This patch add a field of 64-bit physical pointer to NULL terminated
single linked list of struct setup_data to real-mode kernel
header. This is used as a more extensible boot parameters passing
mechanism.

This patch has been tested against 2.6.23-rc6-mm1 kernel on x86_64. It
is based on the proposal of Peter Anvin.


Known Issues:

1. Where is safe to place the linked list of setup_data?
Because the length of the linked list of setup_data is variable, it
can not be copied into BSS segment of kernel as that of "zero
page". We must find a safe place for it, where it will not be
overwritten by kernel during booting up. The i386 kernel will
overwrite some pages after _end. The x86_64 kernel will overwrite some
pages from 0x1000 on.


ChangeLog:

-- v2 --

- Increase the boot protocol version number.
- Check version number before parsing setup_data.


Signed-off-by: Huang Ying <ying.huang@intel.com>

---

 arch/i386/Kconfig            |    3 ---
 arch/i386/boot/header.S      |    8 +++++++-
 arch/i386/kernel/setup.c     |   22 ++++++++++++++++++++++
 arch/x86_64/kernel/setup.c   |   21 +++++++++++++++++++++
 include/asm-i386/bootparam.h |   15 +++++++++++++++
 include/asm-i386/io.h        |    7 +++++++
 6 files changed, 72 insertions(+), 4 deletions(-)

Index: linux-2.6.23-rc6/include/asm-i386/bootparam.h
===================================================================
--- linux-2.6.23-rc6.orig/include/asm-i386/bootparam.h	2007-09-19 10:22:02.000000000 +0800
+++ linux-2.6.23-rc6/include/asm-i386/bootparam.h	2007-09-19 16:41:57.000000000 +0800
@@ -9,6 +9,17 @@
 #include <asm/ist.h>
 #include <video/edid.h>
 
+/* setup data types */
+#define SETUP_NONE			0
+
+/* extensible setup data list node */
+struct setup_data {
+	u64 next;
+	u32 type;
+	u32 len;
+	u8 data[0];
+} __attribute__((packed));
+
 struct setup_header {
 	u8	setup_sects;
 	u16	root_flags;
@@ -41,6 +52,10 @@
 	u32	initrd_addr_max;
 	u32	kernel_alignment;
 ...
[ message continues ]

The latter is definitely not safe, since the space below 640K is the
documented place to put the command line (and presumably where the
bootloader would put other auxilliary chunks.)

I'll try to do a full review of this later today.  Haven't had time yet
to look at this anything than but piecemeal.

	-hpa
-

[ message continues ]

Hi, Peter,


Do you think this patch and the 32-bit boot protocol patch are ready to
merge for -mm? If not, I can revise them.

Best Regards,
Huang Ying
-

[ message continues ]

Sorry, haven't had a chance to look at it in proper detail yet, mostly
due to debugging, but one thing I'd like to see is both the boot_params
structure as well as all the chained information pointers exported into
sysfs.  The experience with the x86 setup code has shown that it would
help immensely with debugging, not to mention being available to tools
like kexec.

	-hpa
-

[ message continues ]

Do you actually need a linked list of data?  This is similar to the
changes to bzImage to support booting bzImage a paravirt environment,
but I just proposed a pointer to a single info structure, along with a
field to identify the boot environment (ie, native/lguest/xen etc).  It
would be easy to extend this to handle EFI as just another boot
environment, and you could hang a list of structures off the pointer.


    J
-

[ message continues ]

*He* doesn't, but *we* do.  We have already run into at least one case
where the existing structure is insufficient (EDD overhaul), and so we
need to do something extensible.

	-hpa
-

[ message continues ]