Hi,
one of the big issue on the device tree will be to update it at
runtime from the bootloader
As example the same board have 2 version with only one difference the
Main Oscilator. So to support it easly without having 2 DT will want
to update the default DT with the proper Osc value from the
bootlaoder.
But as the bootloader will not be able to be udpate in the futur on
production board, we will have to never change to format of the DT
otherwise the bootloder will not be able to work on never kernel.
This is one of issue that nearly all of the PowerPC kernel dev meet
ofen with the couple u-boot + kernel + DT
To avoid this we may use a script to update the DT and then specify
how we get the information from the bootloader. So in this case, the
script we will manage the dependency of the current format version of
the DT.
Best Regards,
J.
HI,
Barebox flow basecly the linux kernel architecture
We use as linux a file system (devfs) to manage the device
such as NOR, NAND, SPI Flash, Memory, phy, SD etc...
As have a pseudo POSIX API to implement command application
such as ls, rm, mount, boot etc...
You can see barebox as "Small Linux like" bootlader
To extend it's fonctionality at runtime we use modules (same as in the
kernel).
We do not support yet ABI Application
With start from anywhere and they relocate our self at the right
addres in DDR after have init the soc and the board. Which include the
DDR.
After we use the initcall to init:
1: the core
Usualy clock and pio
2: console
uart
3: core device
using the device / driver model of the kernel
4: Filesystem
register filesytem
5: device
using the device / driver model of the kernel
as Frambuffer, and other
Then we mount a ramfs and devfs and execute a script /env/bin/init
all the boot sequence is manage by script based on Hush from busybox
we have a default scripting implementation to boot but you are free to
manage it how you want
we an boot from any media and from tftp, nfs with uImage or zImage or
Image
You can manage your boot loader configuration and boot from via a Menu
which can be manage in C or in shell
Over uart and soon over Framebuffer
Best Regards,
J.
I have tried tonight on my personal computer with the latest revision of Tianocore (rev1199).
Actually, there is two versions of the BeagleBoard UFEI upstream. There is the original version.
And the version we are working on which reuses most of the framework we have introduces for our development platform to avoid code duplication and flexibilty.
This is the version I have tested, to build it you must use './build-next.sh' instead of './build.sh'
Unfortunately, there are two patches you will need to apply. These patches are pending and wait to be approved by the maintainers of their respective packages.
One of the patch in the source tree and the other one is attached to this email.
Configuration:
--------------
- ARM GCC CodeSourcery arm-none-eabi-2010q3
- qEmu-linaro:
commit 2d601b5fb663bb2876b85bec255d73bba01e38e6
Author: Peter Maydell <peter.maydell(a)linaro.org>
Date: Wed Jun 15 15:08:48 2011 +0000
- Tianocore EDK2: revision 11999
Build process:
--------------
svn co https://edk2.svn.sourceforge.net/svnroot/edk2/trunk/edk2 edk2 --username guest -r 11999
cd edk2
svn co https://edk2-fatdriver2.svn.sourceforge.net/svnroot/edk2-fatdriver2/trunk/F… FatPkg --username guest
patch -p1 < ArmPlatformPkg/Documentation/patches/BaseTools-Pending-Patches.patch
patch -p1 < 0001-MdeModulePkg-DxeCore-Fix-the-loop-to-find-the-highes.patch
cd BeagleBoardPkg
I add to fix 2 easy warnings to make it build
./build-next.sh RELEASE
To test on qEmu:
----------------
./qemu-system-arm -M beagle -mtdblock ~/dev/edk2/Build/BeagleBoard/RELEASE_ARMGCC/FV/BeagleBoard_EFI_flashboot.fd -serial stdio -sd ~/dev/linaro-image-tools-0.4.8/beagle_sd.img
Log: starting an ATAG kernel from UEFI (after editing its filepath):
------------------------------------------------------------------------------
VNC server running on `127.0.0.1:5900'
The default boot selection will start in 8 seconds
[1] Linux from SD
[2] EBL
[3] Boot Manager
Start: 1
ERROR: Did not find Linux kernel.
[1] Linux from SD
[2] EBL
[3] Boot Manager
Start: 3
[1] Add Boot Device Entry
[2] Update Boot Device Entry
[3] Remove Boot Device Entry
[4] Return to main menu
Choice: 2
[1] Linux from SD
Update entry: 1
File path of the EFI Application or the kernel: zImage-atag
Has FDT support? [y/n] n
Arguments to pass to the binary:
Description for this new Entry: Linux from SD
[1] Add Boot Device Entry
[2] Update Boot Device Entry
[3] Remove Boot Device Entry
[4] Return to main menu
Choice: 4
[1] Linux from SD
[2] EBL
[3] Boot Manager
Start: 1
PEI 160 ms
DXE 622 ms
BDS 22922 ms
BDS 1418980260529 ms
Total Time = 1418980284234 ms
Uncompressing Linux... done, booting the kernel.
[ 0.000000] Linux version 2.6.38.7 (olivier@olivier-laptop) (gcc version 4.5.1 (Sourcery G++ Lite 2010.09-51) ) #1 Sun Jul 3 15:42:26 BST 2011
[ 0.000000] CPU: ARMv7 Processor [412fc083] revision 3 (ARMv7), cr=10c53c7f
[ 0.000000] CPU: VIPT nonaliasing data cache, VIPT nonaliasing instruction cache
[ 0.000000] Machine: OMAP3 Beagle Board
[ 0.000000] Reserving 33554432 bytes SDRAM for VRAM
[ 0.000000] Memory policy: ECC disabled, Data cache writeback
[ 0.000000] OMAP3430/3530 ES3.1 (iva sgx neon isp )
[ 0.000000] SRAM: Mapped pa 0x40200000 to va 0xfe400000 size: 0x10000
[ 0.000000] Clocking rate (Crystal/Core/MPU): 26.0/332/500 MHz
[ 0.000000] Reprogramming SDRC clock to 332000000 Hz
[ 0.000000] Built 1 zonelists in Zone order, mobility grouping on. Total pages: 24320
[ 0.000000] Kernel command line: root=/dev/mmcblk0p2 rootwait console=ttyO2,115200
[ 0.000000] PID hash table entries: 512 (order: -1, 2048 bytes)
[ 0.000000] Dentry cache hash table entries: 16384 (order: 4, 65536 bytes)
[ 0.000000] Inode-cache hash table entries: 8192 (order: 3, 32768 bytes)
[ 0.000000] Memory: 96MB = 96MB total
(...)
Log: Boot from a FDT kernel added to the Boot Menu:
-----------------------------------------------------------------
UEFI firmware built at 23:09:40 on Jul 6 2011
omap_badwidth_read32: 32-bit register 0x00000000
The default boot selection will start in 8 seconds
[1] Linux from SD
[2] EBL
[3] Boot Manager
Start: 3
[1] Add Boot Device Entry
[2] Update Boot Device Entry
[3] Remove Boot Device Entry
[4] Return to main menu
Choice: 1
[1] SemihostFs (0 MB)
[2] boot (51 MB)
[3] VenHw(4D00EF14-C4E0-426B-81B7-30A00A14AAD6)
Select the Boot Device: 2
File path of the EFI Application or the kernel: zImage-fdt
Has FDT support? [y/n] y
Arguments to pass to the binary:
Description for this new Entry: FDT Kernel from SD
[1] Add Boot Device Entry
[2] Update Boot Device Entry
[3] Remove Boot Device Entry
[4] Return to main menu
Choice: 4
[1] Linux from SD
[2] FDT Kernel from SD
[3] EBL
[4] Boot Manager
Start: 2
PEI 149 ms
DXE 609 ms
BDS 222 ms
Total Time = 981 ms
omap2_inth_read: Bad register 0x00000020
[ 0.000000] Linux version 2.6.39.1 (cosgor01@cam-vm-424) (gcc version 4.5.1 (Sourcery G++ Lite 2010.09-51) ) #2 SMP Thu Jun 30 18:55:24 BST 2011
[ 0.000000] CPU: ARMv7 Processor [412fc083] revision 3 (ARMv7), cr=10c53c7f
[ 0.000000] CPU: VIPT nonaliasing data cache, VIPT nonaliasing instruction cache
[ 0.000000] Machine: OMAP3 Beagle Board, model: TI OMAP3 BeagleBoard
[ 0.000000] Memory policy: ECC disabled, Data cache writeback
[ 0.000000] OMAP3430/3530 ES3.1 (iva sgx neon isp )
[ 0.000000] SRAM: Mapped pa 0x40200000 to va 0xfe400000 size: 0x10000
[ 0.000000] Clocking rate (Crystal/Core/MPU): 26.0/332/500 MHz
[ 0.000000] Reprogramming SDRC clock to 332000000 Hz
[ 0.000000] PERCPU: Embedded 7 pages/cpu @c0cc7000 s8160 r8192 d12320 u32768
[ 0.000000] Built 1 zonelists in Zone order, mobility grouping on. Total pages: 32512
[ 0.000000] Kernel command line: root=/dev/mmcblk0p2 rootwait console=ttyO2,115200n8
[ 0.000000] PID hash table entries: 512 (order: -1, 2048 bytes)
[ 0.000000] Dentry cache hash table entries: 16384 (order: 4, 65536 bytes)
[ 0.000000] Inode-cache hash table entries: 8192 (order: 3, 32768 bytes)
(...)
-- IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
To be frank, I spent most of my time since the last meeting to get Tianocore
working on qEmu.
Unfortunately, I have not been able to reproduce a consistent stable
firmware on qEmu.
It seems to work fine on my computer at home. I can boot Linux with ATAG
support or FDT support from UEFI on qEmu.
Booting UEFI from NOR Flash works. I also successfully 'hacked' a SD card
created with the linaro-media-create to replace u-boot by UEFI to get the
following boot stage ROM -> x-loader -> UEFI -> Linux.
I am using ARM GCC with the latest qEmu-linaro tree.
But on my machines at work with the same configuration, qEmu crashes during
the UEFI boot up with a 'Segmentation Fault'. It is the first time I work
with qEmu, I do not know if an 'Exception Fault' could be expected in case
of bug in the emulated binary or a problem in qEmu itself. I have not enough
element yet to report the issue to the qEmu team.
Anyway, these are the build instructions to build the Tianocore project
(UEFI Open Source implementation) and to test on qEmu:
http://edk2.svn.sourceforge.net/viewvc/edk2/trunk/edk2/BeagleBoardPkg/readme
.txt?revision=11997
Olivier
I forgot to mention but Tianocore and other UEFI Open Source projects have implemented shells for UEFI firmwares.
Some people are even working on porting Python to UEFI.
Feel free to ask questions or clarifications about this overview.
Mainly if you had thoughts about the way UEFI works which does not correspond to the big picture I gave.
Olivier
-----Original Message-----
From: boot-architecture-bounces(a)lists.linaro.org [mailto:boot-architecture-bounces@lists.linaro.org] On Behalf Of Olivier Martin
Sent: 06 July 2011 20:29
To: boot-architecture(a)lists.linaro.org
Subject: UEFI Overview
This email describes the way UEFI Tianocore implementation works. I limit
the scope to the UEFI drivers and the 'boot manager'. The implementation and
the platform specific bits are skipped.
Boot sequence:
--------------
1) After the platform modules (part of the PI (Platform Initialization)
phase) have finished the bringup initialization, the DXE (Driver Execution
Environment) Core is loaded.
2) The DXE core scans the different firmwares available on the platform and
calls the entrypoint of all the drivers that are expected to start (either
they request to be initialized during this sequence or they have their
dependencies solved).
Following the recommendation of the UEFI specification, the drivers should
do not touch the hardware controllers during this phase. They should only
declares their protocols (UEFI name for 'interfaces') to the DXE core.
3) Once all the drivers are dispatched, the DXE core passes its control to
the driver that exports the Boot Device Selection (BDS) interface.
This BDS driver should implement the 'Boot Manager' of the UEFI firmware
defined by the Chapter 3 of the UEFI spec.
This chapter defines non volatile variables that define among other things:
- The Boot Device entries ('Boot###' variables) (Name, location of the OS
loader, arguments) and their order ('BootOrder' variable)
- The devices responsible for the Console Input/Output
- The 'Timeout' before the BDS starts the default boot selection
- if defined, 'BootNext' defines the OS loader/EFI application to start the
next boot only (the variable is erased at the next boot)
4) Following the non variables values, the BDS will either automatically
starts the boot device (case 'BootNext' is defined or 'TimeOut' reached 0)
or displays the boot menu (with the Boot Device entries defined by
'Boot###' listed in the order defined by 'BootOrder'). Then the user chooses
in this menu which entry he/she wants to start.
5) With our implementation of the Boot Manager, we currently defines 3 types
of boot devices. The boot device can be either an EFI application (eg: an OS
loader) or a ATAG or FDT Linux kernel image. In case the BDS starts a Linux
kernel images, our BDS will automatically fills the required ATAG or loads
the Device Tree blob.
Note: One of our partner has been asking us to provide a reference EFI
Application to boot Linux. It is expecting some platforms will have their
own BDS and will use an EFI application to load the Linux kernel.
Actually, we are waiting the output of the Boot Architecture 'forum' to
provide such 'reference' EFI application.
UEFI Device Path
----------------
UEFI has the concept of 'Device Path'. Every hardware devices supported by
the UEFI firmware on the platform have a representation that defined the
hardware path to access to this device and their properties.
Some examples:
# Device path for a kernel on PCI SATA Hard Drive:
PciRoot(0)/Pci(0|0)/Pci(0|0)/Pci(5|0)/Pci(0|0)/Sata(0,0,0)/HD(2,MBR,0x000767
30,0x1F21BF,0x1F21BF)\boot\zImage
- PciRoot(0) define the first PCI Root complex.
- PCI(5|0) = PCI controller with Device = 5 and Function = 0
- HD(Partition,Type,Signature,Start, Size) = hard disk partition
- \boot\zImage : filepath on the partition
# Device path for a Linux kernel on the first partition of the SD card:
VenHw(B615F1F5-5088-43CD-809C-A16E52487D00)/HD(1,MBR,0x00000000,0x3F,0x19FC0
)/zImage
- VenHw(guid) is used when the device path node is not defined by the UEFI
specification (generally memory mapped controllers)
The Device Path allows to describe hardware devices or media even if they
have been initialized. When hardware drivers are started they can declare
the device path(s) of the hardware they support.
The advantage is if we do not need some specific hardwares during the UEFI
boot up, we could skip their initialization in UEFI.
This device path complexity can be hidden by the Boot Manager. We have
implemented our own interactive boot menu to help us to define the boot
devices.
_______________________________________________
boot-architecture mailing list
boot-architecture(a)lists.linaro.org
http://lists.linaro.org/mailman/listinfo/boot-architecture
-- IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
This email describes the way UEFI Tianocore implementation works. I limit
the scope to the UEFI drivers and the 'boot manager'. The implementation and
the platform specific bits are skipped.
Boot sequence:
--------------
1) After the platform modules (part of the PI (Platform Initialization)
phase) have finished the bringup initialization, the DXE (Driver Execution
Environment) Core is loaded.
2) The DXE core scans the different firmwares available on the platform and
calls the entrypoint of all the drivers that are expected to start (either
they request to be initialized during this sequence or they have their
dependencies solved).
Following the recommendation of the UEFI specification, the drivers should
do not touch the hardware controllers during this phase. They should only
declares their protocols (UEFI name for 'interfaces') to the DXE core.
3) Once all the drivers are dispatched, the DXE core passes its control to
the driver that exports the Boot Device Selection (BDS) interface.
This BDS driver should implement the 'Boot Manager' of the UEFI firmware
defined by the Chapter 3 of the UEFI spec.
This chapter defines non volatile variables that define among other things:
- The Boot Device entries ('Boot###' variables) (Name, location of the OS
loader, arguments) and their order ('BootOrder' variable)
- The devices responsible for the Console Input/Output
- The 'Timeout' before the BDS starts the default boot selection
- if defined, 'BootNext' defines the OS loader/EFI application to start the
next boot only (the variable is erased at the next boot)
4) Following the non variables values, the BDS will either automatically
starts the boot device (case 'BootNext' is defined or 'TimeOut' reached 0)
or displays the boot menu (with the Boot Device entries defined by
'Boot###' listed in the order defined by 'BootOrder'). Then the user chooses
in this menu which entry he/she wants to start.
5) With our implementation of the Boot Manager, we currently defines 3 types
of boot devices. The boot device can be either an EFI application (eg: an OS
loader) or a ATAG or FDT Linux kernel image. In case the BDS starts a Linux
kernel images, our BDS will automatically fills the required ATAG or loads
the Device Tree blob.
Note: One of our partner has been asking us to provide a reference EFI
Application to boot Linux. It is expecting some platforms will have their
own BDS and will use an EFI application to load the Linux kernel.
Actually, we are waiting the output of the Boot Architecture 'forum' to
provide such 'reference' EFI application.
UEFI Device Path
----------------
UEFI has the concept of 'Device Path'. Every hardware devices supported by
the UEFI firmware on the platform have a representation that defined the
hardware path to access to this device and their properties.
Some examples:
# Device path for a kernel on PCI SATA Hard Drive:
PciRoot(0)/Pci(0|0)/Pci(0|0)/Pci(5|0)/Pci(0|0)/Sata(0,0,0)/HD(2,MBR,0x000767
30,0x1F21BF,0x1F21BF)\boot\zImage
- PciRoot(0) define the first PCI Root complex.
- PCI(5|0) = PCI controller with Device = 5 and Function = 0
- HD(Partition,Type,Signature,Start, Size) = hard disk partition
- \boot\zImage : filepath on the partition
# Device path for a Linux kernel on the first partition of the SD card:
VenHw(B615F1F5-5088-43CD-809C-A16E52487D00)/HD(1,MBR,0x00000000,0x3F,0x19FC0
)/zImage
- VenHw(guid) is used when the device path node is not defined by the UEFI
specification (generally memory mapped controllers)
The Device Path allows to describe hardware devices or media even if they
have been initialized. When hardware drivers are started they can declare
the device path(s) of the hardware they support.
The advantage is if we do not need some specific hardwares during the UEFI
boot up, we could skip their initialization in UEFI.
This device path complexity can be hidden by the Boot Manager. We have
implemented our own interactive boot menu to help us to define the boot
devices.
Hey
QEMU is really handy to try out OMAP and Versatile Express boot stuff;
the Linaro QEMU version has patches (progressively being upstreamed)
for OMAP support and allows emulating vexpress, overo, beaglexm and
beagle boards which is a good collection already. It's based of a
modern QEMU commit and contains plenty of good fixes (all on their way
to mainline).
Get a tarball from:
https://launchpad.net/qemu-linaro
Source code at:
http://git.linaro.org/gitweb?p=qemu/qemu-linaro.git;a=summarygit://git.linaro.org/qemu/qemu-linaro.git
(This is like a regular qemu git tree or tarball; usual QEMU
documentation applies.)
Binaries are available in Ubuntu and backports are in the
linaro-maintainers/tools PPA.
Some QEMU HowTos are hosted on the Linaro wiki, but they are a bit
specific to Linaro images and tools to manipulate them. Basically you
can run a beagle xm SD image with:
qemu-system-arm -M beaglexm -sd your.img
(You might want to throw the serial line output directly on your
terminal or disable graphics or various other things.)
For vexpress, you can only boot by passing a kernel or ELF image to
QEMU:
qemu-system-arm -M vexpress-a9 -kernel u-boot.bin
Cheers,
--
Loïc Minier
Hey
(followup to today's call0
This is an attempt to write down an use case which expresses the
concept of "modules" that the boot architecture would cover. I don't
really like the name modules, if you have better names please propose!
So here's the writeup; this is completely fictional, and is just meant
to illustrate modules.
Here are the steps that could be taken to load a plain linux kernel:
a) your SoC starts up in a SoC-specific way
b) the boot architecture mandates how vmlinuz is loaded and started
c) control is passed to linux
d) profit!1!
in a very dumbed down version of our boot architecture, we could say
for b):
a file named "bootme" is loaded from the first FAT partition of the
SD card; its type is detected and if it's a linux kernel, it's
started with a device tree blob provided by the firmware to the
kernel
Good:
* easy to install or update a kernel by replacing the vmlinuz file
Bad:
* miss way to pass kernel cmdline
* miss way to pass (optional) initrd
* miss way to load/replace the DT
all of the above bad points could be alleviated by pushing this data in
the vmlinuz file, but it's not as practical as the modern linux systems
we know.
Another approach would be to put U-Boot or Barebox in the "bootme"
file, but then U-Boot or Barebox would have to decide how to load the
kernel and implement drivers for the SD card and this wouldn't be a
SoC-agnostic boot image anymore.
So the only possible approach I see in this case is to put a general
purpose ARM bootloader in the "bootme" file which will talk to firmware
to load additional files from the SD card (thanks to non-resident
support from the firmware); this would then go away once linux runs.
Now to "modules": instead of defining b) as just loading a single
"bootme" file, we'd define a config file which lists what to load (e.g.
something like grub.cfg/menu.lst) so that we could pass combinations
such as:
- vmlinuz + initrd + DT
- vmlinuz + DT
- Xen + DT + the OS modules it should load (linux and initrd)
it would also allow setting things like kernel cmdline or other OS
specific things.
Maybe I misunderstood Grant's idea, but it felt like he was mandating a
single payload from being loaded and just deferring to that payload. I
feel that if we go that route, we've specified too little to be able to
construct useful images and have deffered to an implementation specific
architecture (even if it could support multiple SoCs).
So my conclusion is that we should specify:
* either a single payload to start which would be something like a
generic ARM bootloader talking to the firmware to load the other bits
(e.g. via UEFI), but then also how it loads the next things
* or a config file which points at "modules" (by lack of a better name)
to load; these will get loaded by SoC specific code (likely an UEFI
implementation)
the main difference is whether we specify how a generic ARM bootloader
is loaded and then defer to it to decide how to load the OS, or whether
we specify OS-ish bits to load, and leave the loading to the
implementation.
Does that make any sense?
Cheers,
--
Loïc Minier
On Thu, Jun 23, 2011 at 5:25 AM, Olivier Martin <olivier.martin(a)arm.com> wrote:
> Thanks Nicolas for the link. You right the zImage has a signature. I read
> again the code I wrote and it is the signature for the non compressed image
> I have not found.
> My function was able to detect uImage and zImage but I had to assume if it
> is not one of these formats the image was a non-compressed image. Which it
> is ok as far as we have got only 3 formats, but that could cause trouble for
> any additional formats that would have some requirements from the boot
> loader.
>
>
> FYI, our UEFI implementation already supports ATAG (and partially DT). We
> can pass a DT to the kernel but we have not implemented yet the update of
> the tree in the firmware; for example to pass new arguments to the kernel.
> Again same argument as the binary format, if in the future a new format
> incompatible with the DT format is introduced to answer some limitations of
> the Flat Device Tree at this time; should the ARM boot 'standardization' be
> rewriting again and break the legacy mechanism ?
>
>
> I had a look again to this page:
> https://wiki.linaro.org/OfficeofCTO/BootArchitecture
> It seems the current concern is more about the firmware itself than the
> interface between the OS loader and the kernel. I guess this Blue Print has
> been created to solve the issues of the boot fragmentation in the ARM world.
> I suppose the idea is to introduce some requirements for booting ARM
> platforms.
> But should we only limit our view to the requirements of booting Linux 3.0
> (zImage? and Device Tree). Could we also think about defining some
> 'standards' in the interface between the OS Loader and the Linux kernel to
> leave some flexibility for any future technology that involves the boot
> firmware to initialize the platform for the Linux kernel.
I think in general, yes we should constrain our view to 3.0+. That
doesn't preclude firmware from supporting what it already supports.
Indeed it would be crazy to ask firmware projects to remove things
that it currently supports. Rather, I think that the process of
creating a standard implementable by any firmware should be /focused/
on new kernels since we cannot change the stuff that is already
deployed anyway.
> As I suggested in my first email, defining a signature (binary format:zImage
> + machine type:DT) could be a way to define our current requirements and
> leave some place for any future requirements
I've added signature/CRC/identification as a topic to the boot
architecture page.
g.
On Tue, 21 Jun 2011, Olivier Martin wrote:
> Thanks Grant for adding me to the list,
> To introduce myself, I am part of the UEFI team at ARM Ltd. I am also the
> maintainer of the ARM Packages in the Open Source implementation of UEFI
> (http://www.tianocore.org/ ).
>
> Maybe an issue this "ARM Boot Architecture" topic should solve is the
> interface between the boot loader and the Linux kernel binary.
>
> Recently, I was trying to extend our ARM UEFI boot loader to detect the type
> of the Linux kernel binary:
> - zImage, non-compressed Image, uImage format
> - ATAG or FDT support
> ... to start any kernel by passing the correct parameters and jump to the
> right offset.
>
> Unfortunately, it was not as easy as it should be. In fact it does not look
> to be possible right now. Defining a signature (eg: kernel binary format +
> machine type) in the kernel binary would make the task easier. Some of the
> formats have a signature (eg: u-boot) but not all the formats.
The zImage format has a signature too. I documented it here a _long_
time ago:
http://lists.arm.linux.org.uk/lurker/message/20020226.165754.4c7e6539.en.ht…
Of course some things have changed since then. For a while now, the
zImage is fully position independent. I can be loaded anywhere in
memory and branched directly to it as it takes care of relocating itself
when needed. Therefore the load address at offset 0x28 is now always
zero, except when the zImage is linked to be executed directly from
flash memory (but I doubt that case is relevant to UEFI).
For that matter, the uImage format is broken. It insists on absolute
addresses in its header. This is therefore not a good format if we want
to support multiple SoCs with the same kernel binary, as some of them
have their physical RAM at different addresses. You may want to support
it if you want, but it is likely not to be suitable anymore in the
future. Hence having the ability to boot a zImage directly is likely to
be the most flexible option.
> Actually, if we try to boot a Linux kernel with the wrong settings then the
> kernel would crash or raise an error. And it is not be possible to recover
> from this state without restarting the platform.
What kind of parameters do you have in mind?
> Making these information available to the boot loader (u-boot, uefi, etc)
> would also make easier the migration from the legacy ATAG to the Flat Device
> Tree (FDT) support.
How would it make that transition easier? The link between those issues
is not obvious to me.
With regard to ATAGs, I don't think UEFI should bother with them. If we
really want to move to device tree, and it seems we do, then ATAGs
should be left behind. We do have a compatibility layer going into
zImage to convert ATAGs into DT properties at run time to allow booting
a DT kernel on legacy bootloaders. Adding ATAg support to UEFI might
impair the move to DT in the kernel by making it a lower priority.
Nicolas