Hi Bin,
On 26 November 2014 at 18:44, Bin Meng bmeng.cn@gmail.com wrote:
Hi Simon,
On Wed, Nov 26, 2014 at 5:51 AM, Simon Glass sjg@chromium.org wrote:
Hi Bin (and others interested in U-Boot on x86),
I've applied the remaining x86 patches to u-boot-x86. It runs on chromebook_link (Pixel) with support for most hardware relevant to a boot loader: SDRAM, SPI, PCI, USB (and USB Ethernet), SATA (internal 32GB SSD), SD card, LCD, UART, keyboard, EC.
Bin this should be a good base for you to send patches for your Atom platform and I have no major work pending now so should not get in your way.
This is great! Thanks for applying your patch series into the mainline so quickly. I will start working on my patches soon.
Instructions on how to build and run are here:
http://www.denx.de/wiki/U-Boot/X86
For this platform 4 binary blobs are needed. This is an unavoidable feature of the platform at present. The blobs cover flash descriptor, SDRAM init, video init and Management Engine. Instructions on how to get these are on the same page.
Here is a list of some missing features:
- README.x86 in the source (mostly the content from the Wiki page
would be a good start)
- MTRR support (for performance)
- Audio
- Chrome OS verified boot (only a rough rebase has been done, I'm not
sure how to track mainline anyway)
- SMI and ACPI support, to provide platform info and facilities to Linux
One question related to ACPI, do we need support pre-ACPI protocols for handling over resource allocations and interrupt vector assignment information to the OS? I mean the PIRQ table and MP table. These specs are really old nowadays, and even commercial BIOS does not always get those tables correct, but as far as I can see, ACPI tables are more reliable. I think this is largely because they validate ACPI support with Windows and Linux which always use ACPI.
I vote no. ACPI should be enough.
Regards, Simon