8.18. x86_64

8.18.1. amd64

This BSP offers two variants: amd64 and amd64efi. The BSP can run on UEFI-capable systems by using the FreeBSD bootloader in the case of amd64 or a multiboot2 compliant bootloader in the case of amd64efi. The main difference of amd64efi is that it utilizes the UEFI Boot Services for its functionality.

Currently the console driver, clock driver, and context switching are functional. ACPI functionality is supported through ACPICA and used for SMP (only supported in amd64).

8.18.1.1. Build Configuration Options

There are no BSP configuration options available at build time.

8.18.1.2. Testing with QEMU

To test with QEMU, we need to:

  • Build / install QEMU (most distributions should have it available on the package manager).

  • Build UEFI firmware that QEMU can use to simulate an x86-64 system capable of booting a UEFI-aware kernel, through the --bios flag.

8.18.1.2.1. Building TianoCore’s UEFI firmware, OVMF

Complete detailed instructions are available at TianoCore’s Github’s wiki.

Quick instructions (which may fall out of date) are:

$ git clone git://github.com/tianocore/edk2.git
$ cd edk2
$ make -C BaseTools
$ . edksetup.sh

Then edit Conf/target.txt to set:

ACTIVE_PLATFORM       = OvmfPkg/OvmfPkgX64.dsc
TARGET                = DEBUG
TARGET_ARCH           = X64
# You can use GCC46 as well, if you'd prefer
TOOL_CHAIN_TAG        = GCC5

Then run build in the edk2 directory - the output should list the location of the OVMF.fd file, which can be used with QEMU to boot into a UEFI shell.

You can find the OVMF.fd file like this as well in the edk2 directory:

$ find . -name "*.fd"
./Build/OvmfX64/DEBUG_GCC5/FV/MEMFD.fd
./Build/OvmfX64/DEBUG_GCC5/FV/OVMF.fd # the file we're looking for
./Build/OvmfX64/DEBUG_GCC5/FV/OVMF_CODE.fd
./Build/OvmfX64/DEBUG_GCC5/FV/OVMF_VARS.fd

8.18.1.2.2. Booting RTEMS in QEMU

The amd64 variant supports being booted through the FreeBSD bootloader (Booting via the FreeBSD bootloader), meanwhile the amd64efi variant supports being booted by a multiboot2 compliant bootloader, such as GRUB (Booting via GRUB).

8.18.1.2.2.1. Booting via the FreeBSD bootloader

Note

The following section describes how to boot RTEMS using a FreeBSD VM, for a more self contained alternative check out Creating a FreeBSD Boot Image

The RTEMS executable produced (an ELF file) needs to be placed in the FreeBSD’s /boot/kernel/kernel’s place.

To do that, we first need a hard-disk image with FreeBSD installed on it. Download FreeBSD’s installer “memstick” image for amd64 and then run the following commands, replacing paths as appropriate.

$ qemu-img create freebsd.img 8G
$ OVMF_LOCATION=/path/to/ovmf/OVMF.fd
$ FREEBSD_MEMSTICK=/path/to/FreeBSD-11.2-amd64-memstick.img
$ qemu-system-x86_64 -m 1024 -serial stdio --bios $OVMF_LOCATION \
    -drive format=raw,file=freebsd.img \
    -drive format=raw,file=$FREEBSD_MEMSTICK

The first time you do this, continue through and install FreeBSD. FreeBSD’s installation guide may prove useful if required.

Once installed, build your RTEMS executable (an ELF file), for eg. hello.exe. We need to transfer this executable into freebsd.img’s filesystem, at either /boot/kernel/kernel or /boot/kernel.old/kernel (or elsewhere, if you don’t mind user FreeBSD’s loader’s prompt to boot your custom kernel).

If your host system supports mounting UFS filesystems as read-write (eg. FreeBSD), go ahead and:

  1. Mount freebsd.img as read-write

  2. Within the filesystem, back the existing FreeBSD kernel up (i.e. effectively cp -r /boot/kernel /boot/kernel.old).

  3. Place your RTEMS executable at /boot/kernel/kernel

If your host doesn’t support mounting UFS filesystems (eg. most Linux kernels), do something to the effect of the following.

On the host

# Upload hello.exe anywhere accessible within the host
$ curl --upload-file hello.exe https://transfer.sh/rtems

Then on the guest (FreeBSD), login with root and

# Back the FreeBSD kernel up
$ cp -r /boot/kernel/ /boot/kernel.old
# Bring networking online if it isn't already
$ dhclient em0
# You may need to add the --no-verify-peer depending on your server
$ fetch https://host.com/path/to/rtems/hello.exe
# Replace default kernel
$ cp hello.exe /boot/kernel/kernel
$ reboot

After rebooting, the RTEMS kernel should run after the UEFI firmware and FreeBSD’s bootloader. The -serial stdio QEMU flag will let the RTEMS console send its output to the host’s stdio stream.

8.18.1.2.2.2. Booting via GRUB

All that is required is for GRUB to be configured to boot the executable through multiboot2. This section simply shows a possible way of achieving this.

We are going to create a single EFI System Partition (ESP) containing the GRUB binary and our executable. First, create the proper file structure for the ESP:

$ mkdir -p RTEMS-GRUB/EFI/BOOT

We are going to need a valid grub.cfg file. The following example will configure GRUB to search for a file named “rtems” in the root of the partition and boot it with multiboot2 instantly:

set timeout=0
set default=0

search --file --set=root /rtems

menuentry 'RTEMS' {
    multiboot2 /rtems
    boot
}

With this in place we can generate a GRUB binary containing the grub.cfg and required modules on EFI/BOOT/BOOTX64.EFI (the default boot loader file for UEFI systems):

$ grub-mkstandalone --format=x86_64-efi --fonts="" --locales="" --themes="" \
    --install-modules="normal search fat multiboot2"                        \
    boot/grub/grub.cfg=grub.cfg -o RTEMS-GRUB/EFI/BOOT/BOOTX64.EFI

And then copy the executable you desire to boot to /rtems (as specified by our grub.cfg) in our ESP:

$ cp ${rtems-executable} RTEMS-GRUB/rtems

With all this in place we will use the makefs tool (which is contained in the x86_64 build set in the RTEMS Source Builder) to create a FAT32 image out of the file structure:

$ makefs -t msdos -s 50m RTEMS-GRUB.img RTEMS-GRUB

And now all that is left is booting the image with QEMU:

$ qemu-system-x86_64 -m 512 -serial stdio --bios $OVMF_LOCATION \
    -drive format=raw,file=RTEMS-GRUB.img

Note

The guide in this section uses the makefs tool to create the final FAT32 image out of the ESP file structure. Using makefs is not required as long as you can create a FAT32 image containing the same exact file structure.

8.18.1.3. Using the RTEMS tester

Both amd64 and amd64efi contain tester configuration files for using the RTEMS tester tool, but they require user configuration.

The amd64 requires the path to the FreeBSD boot image which will be used by the tester tool (Creating a FreeBSD Boot Image). Meanwhile both amd64 and amd64efi require the path to OVMF (Building TianoCore’s UEFI firmware, OVMF).

An example of the user configuration file:

[amd64_qemu]
amd64_ovmf_path = {OVMF_PATH}
amd64_freebsd_boot_image_path = {FREEBSD_BOOT_IMAGE_PATH}

[amd64efi_grub_qemu]
amd64_ovmf_path = {OVMF_PATH}

8.18.1.3.1. Creating a FreeBSD Boot Image

Note

You can instead choose to download a working boot image here. For directly using the boot image with QEMU go to Booting in QEMU Using the Boot Image:

To acquire or build the FreeBSD bootloader a FreeBSD machine or VM is required. You can either copy the files already present under /boot or build them yourself.

To build the bootloader yourself, assuming the FreeBSD source tree is under /usr/src, head over to /usr/src/stand and run the following commands:

$ make
$ make install DESTDIR={bootloader-path}

Note

The directories usr/share/man/man3, usr/share/man/man5, and usr/share/man/man8 must be created under {bootloader-path} before running make install

Next create the EFI disk image with the FreeBSD bootloader under EFI/BOOT/BOOTX64.EFI (the default boot loader file for UEFI systems):

$ mkdir -p efi-image/EFI/BOOT/
$ cp {bootloader-path}/loader.efi efi-image/EFI/BOOT/BOOTX64.EFI
$ makefs -t msdos -s 1m EFI.img efi-image

And then the FreeBSD Root FS disk image:

$ mkdir -p rootfs-image/boot/
$ cp -r {bootloader-path}/defaults rootfs-image/boot/
$ cp -r {bootloader-path}/lua rootfs-image/boot/

The following configuration file will instruct the FreeBSD loader to instantly load the file /rtems contained in the second disk. It should be created under rootfs-image/boot/loader.conf

beastie_disable="YES"
kernel="/rtems"
currdev="disk1"
autoboot_delay="0"

And then we can convert rootfs-image to an UFS disk image and use the mkimg tool to create a singular image with both partitions:

makefs -t ffs -o version=2 ROOTFS.img rootfs-image
mkimg -s gpt -p efi:=EFI.img -p freebsd-ufs:=ROOTFS.img -o FreeBSDBoot.img
8.18.1.3.1.1. Booting in QEMU Using the Boot Image:

You can use the boot image to run any rtems executable with QEMU as such:

$ mkdir rtems-image
$ cp {rtems-executable} rtems-image/rtems
$ makefs -t ffs -o version=2 rtems.img rtems-image
$ qemu-system-x86_64 -m 1024 -serial stdio --bios {OVMF_LOCATION} \
    -drive format=raw,file=FreeBSDBoot.img \
    -drive format=raw,file=rtems.img

8.18.1.4. Paging

During the BSP’s initialization, the paging tables are setup to identity-map the first 512GiB, i.e. virtual addresses are the same as physical addresses for the first 512GiB.

The page structures are set up statically with 1GiB super-pages.

Note

Page-faults are not handled.

Warning

RAM size is not detected dynamically and defaults to 1GiB, if the configuration-time RamSize parameter is not used.

8.18.1.5. Interrupt Setup

Interrupt vectors 0 through 32 (i.e. 33 interrupt vectors in total) are setup as “RTEMS interrupts”, which can be hooked through rtems_interrupt_handler_install.

The Interrupt Descriptor Table supports a total of 256 possible vectors (0 through 255), which leaves a lot of room for “raw interrupts”, which can be hooked through _CPU_ISR_install_raw_handler.

Since the APIC needs to be used for the clock driver, the PIC is remapped (IRQ0 of the PIC is redirected to vector 32, and so on), and then all interrupts are masked to disable the PIC. In this state, the PIC may _still_ produce spurious interrupts (IRQ7 and IRQ15, redirected to vector 39 and vector 47 respectively).

The clock driver triggers the initialization of the APIC and then the APIC timer.

The I/O APIC is not supported at the moment.

Note

IRQ32 is reserved by default for the APIC timer (see following section).

IRQ33 is reserved by default for interprocessor interrupts if SMP is enabled.

IRQ255 is reserved by default for the APIC’s spurious vector.

Warning

Besides the first 33 vectors (0 through 32), and vector 255 (the APIC spurious vector), no other handlers are attached by default.

8.18.1.6. Clock Driver

8.18.1.6.1. amd64

The clock driver currently uses the APIC timer. Since the APIC timer runs at the CPU bus frequency, which can’t be detected easily, the PIT is used to calibrate the APIC timer, and then the APIC timer is enabled in periodic mode, with the initial counter setup such that interrupts fire at the same frequency as the clock tick frequency, as requested by CONFIGURE_MICROSECONDS_PER_TICK.

8.18.1.6.2. amd64efi

The clock driver uses the SetTimer UEFI boot service.

8.18.1.7. Console Driver

8.18.1.7.1. amd64

The console driver defaults to using the COM1 UART port (at I/O port 0x3F8), using the NS16550 polled driver.

8.18.1.7.2. amd64efi

The console driver uses the UEFI Simple Text Output Protocol