Fastboot

Jonas Vautherin jonas.vautherin at gmail.com
Thu Mar 11 01:56:09 CET 2021


Thanks a lot, Sean!

>From you answer, it seems like what I am trying to do (i.e. applying that
patch to am335x.dts to set dr_mode to "peripheral") makes sense, though for
some reason it does not get applied.

I have been careful in checking that the dts file disappears after I
cleansstate u-boot, and that it gets patched in the do_patch step. I have
also checked that u-boot.img disappears from the deploy folder when I
cleansstate, and gets rebuilt after the patch is applied. And finally, I
have made sure that the cleansstate on my image removed it from the deploy
folder.

Moreover, I have `find build | grep am335x.dts` to check if I was patching
the right file (which should be ensured by the fact that I patch it in the
bbappend of the u-boot recipe).

My fear is that for some reason, it does not use am335x.dts to build the
device tree but something else entirely, which would mean that I am
patching something that is not used.

>From u-boot.img, or ideally even from my final image I flash on the sdcard,
is there a way for me to see what is used to define the device tree?

Best,

On Thu, 11 Mar 2021, 00:44 Sean Anderson, <seanga2 at gmail.com> wrote:

> On 3/10/21 12:43 PM, Jonas Vautherin wrote:
> > Hello!
> >
> > I've made some more tests, but I am stuck and I don't really know what
> > I am missing. Would love to get some insights, if somebody has an idea
> > :-). So my root problem is that I would like to run my pocketbeagle in
> > fastboot mode so that I can flash an image from my laptop over USB
> > (it's a learning experience, I'm new to u-boot).
> >
> > When I go into the u-boot prompt and run `fastboot usb 0` (after I set
> > `log level 7`), I get:
> >
> > ```
> > => fastboot usb 0
> > uclass_find_device_by_seq() 0 0
> > uclass_find_device_by_seq()    - not found
> > uclass_find_device_by_seq() 1 0
> > uclass_find_device_by_seq()    - not found
> > cmd_call() Command failed, result=1
> > ```
> >
> > Similar for `fastboot usb 1` (I don't know which one is the microB on
> > the pocketbeagle), except the numbers are `0 1` and `1 1` in the log
> > output.
>
> The numbers come from /aliases in the device tree. You should read the
> datasheet/schematic to figure out which is which :)
>
> >
> > If I `dm tree`, for some reason I see only usb1 at 47401800 (according to
> > the `.dts`, usb0 is @47401000).
> >
> > ```
> >   misc          0  [ + ]   ti-musb-wrapper           `-- usb at 47400000
> >   usb           0  [   ]   ti-musb-host                  `-- usb at 47401800
> > ```
> >
> > Looking at the defconfig (I use Yocto, so `bitbake -c menuconfig
> > u-boot), I see that "Default Device Tree for DT control" is
> > (am335x-evm). Opening
> >
> ./build/tmp/work/pocketbeagle-poky-linux-gnueabi/u-boot/1_2020.07-r0/git/arch/arm/dts/am335x-evm.dts
> > (which I believe is the one corresponding to `am335x-evm` in my
> > defconfig), I see the following for usb:
> >
> > ```
> > &usb {
> > status = "okay";
> > };
> >
> > &usb_ctrl_mod {
> > status = "okay";
> > };
> >
> > &usb0_phy {
> > status = "okay";
> > };
> >
> > &usb1_phy {
> > status = "okay";
> > };
> >
> > &usb0 {
> > status = "okay";
> > };
> >
> > &usb1 {
> > status = "okay";
> > dr_mode = "host";
> > };
> > ```
> >
> > And showing usb0 and usb1 with `fdt print` (after I get the value of
> > $fdt_blob with `bdinfo` and use it with `fdt addr $fdt_blob`), I get:
> >
> > ```
> > => fdt print /ocp/usb at 47400000/usb at 47401000
> > usb at 47401000 {
> >          compatible = "ti,musb-am33xx";
> >          status = "okay";
> >          reg = <0x47401400 0x00000400 0x47401000 0x00000200>;
> >          reg-names = "mc", "control";
> >          interrupts = <0x00000012>;
> >          interrupt-names = "mc";
> >          dr_mode = "otg";
> >          mentor,multipoint = <0x00000001>;
> >          mentor,num-eps = <0x00000010>;
> >          mentor,ram-bits = <0x0000000c>;
> >          mentor,power = <0x000001f4>;
> >          phys = <0x00000037>;
> >          dmas = * 0x9df20aa8 [0x00000168];
> >          dma-names = "rx1", "rx2", "rx3", "rx4", "rx5", "rx6", "rx7",
> "rx8", "rx
> > 9", "rx10", "rx11", "rx12", "rx13", "rx14", "rx15", "tx1", "tx2", "tx3",
> "tx4", "tx5", "tx6", "tx7", "tx8", "tx9", "tx10", "tx11", "tx12", "tx13",
> "tx14", "tx15";
> > };
> > ```
> >
> > and
> >
> > ```
> > => fdt print /ocp/usb at 47400000/usb at 47401800
> > usb at 47401800 {
> >          compatible = "ti,musb-am33xx";
> >          status = "okay";
> >          reg = <0x47401c00 0x00000400 0x47401800 0x00000200>;
> >          reg-names = "mc", "control";
> >          interrupts = <0x00000013>;
> >          interrupt-names = "mc";
> >          dr_mode = "host";
> >          mentor,multipoint = <0x00000001>;
> >          mentor,num-eps = <0x00000010>;
> >          mentor,ram-bits = <0x0000000c>;
> >          mentor,power = <0x000001f4>;
> >          phys = <0x00000039>;
> >          dmas = * 0x9df20e50 [0x00000168];
> >          dma-names = "rx1", "rx2", "rx3", "rx4", "rx5", "rx6", "rx7",
> "rx8", "rx9", "rx10", "rx11", "rx12", "rx13", "rx14", "rx15", "tx1", "tx2",
> "tx3", "tx4", "tx5", "tx6", "tx7", "tx8", "tx9", "tx10", "tx11", "tx12",
> "tx13", "tx14", "tx
> > 15";
> > };
> > ```
> >
> > My understanding is therefore that `usb0` is in "otg" mode, and `usb1`
> > is in "host" mode. But that does not really tell me why `fastboot usb
> > 0` and `fastboot usb 1` failed.
>
> Yes. The U-Boot driver only supports host and peripheral modes, so no
> otg device is created. And fastboot cannot work with a USB host.
>
> >
> > I have also tried to edit `am335x-evm.dts` and set both to
> > "peripheral" (I do that in a Yocto patch step and make sure I `bitbake
> > -c cleansstate u-boot` and `bitbake -c clean <my-image>` before I
> > rebuild):
> >
> > ```
> > &usb0 {
> > status = "okay";
> > dr_mode = "peripheral";
> > };
> >
> > &usb1 {
> > status = "okay";
> > dr_mode = "peripheral";
> > };
> > ```
> >
> > But that does not change anything in the `fdt print` output: usb0 is
> > still "otg" and usb1 is still "host". Which makes me guess I may be
> > modifying the wrong `am335x-evm.dts`, but I really don't know how I
> > could check that.
>
> Does cleaning wipe your changes? There should be some yocto way to apply
> patches; perhaps someone else can comment on that.
>
> --Sean
>
> >
> > Again, any hints are appreciated, I'm happy to DuckDuckGo some new
> keywords or ideas as my goal is to learn about u-boot, but right now I'm
> just stuck :-).
> >
> > Best Regards,
> > Jonas
> >
> > On Sun, Mar 7, 2021 at 7:31 PM Jonas Vautherin <
> jonas.vautherin at gmail.com <mailto:jonas.vautherin at gmail.com>> wrote:
> >
> >     Yes that was it! I had used `log help`, but misunderstood the
> output. My bad... Below is the debug output I receive, where I don't really
> get what's happening except for those two lines:
> >
> >     ```
> >     uclass_find_device_by_seq() - 1 -1 'usb at 47401800'
> >     uclass_find_device_by_seq() - not found
> >     ```
> >
> >     But I don't really understand what it means. I spent a lot of time
> reading the available options in menuconfig, and at that point I cannot say
> what I am doing wrong. Note that I am using the `am335x_evm_defconfig`
> defconfig, which I changed to get more logging outputs.
> >
> >     Here is the output of `usb start`, let me know if I am missing
> something obvious from there:
> >
> >     ```
> >     => usb start
> >     starting USB...
> >     Bus usb at 47401800: __of_translate_address() OF: ** translation for
> device usb at 47401800 **
> >     __of_translate_address() OF: bus is default (na=1, ns=1) on
> usb at 47400000
> >     __of_translate_address() OF: parent bus is default (na=1, ns=1) on
> ocp
> >     of_translate_one() OF: no ranges, 1:1 translation
> >     of_translate_one() OF: with offset: 1195382784
> >     __of_translate_address() OF: parent bus is default (na=1, ns=1) on
> >     of_translate_one() OF: no ranges, 1:1 translation
> >     of_translate_one() OF: with offset: 1195382784
> >     __of_translate_address() OF: reached root node
> >     fdtdec_lookup_phandle() fdtdec_lookup_phandle: phys
> >     fdtdec_lookup_phandle() fdtdec_lookup_phandle: ti,ctrl_mod
> >     fdtdec_get_addr_size_fixed() fdtdec_get_addr_size_fixed: reg:
> fdtdec_get_addr_size_fixed() addr=44e10620
> >     fdtdec_get_int() fdtdec_get_int: mentor,multipoint: fdtdec_get_int()
> 0x1 (1)
> >     fdtdec_get_int() fdtdec_get_int: mentor,num-eps: fdtdec_get_int()
> 0x10 (16)
> >     fdtdec_get_int() fdtdec_get_int: mentor,ram-bits: fdtdec_get_int()
> 0xc (12)
> >     fdtdec_get_int() fdtdec_get_int: mentor,power: fdtdec_get_int()
> 0x1f4 (500)
> >     uclass_find_device_by_seq() 0 1
> >     uclass_find_device_by_seq() - 1 -1 'usb at 47401800'
> >     uclass_find_device_by_seq() - not found
> >     clk_set_defaults() clk_set_defaults(usb at 47401800)
> >     clk_set_default_parents() clk_set_default_parents: could not read
> assigned-clock-parents for 9df24d30
> >     ofnode_read_prop() ofnode_read_prop: assigned-clock-rates:
> ofnode_read_prop() <not found>
> >     musb_core_init() musb-hdrc: hw_ep 0shared, max 64
> >     musb_core_init() musb-hdrc: hw_ep 1tx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 1rx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 2tx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 2rx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 3tx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 3rx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 4tx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 4rx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 5tx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 5rx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 6tx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 6rx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 7tx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 7rx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 8tx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 8rx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 9tx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 9rx, max 512
> >     musb_core_init() musb-hdrc: hw_ep 10tx, max 256
> >     musb_core_init() musb-hdrc: hw_ep 10rx, max 64
> >     musb_core_init() musb-hdrc: hw_ep 11tx, max 256
> >     musb_core_init() musb-hdrc: hw_ep 11rx, max 64
> >     musb_core_init() musb-hdrc: hw_ep 12tx, max 256
> >     musb_core_init() musb-hdrc: hw_ep 12rx, max 64
> >     musb_core_init() musb-hdrc: hw_ep 13shared, max 4096
> >     musb_core_init() musb-hdrc: hw_ep 14shared, max 1024
> >     musb_core_init() musb-hdrc: hw_ep 15shared, max 1024
> >     musb_start() <== devctl 80
> >     musb_stop() HDRC disabled
> >     Port not available.
> >     usb_init() scan end
> >     ```
> >
> >     Best,
> >
> >     On Sun, Mar 7, 2021 at 3:04 AM Sean Anderson <seanga2 at gmail.com
> <mailto:seanga2 at gmail.com>> wrote:
> >
> >         On 3/6/21 8:38 PM, Jonas Vautherin wrote:
> >          > Thanks a lot Sean!
> >          >
> >          > I tried to enable logging by using the following CONFIG_:
> >          >
> >          > ```
> >          > CONFIG_LOG=y
> >          > CONFIG_SPL_LOG=y
> >          > CONFIG_TPL_LOG=y
> >          > CONFIG_LOG_MAX_LEVEL=5
> >
> >         You need to increase this. Log levels are only compiled-in if
> they are
> >         less than the max log level.
> >
> >          > CONFIG_SPL_LOG_MAX_LEVEL=3
> >          > CONFIG_TPL_LOG_MAX_LEVEL=3
> >          > CONFIG_LOG_DEFAULT_LEVEL=7
> >          > CONFIG_LOG_CONSOLE=y
> >          > CONFIG_SPL_LOG_CONSOLE=y
> >          > CONFIG_TPL_LOG_CONSOLE=y
> >          >
> >          > CONFIG_CMD_LOG=y
> >          > ```
> >          >
> >          > And I got the "log" function indeed. But it did not have the
> effect I was hoping for:
> >          >
> >          > ```
> >          > => log level set 7
> >
> >         Use the help function ;)
> >
> >         The correct syntax is "log level 7"
> >
> >         If you use a newer U-Boot (newer than 2021.01) then the output
> will be
> >         more verbose. E.g.
> >
> >         => log level
> >         EMERG
> >         ALERT
> >         CRIT
> >         ERR
> >         WARNING
> >         NOTICE
> >         INFO (default)
> >         DEBUG
> >
> >         --Sean
> >
> >          > => log level
> >          > Default log level: 0
> >          > => usb start
> >          > starting USB...
> >          > Bus usb at 47401800: Port not available.
> >          > ```
> >          >
> >          > First, I don't get why it says "Default log level: 0" after I
> set it (to whatever number, apparently). And second, it does not seem to
> bring me more output about the usb issue :-/.
> >          >
> >          > Am I missing something?
> >          >
> >          > Best,
> >          > Jonas
> >          >
> >          > On Sat, Mar 6, 2021 at 3:20 AM Sean Anderson <
> seanga2 at gmail.com <mailto:seanga2 at gmail.com> <mailto:seanga2 at gmail.com
> <mailto:seanga2 at gmail.com>>> wrote:
> >          >
> >          >     On 3/5/21 7:10 PM, Jonas Vautherin wrote:
> >          >      > Hello!
> >          >      >
> >          >      > I am learning about Fastboot, and I would like to run
> it from u-boot and
> >          >      > detect it from my host computer (I have the tools on
> my host, that's
> >          >      > working with other devices).
> >          >      >
> >          >      > I am using a pocketbeagle and I built an image using
> Yocto. I can get into
> >          >      > the u-boot prompt at boot time, and I get the
> following output:
> >          >      >
> >          >      > ```
> >          >      > Rebooting... reboot: Restarting system
> >          >      >
> >          >      > U-Boot SPL 2020.07 (Jul 06 2020 - 19:22:53 +0000)
> >          >      > Trying to boot from MMC1
> >          >      >
> >          >      >
> >          >      > U-Boot 2020.07 (Jul 06 2020 - 19:22:53 +0000)
> >          >      >
> >          >      > CPU  : AM335X-GP rev 2.1
> >          >      > Model: TI AM335x PocketBeagle
> >          >      > DRAM:  512 MiB
> >          >      > WDT:   Started with servicing (60s timeout)
> >          >      > NAND:  0 MiB
> >          >      > MMC:   OMAP SD/MMC: 0
> >          >      > Loading Environment from FAT... *** Warning - bad CRC,
> using default
> >          >      > environment
> >          >      >
> >          >      > <ethaddr> not set. Validating first E-fuse MAC
> >          >      > Net:   No ethernet found.
> >          >      > Hit any key to stop autoboot:  2
> >          >      > ```
> >          >      >
> >          >      > If I hit a key here, I get the u-boot prompt (`=>`)
> where I find
> >          >      > "fastboot". The thing is that none of the fastboot
> commands I run have any
> >          >      > output:
> >          >
> >          >     This is normal. There will be output only when you issue
> commands.
> >          >
> >          >      >
> >          >      > * fastboot usb 0
> >          >      > * fastboot usb
> >          >      > * fastboot udp
> >          >      >
> >          >      > Now, fastboot udp is not enabled in my defconfig, but
> the udp gadget
> >          >      > support is, and the usb fastboot gadget is enabled,
> too. From my host, I
> >          >      > just don't see anything in the `dmesg` output.
> >          >      >
> >          >      > Here are a few questions:
> >          >      >
> >          >      > * Does fastboot usb go over ethernet? If yes, I guess
> that the "Net: No
> >          >      > ethernet found." message is an issue that I need to
> fix such that a new usb
> >          >      > network interface is created by my host, but from my
> tests with this
> >          >      > proprietary device I own, it does not seem to be
> enumerated as a network
> >          >      > interface.
> >          >
> >          >     Only udp goes over ethernet (though I suppose you could
> use a
> >          >     usb-ethernet gadget/dongle).
> >          >
> >          >      > * In readthedocs [1], it mentions that "The
> CONFIG_USB_GADGET_VENDOR_NUM
> >          >      > must be one of the numbers supported by the fastboot
> client." In my setup,
> >          >      > it is set to 0x0451, and I did not find what is
> supposed to be supported in
> >          >      > fastboot.cpp [2]. So from the host, I also tried
> `fastboot -i 0x0451
> >          >      > devices`. Is that an issue?
> >          >
> >          >     The fastboot client detects devices with two bulk
> endpoints and the
> >          >     correct interface (see match_fastboot and
> filter_usb_device). Generally,
> >          >     you will only have to specify the vid/pid/serial if you
> have multiple
> >          >     devices plugged in at once.
> >          >
> >          >      > * I thought that maybe I should start USB, because the
> output of `usb tree`
> >          >      > says "USB is stopped. Please issue 'usb start'
> first.". But running `usb
> >          >      > start` errors with "Bus usb at 47401800: Port not
> available.". Is that related
> >          >      > to fastboot at all, i.e. should I be concerned about
> that? And how can I
> >          >      > debug it?
> >          >
> >          >     That's probably why your fastboot isn't working :)
> >          >
> >          >     You can try debugging it by enabling debug messages. If
> you want less than
> >          >     a firehose, you can try setting the default log level to
> info, and then
> >          >     setting it to debug before running usb start. See [1] for
> details.
> >          >
> >          >     [1]
> https://u-boot.readthedocs.io/en/latest/develop/logging.html
> >          >
> >          >     --Sean
> >          >
> >          >      > Thanks in advance for the help!
> >          >      > Jonas
> >          >      >
> >          >      > [1]:
> https://u-boot.readthedocs.io/en/latest/android/fastboot.html
> >          >      > [2]:
> >          >      >
> https://android.googlesource.com/platform/system/core/+/master/fastboot/fastboot.cpp
> >          >      >
> >          >
> >
>
>


More information about the U-Boot mailing list