On Wed, 15 Dec 2021 12:44:16 +0000 Sudeep Holla <sudeep.ho...@arm.com> wrote:
Hi Sudeep, > On Tue, Dec 14, 2021 at 05:55:39PM +0000, Andre Przywara wrote: > > The Juno Arm development board is an open, vendor-neutral, Armv8-A > > development platform. > > Add documentation that briefly outlines the hardware, and describes > > building and installation of U-Boot. > > > > Signed-off-by: Andre Przywara <andre.przyw...@arm.com> > > --- > > doc/board/armltd/index.rst | 1 + > > doc/board/armltd/juno.rst | 117 +++++++++++++++++++++++++++++++++++++ > > 2 files changed, 118 insertions(+) > > create mode 100644 doc/board/armltd/juno.rst > > > > diff --git a/doc/board/armltd/index.rst b/doc/board/armltd/index.rst > > index caa6fd2bb0..68d938c647 100644 > > --- a/doc/board/armltd/index.rst > > +++ b/doc/board/armltd/index.rst > > @@ -8,3 +8,4 @@ ARM Ltd. boards and emulated systems > > :maxdepth: 2 > > > > fvp64 > > + juno > > diff --git a/doc/board/armltd/juno.rst b/doc/board/armltd/juno.rst > > new file mode 100644 > > index 0000000000..f37bc2c78e > > --- /dev/null > > +++ b/doc/board/armltd/juno.rst > > @@ -0,0 +1,117 @@ > > +.. SPDX-License-Identifier: GPL-2.0 > > +.. Copyright (C) 2021 Arm Ltd. > > + > > +Arm Juno board > > +============== > > + > > +The `Juno development board`_ is an open, vendor-neutral, Armv8-A > > +development platform, made by Arm Ltd. It is based on the former Versatile > > +Express series. > > +There are three revisions of the board: > > + > > +* Juno r0, with two Cortex-A57 and four Cortex-A53 cores, without PCIe. > > +* Juno r1, with two Cortex-A57 and four Cortex-A53 cores, in later silicon > > + revisions, and with PCIe slots, Gigabit Ethernet and two SATA ports. > > +* Juno r2, with two Cortex-A72 and four Cortex-A53 cores, otherwise the > > + same as r1. > > + > > +Among other things, the motherboard contains a management controller > > (MCP), > > IIRC the MCP is new and inside the SoC. You may refer [1], [2] and use the > terminologies from there to be consistent with the documentation. IIRC this > one is referred as MCC. So I would prefer s/MCP/MCC or MB throughout this > document. True, I managed to mix that up, thanks for pointing this out! > > +an FPGA providing I/O interfaces (IOFPGA) and 64MB of NOR flash. The > > provided > > +platform devices resemble the VExpress peripherals. > > +The actual SoC also contains a Cortex-M3 based System Control Processor > > (SCP). > > + > > +U-Boot build > > +------------ > > +There is only one defconfig and one binary build that covers all three > > board > > +revisions, so to generate the needed ``u-boot.bin``: > > + > > +.. code-block:: bash > > + > > + $ make vexpress_aemv8a_juno_defconfig > > + $ make > > + > > +The automatic distro boot sequence looks for UEFI boot applications and > > +``boot.scr`` scripts on various boot media, starting with USB, then on > > disks > > +connected to the two SATA ports, PXE, DHCP and eventually on the NOR flash. > > + > > +U-Boot installation > > +------------------- > > +This assumes there is some firmware on the SD card or NOR flash (see below > > +for more details). The U-Boot binary is included in the Trusted Firmware > > +FIP image, so after building U-Boot, this needs to be repackaged or > > recompiled. > > + > > +The NOR flash will be updated by the MCP, based on the content of a > > micro-SD > > +card, which will be exported as a USB mass storage device via the rear > > USB-B > > +socket. So to access that SD card, connect a USB-A->USB-B cable between > > some > > +host computer and the board, and mount the FAT partition on the UMS device. > > +If there is no device, check the upper serial port for a prompt, and > > +explicitly enable the USB interface:: > > + > > + Cmd> usb_on > > + Enabling debug USB... > > + > > Not sure if you need these details(above one paragraph) here if we can direct > to one of the pages I have pointed out or specifically [3]. I guess you can > add other topics from there and links to those subsections if you need more > details. I am fine either way. Yeah, I will add a link to the Juno TRM, and refer to that. But at least on the Linux side I have seen pushback against deep links to manufacturer websites in documentation, since those URLs tend to 404 sooner or later. So I am tempted to keep at least that usb_on command in, since I needed to use that at times, and it's a quick solution to a common problem. Thanks for your comments! Andre
