
 
-	map_len = max(PAGE_SIZE - (initial_dtb & ~PAGE_MASK), (u64)128);
-
-	initial_boot_params = dt = early_memremap(initial_dtb, map_len);
-	size = of_get_flat_dt_size();
-	if (map_len < size) {
-		early_iounmap(dt, map_len);
-		initial_boot_params = dt = early_memremap(initial_dtb, size);
-		map_len = size;
-	}
-
+	initial_boot_params = dt = early_memremap(initial_dtb,
+						  initial_dtb_size);
 	unflatten_and_copy_device_tree();
-	early_iounmap(dt, map_len);
-}
-#else
-static inline void x86_flattree_get_config(void) { }
+	early_memunmap(dt, initial_dtb_size);
+
+	memblock_free(initial_dtb, initial_dtb_size);
 #endif
+}
 
 void __init x86_dtb_init(void)
 {
-- 
1.8.4.5

[PATCH 5/8] x86, boot: Add add_pci handler for SETUP_PCI

[Yinghai Lu]

Let it reserve setup_data, and keep it's own list.

Also clear the hdr.setup_data, as all handler will handle or
reserve setup_data locally already.

Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Matt Fleming <matt.fleming@intel.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
---
 arch/x86/include/asm/pci.h |  2 ++
 arch/x86/kernel/setup.c    |  6 ++++++
 arch/x86/pci/common.c      | 42 ++++++++++++++++++++++++++++--------------
 3 files changed, 36 insertions(+), 14 deletions(-)

diff --git a/arch/x86/include/asm/pci.h b/arch/x86/include/asm/pci.h
index 4e370a5..aa25a22 100644
--- a/arch/x86/include/asm/pci.h
+++ b/arch/x86/include/asm/pci.h
@@ -155,4 +155,6 @@ struct pci_setup_rom {
 	uint8_t romdata[0];
 };
 
+void add_pci(u64 pa_data);
+
 #endif /* _ASM_X86_PCI_H */
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index c9b3e2f..93c0adb 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -460,6 +460,9 @@ static void __init parse_setup_data(void)
 		case SETUP_DTB:
 			add_dtb(pa_data);
 			break;
+		case SETUP_PCI:
+			add_pci(pa_data);
+			break;
 		case SETUP_EFI:
 			parse_efi_setup(pa_data, data_len);
 			break;
@@ -467,10 +470,13 @@ static void __init parse_setup_data(void)
 			parse_kaslr_setup(pa_data, data_len);
 			break;
 		default:
+			pr_warn("Unknown setup_data type: %d ignored!\n",
+				data_type);
 			break;
 		}
 		pa_data = pa_next;
 	}
+	boot_params.hdr.setup_data = 0; /* all done */
 }
 
 static void __init memblock_x86_reserve_range_setup_data(void)
diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c
index 3d2612b..4846db7 100644
--- a/arch/x86/pci/common.c
+++ b/arch/x86/pci/common.c
@@ -9,6 +9,7 @@
 #include <linux/pci-acpi.h>
 #include <linux/ioport.h>
 #include <linux/init.h>
+#include <linux/memblock.h>
 #include <linux/dmi.h>
 #include <linux/slab.h>
 
@@ -667,31 +668,44 @@ unsigned int pcibios_assign_all_busses(void)
 	return (pci_probe & PCI_ASSIGN_ALL_BUSSES) ? 1 : 0;
 }
 
+static u64 pci_setup_data;
+void __init add_pci(u64 pa_data)
+{
+	struct setup_data *data;
+
+	data = early_memremap(pa_data, sizeof(*data));
+	memblock_reserve(pa_data, sizeof(*data) + data->len);
+	data->next = pci_setup_data;
+	pci_setup_data = pa_data;
+	early_memunmap(data, sizeof(*data));
+}
+
 int pcibios_add_device(struct pci_dev *dev)
 {
 	struct setup_data *data;
 	struct pci_setup_rom *rom;
 	u64 pa_data;
 
-	pa_data = boot_params.hdr.setup_data;
+	pa_data = pci_setup_data;
 	while (pa_data) {
 		data = ioremap(pa_data, sizeof(*rom));
 		if (!data)
 			return -ENOMEM;
 
-		if (data->type == SETUP_PCI) {
-			rom = (struct pci_setup_rom *)data;
-
-			if ((pci_domain_nr(dev->bus) == rom->segment) &&
-			    (dev->bus->number == rom->bus) &&
-			    (PCI_SLOT(dev->devfn) == rom->device) &&
-			    (PCI_FUNC(dev->devfn) == rom->function) &&
-			    (dev->vendor == rom->vendor) &&
-			    (dev->device == rom->devid)) {
-				dev->rom = pa_data +
-				      offsetof(struct pci_setup_rom, romdata);
-				dev->romlen = rom->pcilen;
-			}
+		rom = (struct pci_setup_rom *)data;
+
+		if ((pci_domain_nr(dev->bus) == rom->segment) &&
+		    (dev->bus->number == rom->bus) &&
+		    (PCI_SLOT(dev->devfn) == rom->device) &&
+		    (PCI_FUNC(dev->devfn) == rom->function) &&
+		    (dev->vendor == rom->vendor) &&
+		    (dev->device == rom->devid)) {
+			dev->rom = pa_data +
+			      offsetof(struct pci_setup_rom, romdata);
+			dev->romlen = rom->pcilen;
+			dev_printk(KERN_DEBUG, &dev->dev, "set rom to [%#010lx, %#010lx] via SETUP_PCI\n",
+				   (unsigned long)dev->rom,
+				   (unsigned long)(dev->rom + dev->romlen - 1));
 		}
 		pa_data = data->next;
 		iounmap(data);
-- 
1.8.4.5

[PATCH 6/8] x86: kill not used setup_data handling code

[Yinghai Lu]

Cc: Matt Fleming <matt.fleming@intel.com>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
---
 arch/x86/kernel/kdebugfs.c | 142 ---------------------------------------------
 arch/x86/kernel/setup.c    |  17 ------
 2 files changed, 159 deletions(-)

diff --git a/arch/x86/kernel/kdebugfs.c b/arch/x86/kernel/kdebugfs.c
index dc1404b..c8ca86c 100644
--- a/arch/x86/kernel/kdebugfs.c
+++ b/arch/x86/kernel/kdebugfs.c
@@ -21,142 +21,6 @@ struct dentry *arch_debugfs_dir;
 EXPORT_SYMBOL(arch_debugfs_dir);
 
 #ifdef CONFIG_DEBUG_BOOT_PARAMS
-struct setup_data_node {
-	u64 paddr;
-	u32 type;
-	u32 len;
-};
-
-static ssize_t setup_data_read(struct file *file, char __user *user_buf,
-			       size_t count, loff_t *ppos)
-{
-	struct setup_data_node *node = file->private_data;
-	unsigned long remain;
-	loff_t pos = *ppos;
-	struct page *pg;
-	void *p;
-	u64 pa;
-
-	if (pos < 0)
-		return -EINVAL;
-
-	if (pos >= node->len)
-		return 0;
-
-	if (count > node->len - pos)
-		count = node->len - pos;
-
-	pa = node->paddr + sizeof(struct setup_data) + pos;
-	pg = pfn_to_page((pa + count - 1) >> PAGE_SHIFT);
-	if (PageHighMem(pg)) {
-		p = ioremap_cache(pa, count);
-		if (!p)
-			return -ENXIO;
-	} else
-		p = __va(pa);
-
-	remain = copy_to_user(user_buf, p, count);
-
-	if (PageHighMem(pg))
-		iounmap(p);
-
-	if (remain)
-		return -EFAULT;
-
-	*ppos = pos + count;
-
-	return count;
-}
-
-static const struct file_operations fops_setup_data = {
-	.read		= setup_data_read,
-	.open		= simple_open,
-	.llseek		= default_llseek,
-};
-
-static int __init
-create_setup_data_node(struct dentry *parent, int no,
-		       struct setup_data_node *node)
-{
-	struct dentry *d, *type, *data;
-	char buf[16];
-
-	sprintf(buf, "%d", no);
-	d = debugfs_create_dir(buf, parent);
-	if (!d)
-		return -ENOMEM;
-
-	type = debugfs_create_x32("type", S_IRUGO, d, &node->type);
-	if (!type)
-		goto err_dir;
-
-	data = debugfs_create_file("data", S_IRUGO, d, node, &fops_setup_data);
-	if (!data)
-		goto err_type;
-
-	return 0;
-
-err_type:
-	debugfs_remove(type);
-err_dir:
-	debugfs_remove(d);
-	return -ENOMEM;
-}
-
-static int __init create_setup_data_nodes(struct dentry *parent)
-{
-	struct setup_data_node *node;
-	struct setup_data *data;
-	int error;
-	struct dentry *d;
-	struct page *pg;
-	u64 pa_data;
-	int no = 0;
-
-	d = debugfs_create_dir("setup_data", parent);
-	if (!d)
-		return -ENOMEM;
-
-	pa_data = boot_params.hdr.setup_data;
-
-	while (pa_data) {
-		node = kmalloc(sizeof(*node), GFP_KERNEL);
-		if (!node) {
-			error = -ENOMEM;
-			goto err_dir;
-		}
-
-		pg = pfn_to_page((pa_data+sizeof(*data)-1) >> PAGE_SHIFT);
-		if (PageHighMem(pg)) {
-			data = ioremap_cache(pa_data, sizeof(*data));
-			if (!data) {
-				kfree(node);
-				error = -ENXIO;
-				goto err_dir;
-			}
-		} else
-			data = __va(pa_data);
-
-		node->paddr = pa_data;
-		node->type = data->type;
-		node->len = data->len;
-		error = create_setup_data_node(d, no, node);
-		pa_data = data->next;
-
-		if (PageHighMem(pg))
-			iounmap(data);
-		if (error)
-			goto err_dir;
-		no++;
-	}
-
-	return 0;
-
-err_dir:
-	debugfs_remove(d);
-	return error;
-}
-
 static struct debugfs_blob_wrapper boot_params_blob = {
 	.data		= &boot_params,
 	.size		= sizeof(boot_params),
@@ -181,14 +45,8 @@ static int __init boot_params_kdebugfs_init(void)
 	if (!data)
 		goto err_version;
 
-	error = create_setup_data_nodes(dbp);
-	if (error)
-		goto err_data;
-
 	return 0;
 
-err_data:
-	debugfs_remove(data);
 err_version:
 	debugfs_remove(version);
 err_dir:
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 93c0adb..b9f1687 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -479,20 +479,6 @@ static void __init parse_setup_data(void)
 	boot_params.hdr.setup_data = 0; /* all done */
 }
 
-static void __init memblock_x86_reserve_range_setup_data(void)
-{
-	struct setup_data *data;
-	u64 pa_data;
-
-	pa_data = boot_params.hdr.setup_data;
-	while (pa_data) {
-		data = early_memremap(pa_data, sizeof(*data));
-		memblock_reserve(pa_data, sizeof(*data) + data->len);
-		pa_data = data->next;
-		early_iounmap(data, sizeof(*data));
-	}
-}
-
 /*
  * --------- Crashkernel reservation ------------------------------
  */
@@ -999,9 +985,6 @@ void __init setup_arch(char **cmdline_p)
 
 	x86_report_nx();
 
-	/* after early param, so could get panic from serial */
-	memblock_x86_reserve_range_setup_data();
-
 	if (acpi_mps_check()) {
 #ifdef CONFIG_X86_LOCAL_APIC
 		disable_apic = 1;
-- 
1.8.4.5

[PATCH 6/8] x86: kill not used setup_data handling code

[Yinghai Lu]

Cc: Matt Fleming <matt.fleming-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Yinghai Lu <yinghai-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
---
 arch/x86/kernel/kdebugfs.c | 142 ---------------------------------------------
 arch/x86/kernel/setup.c    |  17 ------
 2 files changed, 159 deletions(-)

diff --git a/arch/x86/kernel/kdebugfs.c b/arch/x86/kernel/kdebugfs.c
index dc1404b..c8ca86c 100644
--- a/arch/x86/kernel/kdebugfs.c
+++ b/arch/x86/kernel/kdebugfs.c
@@ -21,142 +21,6 @@ struct dentry *arch_debugfs_dir;
 EXPORT_SYMBOL(arch_debugfs_dir);
 
 #ifdef CONFIG_DEBUG_BOOT_PARAMS
-struct setup_data_node {
-	u64 paddr;
-	u32 type;
-	u32 len;
-};
-
-static ssize_t setup_data_read(struct file *file, char __user *user_buf,
-			       size_t count, loff_t *ppos)
-{
-	struct setup_data_node *node = file->private_data;
-	unsigned long remain;
-	loff_t pos = *ppos;
-	struct page *pg;
-	void *p;
-	u64 pa;
-
-	if (pos < 0)
-		return -EINVAL;
-
-	if (pos >= node->len)
-		return 0;
-
-	if (count > node->len - pos)
-		count = node->len - pos;
-
-	pa = node->paddr + sizeof(struct setup_data) + pos;
-	pg = pfn_to_page((pa + count - 1) >> PAGE_SHIFT);
-	if (PageHighMem(pg)) {
-		p = ioremap_cache(pa, count);
-		if (!p)
-			return -ENXIO;
-	} else
-		p = __va(pa);
-
-	remain = copy_to_user(user_buf, p, count);
-
-	if (PageHighMem(pg))
-		iounmap(p);
-
-	if (remain)
-		return -EFAULT;
-
-	*ppos = pos + count;
-
-	return count;
-}
-
-static const struct file_operations fops_setup_data = {
-	.read		= setup_data_read,
-	.open		= simple_open,
-	.llseek		= default_llseek,
-};
-
-static int __init
-create_setup_data_node(struct dentry *parent, int no,
-		       struct setup_data_node *node)
-{
-	struct dentry *d, *type, *data;
-	char buf[16];
-
-	sprintf(buf, "%d", no);
-	d = debugfs_create_dir(buf, parent);
-	if (!d)
-		return -ENOMEM;
-
-	type = debugfs_create_x32("type", S_IRUGO, d, &node->type);
-	if (!type)
-		goto err_dir;
-
-	data = debugfs_create_file("data", S_IRUGO, d, node, &fops_setup_data);
-	if (!data)
-		goto err_type;
-
-	return 0;
-
-err_type:
-	debugfs_remove(type);
-err_dir:
-	debugfs_remove(d);
-	return -ENOMEM;
-}
-
-static int __init create_setup_data_nodes(struct dentry *parent)
-{
-	struct setup_data_node *node;
-	struct setup_data *data;
-	int error;
-	struct dentry *d;
-	struct page *pg;
-	u64 pa_data;
-	int no = 0;
-
-	d = debugfs_create_dir("setup_data", parent);
-	if (!d)
-		return -ENOMEM;
-
-	pa_data = boot_params.hdr.setup_data;
-
-	while (pa_data) {
-		node = kmalloc(sizeof(*node), GFP_KERNEL);
-		if (!node) {
-			error = -ENOMEM;
-			goto err_dir;
-		}
-
-		pg = pfn_to_page((pa_data+sizeof(*data)-1) >> PAGE_SHIFT);
-		if (PageHighMem(pg)) {
-			data = ioremap_cache(pa_data, sizeof(*data));
-			if (!data) {
-				kfree(node);
-				error = -ENXIO;
-				goto err_dir;
-			}
-		} else
-			data = __va(pa_data);
-
-		node->paddr = pa_data;
-		node->type = data->type;
-		node->len = data->len;
-		error = create_setup_data_node(d, no, node);
-		pa_data = data->next;
-
-		if (PageHighMem(pg))
-			iounmap(data);
-		if (error)
-			goto err_dir;
-		no++;
-	}
-
-	return 0;
-
-err_dir:
-	debugfs_remove(d);
-	return error;
-}
-
 static struct debugfs_blob_wrapper boot_params_blob = {
 	.data		= &boot_params,
 	.size		= sizeof(boot_params),
@@ -181,14 +45,8 @@ static int __init boot_params_kdebugfs_init(void)
 	if (!data)
 		goto err_version;
 
-	error = create_setup_data_nodes(dbp);
-	if (error)
-		goto err_data;
-
 	return 0;
 
-err_data:
-	debugfs_remove(data);
 err_version:
 	debugfs_remove(version);
 err_dir:
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 93c0adb..b9f1687 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -479,20 +479,6 @@ static void __init parse_setup_data(void)
 	boot_params.hdr.setup_data = 0; /* all done */
 }
 
-static void __init memblock_x86_reserve_range_setup_data(void)
-{
-	struct setup_data *data;
-	u64 pa_data;
-
-	pa_data = boot_params.hdr.setup_data;
-	while (pa_data) {
-		data = early_memremap(pa_data, sizeof(*data));
-		memblock_reserve(pa_data, sizeof(*data) + data->len);
-		pa_data = data->next;
-		early_iounmap(data, sizeof(*data));
-	}
-}
-
 /*
  * --------- Crashkernel reservation ------------------------------
  */
@@ -999,9 +985,6 @@ void __init setup_arch(char **cmdline_p)
 
 	x86_report_nx();
 
-	/* after early param, so could get panic from serial */
-	memblock_x86_reserve_range_setup_data();
-
 	if (acpi_mps_check()) {
 #ifdef CONFIG_X86_LOCAL_APIC
 		disable_apic = 1;
-- 
1.8.4.5

[PATCH 7/8] x86, pci: convert SETUP_PCI data to list

[Yinghai Lu]

So we could avoid ioremap every time later.

Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
---
 arch/x86/pci/common.c | 77 +++++++++++++++++++++++++++++++++++++++++----------
 1 file changed, 63 insertions(+), 14 deletions(-)

diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c
index 4846db7..7b6fb94 100644
--- a/arch/x86/pci/common.c
+++ b/arch/x86/pci/common.c
@@ -680,10 +680,41 @@ void __init add_pci(u64 pa_data)
 	early_memunmap(data, sizeof(*data));
 }
 
-int pcibios_add_device(struct pci_dev *dev)
+struct firmware_setup_pci_entry {
+	struct list_head list;
+	uint16_t vendor;
+	uint16_t devid;
+	uint64_t pcilen;
+	unsigned long segment;
+	unsigned long bus;
+	unsigned long device;
+	unsigned long function;
+	phys_addr_t rom;
+};
+
+static LIST_HEAD(setup_pci_entries);
+
+static void __init firmware_setup_pci_add_entry(struct pci_setup_rom *rom,
+					u64 pa_data,
+					struct firmware_setup_pci_entry *entry)
+{
+	entry->segment = rom->segment;
+	entry->bus = rom->bus;
+	entry->device = rom->device;
+	entry->function = rom->function;
+	entry->vendor = rom->vendor;
+	entry->devid = rom->devid;
+	entry->rom = pa_data + offsetof(struct pci_setup_rom, romdata);
+	entry->pcilen = rom->pcilen;
+
+	list_add(&entry->list, &setup_pci_entries);
+}
+
+static __init int fill_setup_pci_entries(void)
 {
 	struct setup_data *data;
 	struct pci_setup_rom *rom;
+	struct firmware_setup_pci_entry *entry;
 	u64 pa_data;
 
 	pa_data = pci_setup_data;
@@ -692,24 +723,42 @@ int pcibios_add_device(struct pci_dev *dev)
 		if (!data)
 			return -ENOMEM;
 
-		rom = (struct pci_setup_rom *)data;
-
-		if ((pci_domain_nr(dev->bus) == rom->segment) &&
-		    (dev->bus->number == rom->bus) &&
-		    (PCI_SLOT(dev->devfn) == rom->device) &&
-		    (PCI_FUNC(dev->devfn) == rom->function) &&
-		    (dev->vendor == rom->vendor) &&
-		    (dev->device == rom->devid)) {
-			dev->rom = pa_data +
-			      offsetof(struct pci_setup_rom, romdata);
-			dev->romlen = rom->pcilen;
+		entry = kzalloc(sizeof(struct firmware_setup_pci_entry),
+				GFP_ATOMIC);
+		if (!entry) {
+			iounmap(data);
+			return -ENOMEM;
+		}
+
+		firmware_setup_pci_add_entry((struct pci_setup_rom *)data,
+					     pa_data, entry);
+		pa_data = data->next;
+		iounmap(data);
+	}
+
+	return 0;
+}
+postcore_initcall(fill_setup_pci_entries);
+
+int pcibios_add_device(struct pci_dev *dev)
+{
+	struct firmware_setup_pci_entry *entry;
+
+	list_for_each_entry(entry, &setup_pci_entries, list) {
+		if ((pci_domain_nr(dev->bus) == entry->segment) &&
+		    (dev->bus->number == entry->bus) &&
+		    (PCI_SLOT(dev->devfn) == entry->device) &&
+		    (PCI_FUNC(dev->devfn) == entry->function) &&
+		    (dev->vendor == entry->vendor) &&
+		    (dev->device == entry->devid)) {
+			dev->rom = entry->rom;
+			dev->romlen = entry->pcilen;
 			dev_printk(KERN_DEBUG, &dev->dev, "set rom to [%#010lx, %#010lx] via SETUP_PCI\n",
 				   (unsigned long)dev->rom,
 				   (unsigned long)(dev->rom + dev->romlen - 1));
 		}
-		pa_data = data->next;
-		iounmap(data);
 	}
+
 	return 0;
 }
 
-- 
1.8.4.5

[PATCH 8/8] x86, pci: export SETUP_PCI data via sysfs

[Yinghai Lu]

So we could let kexec-tools to rebuild SETUP_PCI and pass it to
second kernel.
Now kexec-tools only can build SETUP_EFI and SETUP_E820EXT.

Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
---
 arch/x86/pci/common.c | 188 ++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 188 insertions(+)

diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c
index 7b6fb94..bf0209a 100644
--- a/arch/x86/pci/common.c
+++ b/arch/x86/pci/common.c
@@ -682,6 +682,8 @@ void __init add_pci(u64 pa_data)
 
 struct firmware_setup_pci_entry {
 	struct list_head list;
+	struct kobject kobj;
+	struct bin_attribute *rom_attr;
 	uint16_t vendor;
 	uint16_t devid;
 	uint64_t pcilen;
@@ -762,6 +764,192 @@ int pcibios_add_device(struct pci_dev *dev)
 	return 0;
 }
 
+#ifdef CONFIG_SYSFS
+static inline struct firmware_setup_pci_entry *
+to_setup_pci_entry(struct kobject *kobj)
+{
+	return container_of(kobj, struct firmware_setup_pci_entry, kobj);
+}
+
+static ssize_t vendor_show(struct firmware_setup_pci_entry *entry, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "0x%04llx\n",
+			(unsigned long long)entry->vendor);
+}
+
+static ssize_t devid_show(struct firmware_setup_pci_entry *entry, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "0x%04llx\n",
+			(unsigned long long)entry->devid);
+}
+
+static ssize_t pcilen_show(struct firmware_setup_pci_entry *entry, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "0x%llx\n",
+			(unsigned long long)entry->pcilen);
+}
+
+static ssize_t segment_show(struct firmware_setup_pci_entry *entry, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "0x%04llx\n",
+			(unsigned long long)entry->segment);
+}
+
+static ssize_t bus_show(struct firmware_setup_pci_entry *entry, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "0x%02llx\n",
+			(unsigned long long)entry->bus);
+}
+
+static ssize_t device_show(struct firmware_setup_pci_entry *entry, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "0x%02llx\n",
+			(unsigned long long)entry->device);
+}
+
+static ssize_t function_show(struct firmware_setup_pci_entry *entry, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "0x%1llx\n",
+			(unsigned long long)entry->function);
+}
+
+struct setup_pci_attribute {
+	struct attribute attr;
+	ssize_t (*show)(struct firmware_setup_pci_entry *entry, char *buf);
+};
+
+static inline struct setup_pci_attribute *to_setup_pci_attr(
+							struct attribute *attr)
+{
+	return container_of(attr, struct setup_pci_attribute, attr);
+}
+
+static ssize_t setup_pci_attr_show(struct kobject *kobj,
+				   struct attribute *attr, char *buf)
+{
+	struct firmware_setup_pci_entry *entry = to_setup_pci_entry(kobj);
+	struct setup_pci_attribute *setup_pci_attr = to_setup_pci_attr(attr);
+
+	return setup_pci_attr->show(entry, buf);
+}
+
+static struct setup_pci_attribute setup_pci_vendor_attr = __ATTR_RO(vendor);
+static struct setup_pci_attribute setup_pci_devid_attr = __ATTR_RO(devid);
+static struct setup_pci_attribute setup_pci_pcilen_attr = __ATTR_RO(pcilen);
+static struct setup_pci_attribute setup_pci_segment_attr = __ATTR_RO(segment);
+static struct setup_pci_attribute setup_pci_bus_attr = __ATTR_RO(bus);
+static struct setup_pci_attribute setup_pci_device_attr = __ATTR_RO(device);
+static struct setup_pci_attribute setup_pci_function_attr = __ATTR_RO(function);
+
+/*
+ * These are default attributes that are added for every memmap entry.
+ */
+static struct attribute *def_attrs[] = {
+	&setup_pci_vendor_attr.attr,
+	&setup_pci_devid_attr.attr,
+	&setup_pci_pcilen_attr.attr,
+	&setup_pci_segment_attr.attr,
+	&setup_pci_bus_attr.attr,
+	&setup_pci_device_attr.attr,
+	&setup_pci_function_attr.attr,
+	NULL
+};
+
+static const struct sysfs_ops setup_pci_attr_ops = {
+	.show = setup_pci_attr_show,
+};
+
+static struct kobj_type __refdata setup_pci_ktype = {
+	.sysfs_ops      = &setup_pci_attr_ops,
+	.default_attrs  = def_attrs,
+};
+
+static ssize_t setup_pci_rom_read(struct file *filp, struct kobject *kobj,
+				  struct bin_attribute *bin_attr, char *buf,
+				  loff_t off, size_t count)
+{
+	struct firmware_setup_pci_entry *entry = to_setup_pci_entry(kobj);
+
+	if (off >= entry->pcilen)
+		count = 0;
+	else {
+		unsigned long start_pfn, end_pfn;
+		void *rom;
+
+		if (off + count > entry->pcilen)
+			count = entry->pcilen - off;
+
+		start_pfn = PFN_DOWN(entry->rom + off);
+		end_pfn = PFN_UP(entry->rom + off + count);
+		if (pfn_range_is_mapped(start_pfn, end_pfn)) {
+			rom = phys_to_virt(entry->rom);
+			memcpy(buf, rom + off, count);
+		} else {
+			rom = ioremap(entry->rom + off, count);
+			if (rom) {
+				memcpy_fromio(buf, rom, count);
+				iounmap(rom);
+			} else
+				count = 0;
+		}
+	}
+
+	return count;
+}
+
+static int __init add_sysfs_fw_setup_pci_entry(
+					struct firmware_setup_pci_entry *entry)
+{
+	int retval = 0;
+	static int setup_pci_entries_nr;
+	static struct kset *setup_pci_kset;
+	struct bin_attribute *attr;
+
+	kobject_init(&entry->kobj, &setup_pci_ktype);
+
+	if (!setup_pci_kset) {
+		setup_pci_kset = kset_create_and_add("setup_pci", NULL,
+						     firmware_kobj);
+		if (!setup_pci_kset)
+			return -ENOMEM;
+	}
+
+	entry->kobj.kset = setup_pci_kset;
+	retval = kobject_add(&entry->kobj, NULL, "%d", setup_pci_entries_nr++);
+	if (retval) {
+		kobject_put(&entry->kobj);
+		return retval;
+	}
+
+	attr = kzalloc(sizeof(*attr), GFP_ATOMIC);
+	if (!attr)
+		return -ENOMEM;
+
+	sysfs_bin_attr_init(attr);
+	attr->size = entry->pcilen;
+	attr->attr.name = "rom";
+	attr->attr.mode = S_IRUSR;
+	attr->read = setup_pci_rom_read;
+	retval = sysfs_create_bin_file(&entry->kobj, attr);
+	if (retval)
+		kfree(attr);
+	entry->rom_attr = attr;
+
+	return retval;
+}
+
+static int __init firmware_setup_pci_init(void)
+{
+	struct firmware_setup_pci_entry *entry;
+
+	list_for_each_entry(entry, &setup_pci_entries, list)
+		add_sysfs_fw_setup_pci_entry(entry);
+
+	return 0;
+}
+late_initcall(firmware_setup_pci_init);
+#endif
+
 int pcibios_enable_device(struct pci_dev *dev, int mask)
 {
 	int err;
-- 
1.8.4.5