Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Mac OSX Yosemite problems #69

Open
sg- opened this issue Oct 24, 2014 · 71 comments

Comments

Projects
None yet
@sg-
Copy link
Member

commented Oct 24, 2014

The OS rejects the format of the file system and puts it in a read only mode. Patch is in progress.

@sg- sg- added the bug label Oct 24, 2014

@rei-vilo

This comment has been minimized.

Copy link

commented Oct 24, 2014

Issue confirmed on

  • Nucleo F401RE
  • Freedom KL25Z with offical mbed driver 20140530_k20dx128_kl25z_if_opensda
  • RedBearLab nRF51822 (updated Oct 30, 2014)
@gonzoveliki

This comment has been minimized.

Copy link

commented Oct 24, 2014

this issue is pending also on FRDM-K64F with latest mbed firmware loaded and using
Mac OS X Yosemite 10.10 official, up-to-date

@matt-land

This comment has been minimized.

Copy link

commented Oct 26, 2014

Also ran into this with my K64F today.

@perjg

This comment has been minimized.

Copy link

commented Oct 26, 2014

I'm getting the exact same issue on LPCXpresso LPC1549

@gonzoveliki

This comment has been minimized.

Copy link

commented Oct 26, 2014

for some stupid "security" reasons, yosemite recognizes MBED's removable storage drive to be broken and unrecoverable, then mounts it as read-only filesystem.
i have tested with all available FRDM and Nucleo boards, and all of them are affected by that issue.

@0xc0170

This comment has been minimized.

Copy link
Member

commented Oct 26, 2014

Thank you guys for reporting. Appreciate it 👍

@rei-vilo

This comment has been minimized.

Copy link

commented Oct 30, 2014

Same issue with

  • RedBearLab nRF51822
@DoZh

This comment has been minimized.

Copy link

commented Oct 31, 2014

Also have problem on Nucleo F103RB

@teddokano

This comment has been minimized.

Copy link

commented Nov 1, 2014

I've just shared a small tool for the time being ;)
https://developer.mbed.org/users/okano/notebook/mbed-on-yosemite/

@sg-

This comment has been minimized.

Copy link
Member Author

commented Nov 1, 2014

@teddokano thanks for this. patch is almost complete

@jcw

This comment has been minimized.

Copy link

commented Nov 2, 2014

Looking forward to it!
-jcw

@rolandking

This comment has been minimized.

Copy link

commented Nov 10, 2014

Same issue on Nordic's new nRF51-DK and nRF51-Dongle (PCA10028 and PCA10031)

@rei-vilo

This comment has been minimized.

Copy link

commented Nov 21, 2014

Any news?

@sg-

This comment has been minimized.

Copy link
Member Author

commented Nov 21, 2014

coming along nicely. Its been a complete logic rewrite. Testing on KL25z and only blocker at the moment is debuggiung sector erase as this is required rather than doing a full chip erase. Relates to improvements for pyOCD too.
https://github.com/sg-/CMSIS-DAP/tree/fs_patch

Meanwhile hope you saw the cmd line and script that was created by teddokano a few comments up. You should be out of service so taking the time to make some necessary improvements.

@bbirand

This comment has been minimized.

Copy link

commented Dec 15, 2014

I have the same issue on the Seeedstudio Arch BLE. Looking forward for a solution!

@jockm

This comment has been minimized.

Copy link

commented Dec 15, 2014

It has been 47 days since this issue was reported. This is not a trivial issue, at the very least I would except some regular updates. It has been 24 days since the last update of "coming along nicely"

@bbirand

This comment has been minimized.

Copy link

commented Dec 15, 2014

Apparently the following is a workaround, but I haven't tried it yet:
https://developer.mbed.org/users/okano/notebook/mbed-on-yosemite/

@sg-

This comment has been minimized.

Copy link
Member Author

commented Dec 15, 2014

@jockm Updates come in response to questions :D unless you want to follow the commit history.

https://github.com/sg-/CMSIS-DAP/commits/fs_patch

@jockm

This comment has been minimized.

Copy link

commented Dec 15, 2014

@bbirand Yes it is linked to above and mentioned in @sg's comment. Unfortunately it didn't work for me (Mid 2011 i5 iMac, 16G RAM, OSX Yosemite 10.10.1).

I hope I am the outlier and not the norm. Right now I have to attach my Nordic mBED board to my HTPC and scp the file over there, and I then watch the output via a webcam. This is non-optimal at best

@jockm

This comment has been minimized.

Copy link

commented Dec 15, 2014

I am sorry sg, I really don't like that answer. You have a showstopping bug for OSX and at least some linux users that has been around for a month and a half and you can't be bothered to, say, give a little update once a week or so to let people know when they can use the hardware they paid for?

As someone who has to jump through hoops in order to use y hardware I don't think the smiley is called for.

But since you only respond to question: When can we reasonably expect this to be done?

@rei-vilo

This comment has been minimized.

Copy link

commented Dec 15, 2014

Other development platforms suffered from a similar glitch. As always, Apple doesn't provide much documentation on the changes. Good luck!

@rei-vilo

This comment has been minimized.

Copy link

commented Dec 15, 2014

For the Nucleo boards from ST, give STLINK a try.
Very easy to use and works fine.

@sg-

This comment has been minimized.

Copy link
Member Author

commented Dec 15, 2014

@jockm Working on merging the patch with master, testing and releasing shortly thereafter. Should be done in the next few weeks.

Would you mind explaining where the current work around fails for you and the platform you're using? There maybe someone out there who can spot why.

@jockm

This comment has been minimized.

Copy link

commented Dec 15, 2014

@sg It just fails silently. Pressing reset, or disconnecting and reconnecting, do nothing. I am using the Nordic nRF51 mBED board

My only solution is the one I describe some comments before:

Right now I have to attach my Nordic mBED board to my HTPC and scp the file over there, and I then watch the output via a webcam.

I have a client who wanted the solution to be mBED based since he wants to be able to modify and tweak himself. There are buttons, there are sensors, etc and every time I need to go interact with it I need to get out of my office, walk down the hall, interact with it, and then walk back.

So you will, perhaps, understand that I am not predisposed to being casually accepting of this issue

@rolandking

This comment has been minimized.

Copy link

commented Dec 15, 2014

I had mixed results with the workaround. About 1/4 of the time the device mounted, then unmounted itself again with an error dialog box about ejecting devices when still connected even though it was. I managed to get an image on it a couple of times but it was hit and miss. I finally managed to brick the thing entirely and had to get a windows box on it.

I have one of the Nordic boards which can be MBED or embedded JLink. In JLink mode it also presents as a disk device when plugged in and that suffered from the same issue when 10.10 came out. I contacted Segger (via Nordic) and they fixed that up in a beta a few weeks ago, still however it fails to mount the boot partition properly, so it looks like the fix was only applied to one piece of it.

Having bricked and recovered it once I've gone back to using JLink until the patch proper is done, the workaround wasn't reliable.

@mikew67

This comment has been minimized.

Copy link

commented Dec 28, 2014

Quick note for jockm and others using Nordic mbed board with Mac Yosemite. I'm running a Win7 VM using Parallels on my MacBook Pro. Mounting the BOOTLOADER volume on the Win7 side and doing mbed complies and loads there works fine. Might be a solution for you if you don't have a Windows hardware box around.
HTH,
Mike (who's also waiting for the mbed/OS X fix)
PS: Have not tried mounting the BOOTLOADER volume on the Win VM side and then writing to it from the Mac OS X side (as I can do on the C: HD when the VM is running). Will let you know if that works.

@mikew67

This comment has been minimized.

Copy link

commented Dec 28, 2014

Nope, no luck writing from the Mac OS X Yosemite side to a running Parallels Win7 VM with a mounted Nordic mbed BOOTLOADER volume. Got the following error:

Mike
screen shot 2014-12-27 at 4 23 35 pm

@jockm

This comment has been minimized.

Copy link

commented Dec 28, 2014

I hadn't tried Parallels, but I did try VirtualBox with a Win7 VM, and it refused to even try and mount the device. I get a similar kind of error message on my Windows 8.1 laptop.

I am lucky in that I do have one system in my house that can interact with my mBed board, but it is very clumsy and inconvenient.

@lucaim777

This comment has been minimized.

Copy link

commented Jan 21, 2015

Guys any news on Nucleo F401RE ?

@cdwilson

This comment has been minimized.

Copy link

commented Jan 22, 2015

@sg- how do ST based devices fit into the roadmap? I'm just curious who is responsible for updating the ST-Link based devices like the Nucleo F411RE, you guys or ST?

@lucaim777 I'm using https://developer.mbed.org/users/okano/notebook/mbed-on-yosemite/ on OS X Yosemite and it works fine after making the fixes I posted in the comments at the bottom of the page (https://developer.mbed.org/users/okano/notebook/mbed-on-yosemite/?compage=1#c15249)

@0xc0170

This comment has been minimized.

Copy link
Member

commented Jan 22, 2015

@cdwilson ST is in charge of ST Link, as you could have noticed they haven't contributed their work. You can write to their support and share it with mbed community if you get any response.

@cdwilson

This comment has been minimized.

Copy link

commented Jan 22, 2015

@0xc0170 that's what I would assume, but I wasn't sure, because the answer from the Austin MBED team in http://developer.mbed.org/questions/5001/Can-not-write-binaries-into-NUCLEO-disk-/ seems to indicate that MBED is fixing it.

@KleinYuan

This comment has been minimized.

Copy link

commented Jan 27, 2015

Same problem with nrf51822, BLE NANO.
Tried to use the command solution : https://developer.mbed.org/users/okano/notebook/mbed-on-yosemite/
Ending up with Device not configured error...
Waiting for firmware update

@xiongyihui

This comment has been minimized.

Copy link
Contributor

commented Jan 28, 2015

The update of Seeed Arch Pro is available now.
https://developer.mbed.org/platforms/Seeeduino-Arch-Pro/#update-or-restore-firmware

The source code is on https://github.com/xiongyihui/CMSIS-DAP/tree/arch_pro
The branch is based on sam's fs_patch branch with some changes of stack size and RAM usage.

@sg-
The task running USBD_RTX_MSC_EP_BULK_Event() requires a lot of memory. In my test, it takes at least 308 bytes of its stack. Probably, we should use a user defined stack for it.

Tasks RAM usage from my experience:

  • stk_timer_30_task > 64
  • stk_dap_task > 168
  • stk_serial_task > 240
  • main_task > 88
  • USBD_RTX_MSC_EP_BULK_Event > 308
@sg-

This comment has been minimized.

Copy link
Member Author

commented Jan 28, 2015

@xiongyihui Thanks for this noting this. Will you create a new task here so once this is closed this doesnt get lost. Once past this patch and some new tools added - when all is merged optimizing for RAM and a list of other things todo will be public for anyone who wants to contribute :D

@cdwilson

This comment has been minimized.

Copy link

commented Jan 29, 2015

@lucaim777 (and anybody else with NUCLEO mbed platform)

FYI, I've been talking with one of the ST FAEs who supports our group and he said that this issue "should be resolved in about a month" for the ST-LINK/V2-1 chipset on the NUCLEO platforms.

@lucaim777

This comment has been minimized.

Copy link

commented Jan 29, 2015

@cdwilson Thanks for the update, can't wait for the release !

@xiongyihui

This comment has been minimized.

Copy link
Contributor

commented Jan 31, 2015

The Seeed Arch BLE gets a update to support Mac OS X Yosemite. Follow the following link to update your Arch BLE : ) @bbirand
https://developer.mbed.org/platforms/Seeed-Arch-BLE/#update-or-restore-firmware

The source code is on https://github.com/xiongyihui/CMSIS-DAP/tree/arch_ble. It needs some clean-up.

@sg-
We should find a time to merge some branches. There are so many branches around. Not easy to maintain.

@sg-

This comment has been minimized.

Copy link
Member Author

commented Feb 2, 2015

Thanks @xiongyihui - Last thing to complete before the merge is hex support for all targets. Great job working this in. The branches I know that need merging are (all based on the /sg-/fs_patch branch):

I'll start merging this if you're feature complete and tested. Can you confirm the following hex files program correctly?

  1. standard linear records
  2. non-continuos record addressing
  3. 16, 32 and file of records of any size in-between?

On a separate note: I've started coming across odd behavior where swd_write_memory doesn't actually write properly. Typical symptom is flash_init failing when programming begins. The BKPT instruction never halts because the algo blob isnt downloaded correctly. Verified by adding verification to these writes for testing. Interestingly here, doing swd_write_byte (very slow) works fine. This leads me to believe there is a pointer that is not being de-referenced properly somewhere in swd_host. I've now seen this on builds for the (K20 -> KL25, K20 -> KL46, and ATSAM3U -> nRF51822mkit).

Another troubling thing is a verification dump from the target when programming is complete fails (swd_read_memory). Have you seen this behavior? sg-@da859da#diff-2690f2d73aa409d1bfd8ab9090c6f302R244

Anyways looking forward to merging once these few things are worked out and then moving in the rest of the tools changes from other dev branches and lots more 👍

@xiongyihui

This comment has been minimized.

Copy link
Contributor

commented Feb 2, 2015

@sg-
just updated https://github.com/xiongyihui/CMSIS-DAP/tree/arch_ble. It should supports:

  1. standard linear records
  2. non-continuos record addressing
  3. 16 or 32 bytes data record (depends on line buffer size)

At the beginning. target_flash_init failed too. Then I saw @toyowata added the following lines in target_flash_init.

    PORT_SWD_SETUP();
    if (!target_set_state(RESET_PROGRAM)) {
        return TARGET_FAIL_RESET;
    }

It solved the issue. It seems that we should halt the target before we write the algo blob to the target's RAM.

@aufflick

This comment has been minimized.

Copy link

commented Feb 4, 2015

I'm a little confused by some of the update/board names. Is there an update yet for the Nordic branded nRF51822-mKIT dev board?

@rolandking

This comment has been minimized.

Copy link

commented Feb 4, 2015

@aufflick - don't believe there's anything for the nrf51822 released yet.

@sg - would it be possible to have a quick summary of the boards / host chips for which updates are done and a list of those still being worked on? It would be a useful milestone for those of us waiting. I'm assuming for instance that basic ATSAM3U has to be done and then updates for each of the target boards using that will drop in sequence sometime afterwards and the same for other chips.

And can anyone just confirm that without a windows box and access to a full Keil environment, it's not possible to help in the effort, else I would try.

@KleinYuan

This comment has been minimized.

Copy link

commented Feb 4, 2015

@aufflick Hey . For NRF51822, though firmware not updated officially. When I tried use Terminal to cp .hex file to MBED drive, though still showing deviced not configured, it actually works.
Just use: sudo -mount -u -w -o sync /Volumes/MBED; cp -X (Drag your file here) /Volumes/MBED

@aufflick

This comment has been minimized.

Copy link

commented Feb 4, 2015

@KleinYuan yep, but there must have been something deeper odd with my board - it mounted a JLINK volume not an MBED volume. I then stupidly copied a hex file to the BOOTLOADER volume as per comments above (the BOOTLOADER volume mounts if you power on the board holding down the reset button) and now my board is bricked (presumably because the BOOTLOADER volume actually does what it says...) :(

@xiongyihui

This comment has been minimized.

Copy link
Contributor

commented Feb 4, 2015

The Arch BLE board is a nRF51822 board with LPC11U35. The update is available. It can be used for the mbed HRM1017 which is also LPC11U35 + nRF51822.

The mKit and Redbear's board is different.

@rei-vilo

This comment has been minimized.

Copy link

commented Feb 26, 2015

Any news for Nucleo, Freedom and RedBearLab boards?

@rolandking

This comment has been minimized.

Copy link

commented Feb 26, 2015

... and Nordic using the Atmel chip?

@jmarcelino

This comment has been minimized.

Copy link

commented Mar 1, 2015

I'm using the latest Yosemite beta (10.10.3 build 14D87h) and it no longer displays this issue on my unpatched MBED boards (tested with the Nordic and ST Nucleo)

The MBED drive now mounts normally.

Of course I'm not sure when this beta will be released and if the fix will still be on the released version, but it might be a faster option than waiting for vendors to update their firmware.

@rolandking

This comment has been minimized.

Copy link

commented Mar 1, 2015

@jmarcelino - I only had 10.10.3 on one of my machines and hadn't tested it with my Nordic boards, now I have, I concur the drive now mounts normally. Not only does it work with the current un-patched MBED firmware, but for the dual-firmware boards which can be MBED or JLINK OB, the JLINK bootloader which suffered from the same issue, now also works properly.

That's great news, I was losing hope of ever being able to use the MBED piece of the boards again.

Hopefully the fix will persist into the release version, if it doesn't, I'll file a bug report that they broke it again!

I tested the Nordic- NRF51-DK and Nordic Dongle (PCA10028 and PCA100031).

@rei-vilo

This comment has been minimized.

Copy link

commented Mar 1, 2015

@jmarcelino @rolandking Thank you both for your feed-back.

I'm awaiting the official release of Mac OS X 10.10.3 as I can't risk installing a beta on the computer I use for production.

@garfield38

This comment has been minimized.

Copy link

commented Mar 31, 2015

You don't need to update Yosemite, ST released a new firmware that fix the issue and it's work on my 10.10.2

@rei-vilo

This comment has been minimized.

Copy link

commented Mar 31, 2015

@garfield38 Thank you for sharing. I'll give a try. It works!

Any news from the other mbed-enabled boards manufacturers?

@lucaim777

This comment has been minimized.

Copy link

commented Mar 31, 2015

@garfield38 Confirmed, it's working on Nucleo F401RE and 10.10.2. Thanks !

@cdwilson

This comment has been minimized.

Copy link

commented Apr 2, 2015

Awesome, ST-Link upgrader now supports Mac/Linux 👍

@rolandking

This comment has been minimized.

Copy link

commented Apr 10, 2015

OSX 10.10.3 is now released and, like the betas, still appears to mount both the drive and bootloader without error. So for those people with boards which never received an update for this issue, which is at least the Nordic boards with Atmel chips and may be others, an upgrade to 10.10.3 should have you working again.

@rei-vilo

This comment has been minimized.

Copy link

commented Apr 15, 2015

A blatant evidence of the very poor support from mbed. I guess all the engineers were focused on the next big thing, mbed OS.

The only chipmaker which updated the firmware was STM.

Needless to say I'm reconsidering the mbed platform fro future projects.

@anpilog

This comment has been minimized.

Copy link

commented Apr 16, 2015

CMSIS-DAP based on LPC11U35 is working fine on OSX 10.10.3 even with very old firmware.

@c1728p9

This comment has been minimized.

Copy link
Contributor

commented Aug 5, 2016

There are several OS X fixes in the DAPLink project so OS X should be fully supported. If any problems are found please open an issue here: https://github.com/mbedmicro/daplink

CC
@anpilog
@rei-vilo
@rolandking
@cdwilson
@lucaim777
@garfield38
@jmarcelino
@xiongyihui
@aufflick
@KleinYuan
@27percent
@toyowata
@jockm
@mikew67
@jcw
@teddokano
@TATQAQTAT
@gonzoveliki
@perjg
@matt-land

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
You can’t perform that action at this time.