1130288 - Enable FOTA to update necessary bluetooth daemon files for v2.2

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Bluetooth, defect)

Description

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Because we moved to bluetooth daemon architecture, Gecko bluetooth module now depends on bluetoothd. If there is any IPC protocol (currently we used BlueZ Protocol) or bluetoothd fix, OTA users cannot get updated. OTA users need to update system partition to get update. This might lead incompatible update since gecko/bluetoothd are mismatched.

I think current OTA mechanism just simply pack /system/b2g folder and replace b2g folder during OTA updating, so it will be easier to pack necessary files into /system/b2g for bluetooth daemon.

Comment 1

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

To achieve this, a few things we need to do:
1. Change Android.mk for bluedroid to put bluetoothd to /system/b2g (maybe it's necessary to move libfdio.so from /system/lib to /system/b2g?)
2. Change init.bluetooth.rc to replace /system/bin/bluetoothd to /system/b2g/bluetoothd

Thomas, do you have any concern or suggestion?

Comment 2

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

ni :kli
Since he had some experience about OTA solution.
It will be good if you can raise any concern if we put bluetoothd under /system/b2g.

Comment 3

[Thomas Zimmermann [:tzimmermann] [:tdz]]

> To achieve this, a few things we need to do:
> 1. Change Android.mk for bluedroid to put bluetoothd to /system/b2g (maybe
> it's necessary to move libfdio.so from /system/lib to /system/b2g?)
> 2. Change init.bluetooth.rc to replace /system/bin/bluetoothd to
> /system/b2g/bluetoothd
> 
> Thomas, do you have any concern or suggestion?

Not in general, but I'm curious how others handle this problem. Other system daemons reside in '/system/bin' AFAIK.

@kli: Is is possible to add things outside of /system/b2g to OTA updates? AFAIK the mechanism can update any file. So it seems rather a policy thing to only update /system/b2g.(In reply to Shawn Huang [:shawnjohnjr] from comment #1)

Comment 4

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Bug 1130293 looks like nfcd doesn't get OTA updates as well.

We also need to move libfdio.so to /system/b2g. This might lead to problems with the dynamic linker if the directory is not in the linker's search path.

Comment 5

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Thomas Zimmermann [:tzimmermann] [:tdz] from comment #3)
> > To achieve this, a few things we need to do:
> > 1. Change Android.mk for bluedroid to put bluetoothd to /system/b2g (maybe
> > it's necessary to move libfdio.so from /system/lib to /system/b2g?)
> > 2. Change init.bluetooth.rc to replace /system/bin/bluetoothd to
> > /system/b2g/bluetoothd
> > 
> > Thomas, do you have any concern or suggestion?
> 
> Not in general, but I'm curious how others handle this problem. Other system
> daemons reside in '/system/bin' AFAIK.
> 
> @kli: Is is possible to add things outside of /system/b2g to OTA updates?
> AFAIK the mechanism can update any file. So it seems rather a policy thing
> to only update /system/b2g.(In reply to Shawn Huang [:shawnjohnjr] from
> comment #1)
I would say if OTA update script can update /system/bin/bluetoothd, /system/lib/libfdio.so, it will be the best solution, so we don't need to change any code.

Comment 6

[Thomas Zimmermann [:tzimmermann] [:tdz]]

I'd prefer this as well.

Comment 7

[Kai-Zhen Li [:seinlin][:kli]]

Now, OTA can only update /system/b2g/*. To update files outside /system/b2g with OTA, I think we need to enhance the current OTA design. But we can use FOTA to update any file under /system/* easily, as FOTA will reboot into recovery ROM and can update files to /boot /data and /system partitions.

There are some options to have bluetoothd/nfcd(bug 1130293) and related library files get updated:

1. move these daemon and library into /system/b2g dir. 
  (This could be a little bit tricky, but less effort)

2. when these daemon and library files need to be updated, we pack and update them in FOTA format.

3. enhance the current OTA design, allow it update files outside /system/b2g/ dir.


One more thing come in my mind about the update of bluetoothd. It is init.bluetooth.rc which is in ramdisk of boot.img and I think it can be updated via FOTA only. I am not sure if there is demand to update this file. If we need to consider it, method 2 could be the most suitable one.

Comment 8

[Kai-Zhen Li [:seinlin][:kli]]

Alexandre and Gabriele, per the above discussion, do you have any other idea on How to have files outside /system/b2g dir get updated?

Comment 9

[Kai-Zhen Li [:seinlin][:kli]]

Fabrice, do you have any other suggestion?

Comment 10

[:gerard-majax]

A couple of things:
 - we can do whatever we want when in FOTA, but we currently do not support everything
 - current FOTA package installation allows to not only touch /system/b2g but anything, this is used on some builds to update fonts (that are living in /system/fonts) ; the directories used are controlled by the B2G_FOTA_DIRS environment variable (gonk-misc/Android.mk)
 - we could flash new partitions (needed for boot.img changes) but this is not supported for now

So, what you can do right now is:
 - make sure your package install target puts your binary and libs into /system/b2g/ ; then those will be picked up by default in OTA and FOTA
 - augment the current edify script applied during FOTA installation to remove the stale binaries you know about (/system/bin/bluetoothd for instance)

What you cannot do:
 - flash new boot.img

As far as I know, the current Flame updates that are distributed are only OTA ones ; this can probably be easily switched server-side.

Comment 11

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Kai-Zhen Li [:seinlin] from comment #7)
> One more thing come in my mind about the update of bluetoothd. It is
> init.bluetooth.rc which is in ramdisk of boot.img and I think it can be
> updated via FOTA only. I am not sure if there is demand to update this file.
> If we need to consider it, method 2 could be the most suitable one.
Honestly FOTA can fix the our problems, there are so many people are complaining bluetooth is dead after having OTA update, and we're trying to announce this information cross mailing lists to ask people full flash images but the results are not good :(

Comment 12

[[:fabrice] Fabrice Desré]

(In reply to Alexandre LISSY :gerard-majax from comment #10)
> A couple of things:
>  - we can do whatever we want when in FOTA, but we currently do not support
> everything
>  - current FOTA package installation allows to not only touch /system/b2g
> but anything, this is used on some builds to update fonts (that are living
> in /system/fonts) ; the directories used are controlled by the B2G_FOTA_DIRS
> environment variable (gonk-misc/Android.mk)
>  - we could flash new partitions (needed for boot.img changes) but this is
> not supported for now
> 
> So, what you can do right now is:
>  - make sure your package install target puts your binary and libs into
> /system/b2g/ ; then those will be picked up by default in OTA and FOTA
>  - augment the current edify script applied during FOTA installation to
> remove the stale binaries you know about (/system/bin/bluetoothd for
> instance)
> 
> What you cannot do:
>  - flash new boot.img

But we need to do that if we want to move the daemons to /system/b2g to update the init.rc files, right?

Comment 13

[:gerard-majax]

That's my understanding also. Being able to flash partition should not be a big issue:
 - we need to embed the partition image in the update zip
 - we need the binary to write the partition (I don't remember the name)
 - we need the target partition device name

Comment 14

[Thomas Zimmermann [:tzimmermann] [:tdz]]

We'll need at least one FOTA update anyway, and if only to update init.b2g.rc.

I'd expect to have more daemons in the future and we probably don't want them to all reside in /system/b2g. So how much work is it to add support for non-b2g directories/files to the OTA script?

Comment 15

[Gabriele Svelto [:gsvelto]]

Clearing my NI as Alexandre already replied in comment 10.

Comment 16

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Hi Chris,
Do you know how much work if OTA (or FOTA) can push binary files (boot.img and /system/bin/bluetoothd, /system/lib/libfdio.so)? There are many people complaining bluetooth is not working after OTA update, would you please answer questions in Comment 14? AFAIK, OTA controlled by Release Engineering team.

Comment 17

[Eric Tsai (no more review request accepted)]

(In reply to Alexandre LISSY :gerard-majax from comment #13)
> That's my understanding also. Being able to flash partition should not be a
> big issue:
>  - we need to embed the partition image in the update zip
>  - we need the binary to write the partition (I don't remember the name)
>  - we need the target partition device name

The update.zip sign key should be the same one in recovery partition, others will fail.
For stable release (e.g. 18D, using T2M own key) with shallow flash to nightly build, I think FOTA will still fail.

Comment 18

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Eric Tsai from comment #17)
> (In reply to Alexandre LISSY :gerard-majax from comment #13)
> > That's my understanding also. Being able to flash partition should not be a
> > big issue:
> >  - we need to embed the partition image in the update zip
> >  - we need the binary to write the partition (I don't remember the name)
> >  - we need the target partition device name
> 
> The update.zip sign key should be the same one in recovery partition, others
> will fail.
> For stable release (e.g. 18D, using T2M own key) with shallow flash to
> nightly build, I think FOTA will still fail.
How can we achieve FOTA update without sign key? It looks like it's impossible to do FOTA update now with this sign key. :(

Comment 19

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Shawn Huang [:shawnjohnjr] from comment #18)
> (In reply to Eric Tsai from comment #17)
> > (In reply to Alexandre LISSY :gerard-majax from comment #13)
> > > That's my understanding also. Being able to flash partition should not be a
> > > big issue:
> > >  - we need to embed the partition image in the update zip
> > >  - we need the binary to write the partition (I don't remember the name)
> > >  - we need the target partition device name
> > 
> > The update.zip sign key should be the same one in recovery partition, others
> > will fail.
> > For stable release (e.g. 18D, using T2M own key) with shallow flash to
> > nightly build, I think FOTA will still fail.
> How can we achieve FOTA update without sign key? It looks like it's
> impossible to do FOTA update now with this sign key. :(

typo: s/with/without

Comment 20

[:gerard-majax]

It's just a matter of who issues/signs the package. If T2M does and then it's getting distributed, there should be no reason.

If people flashed a rebuilt recovery, it should be fine.

Comment 21

[Eric Tsai (no more review request accepted)]

(In reply to Shawn Huang [:shawnjohnjr] from comment #19)
> > How can we achieve FOTA update without sign key? It looks like it's
> > impossible to do FOTA update now with this sign key. :(
> 
> typo: s/with/without

(In reply to Alexandre LISSY :gerard-majax from comment #20)
> It's just a matter of who issues/signs the package. If T2M does and then
> it's getting distributed, there should be no reason.
> 
> If people flashed a rebuilt recovery, it should be fine.

Agree with Alexandre, who provides recovery partition is able to verify the update package.

IMO, we may 1) replace the key in recovery partition by full flash, 2) get manufacturer's key (I don't know if there is any use test key) or 3) let manufacturer sign the package for us. All above need TAM's support by each device, no silver bullet.

Comment 22

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Thanks Alexandre and Eric. And I also talked with Wesley today who is working on V18D_nightly package.
So my understanding now:
1. V18D package from MDN is with manufacturer's key (but for release v2.0)
2. V18D_nightly package from MDN is with test sign key that Mozilla can use it for (F)OTA
3. Naoki is going to pack another version for V18D_nightly, this package contains V18D from T2M + new gecko/gaia + necessary bluetoothd files
4. If user flash V18D_nightly, user will have chance to update via (F)OTA in the future.

Comment 23

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Attachment: Github pull request for platform_build on KK

Comment 24

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Attachment: Github pull request for B2G

Comment 25

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Attachment: Github pull request for gonk-misc

Comment 26

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8564120 [details]
Github pull request for platform_build on KK

I guess this is located separately from the rest of the update scripts because it might be platform specific. I'll send out pull request for the remaining branches after this one got its OK.

Comment 27

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Using this patch set, I'm able to update the individual files of the Bluetooth daemon.

Comment 28

[Gabriele Svelto [:gsvelto]]

Comment on attachment 8564120 [details]
Github pull request for platform_build on KK

Looks good to me, I've just noted a nit in the PR.

Comment 29

[Gabriele Svelto [:gsvelto]]

Comment on attachment 8564125 [details]
Github pull request for gonk-misc

This is also looking good.

Comment 30

[Gabriele Svelto [:gsvelto]]

Comment on attachment 8564124 [details]
Github pull request for B2G

I'm a bit unsure about the optimization you removed: from your comment, and from glancing over the code it seems to be removing directories that shouldn't be part of the update.zip file. I wonder if removing it would cause us to leave unneeded stuff in the zip.

Comment 31

[Thomas Zimmermann [:tzimmermann] [:tdz]]

(In reply to Gabriele Svelto [:gsvelto] from comment #30)
> Comment on attachment 8564124 [details]
> Github pull request for B2G
> 
> I'm a bit unsure about the optimization you removed: from your comment, and
> from glancing over the code it seems to be removing directories that
> shouldn't be part of the update.zip file. I wonder if removing it would
> cause us to leave unneeded stuff in the zip.

Is it a requirement to have minimal scripts and zip files? Because of the changes, the updated files are now processed individually; before the scripts handled complete directories IIRC.

Comment 32

[Thomas Zimmermann [:tzimmermann] [:tdz]]

(In reply to Gabriele Svelto [:gsvelto] from comment #30)
> Comment on attachment 8564124 [details]
> Github pull request for B2G
> 
> I'm a bit unsure about the optimization you removed: from your comment, and
> from glancing over the code it seems to be removing directories that
> shouldn't be part of the update.zip file. I wonder if removing it would
> cause us to leave unneeded stuff in the zip.

Regarding this optimization: this seems to be related to not traversing directories if they don't contain updated files (i.e., if they are part of the updated-files list).

This optimization breaks when updating single files, because the base path of single files is not part of the file list; only the file path itself.

Comment 33

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Ken, Wesley,

we will need new base images for the Flame (master + v2.2) before enabling bluetoothd. Can you help with that?

Comment 34

[Ken Chang[:kenkai]]

I believe that you want to ni? Wesly Huang.

Comment 35

[Gabriele Svelto [:gsvelto]]

Comment on attachment 8564124 [details]
Github pull request for B2G

As per our IRC discussion this is only going to affect the script where we'll have a few more chmod's. That's not something worth bothering about.

Comment 36

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Oh! 'Wesly' and 'Wesley' Huang are not the same person!

Comment 37

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Attachment: Github pull request for platform_build on L

Same code as for KK

Comment 38

[Gabriele Svelto [:gsvelto]]

Comment on attachment 8565893 [details]
Github pull request for platform_build on L

Rubber stamping since it's the same change :)

Comment 39

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Attachment: Github pull request for platform_build on JB

Changes to KK:

  - no support for SE labels and capabilities, so references to |self.selabel| and |se.capabilities| have been removed

Comment 40

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Attachment: Github pull request for platform_build on ICS

Same code as on JB

Comment 41

[Gabriele Svelto [:gsvelto]]

Comment on attachment 8565896 [details]
Github pull request for platform_build on JB

Looking good.

Comment 42

[Gabriele Svelto [:gsvelto]]

Comment on attachment 8565901 [details]
Github pull request for platform_build on ICS

Likewise.

Comment 43

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8564120 [details]
Github pull request for platform_build on KK

Updated github pull request

  - removed an empty line

Comment 44

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8565893 [details]
Github pull request for platform_build on L

Updated Github pull request

  - removed an empty line

Comment 45

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8565896 [details]
Github pull request for platform_build on JB

Updated Github pull request

  - removed an empty line

Comment 46

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8565901 [details]
Github pull request for platform_build on ICS

Updated Github pull request

  - removed an empty line

Comment 47

[Thomas Zimmermann [:tzimmermann] [:tdz]]

This patch set will be required in v2.2 to update the system programs, such as bluetoothd.

Comment 48

[Ryan VanderMeulen [:RyanVM]]

Master: https://github.com/mozilla-b2g/B2G/commit/42dbad3bf1a2c7f8247f7e20094d7b144a31a2e6

Master: https://github.com/mozilla-b2g/gonk-misc/commit/3905180d9e2dbe03d7ce3b3a6debe4d4adc938d1

Master: https://github.com/mozilla-b2g/platform_build/commit/173b3104bfcbd23fc9dccd4b0035fc49aae3d444

b2g-4.3_r2.1: https://github.com/mozilla-b2g/platform_build/commit/4efd19d199ae52656604f794c5a77518400220fd

b2g-4.4.2_r1: https://github.com/mozilla-b2g/platform_build/commit/ef937d1aca7c4cf89ecb5cc43ae8c21c2000a9db

b2g-5.0.0_r6: https://github.com/mozilla-b2g/platform_build/commit/0ca5e1456dfd417c3e4e72bb8eae02069136b99b

Comment 49

[Stephany Wilkes]

Triage is blocking, as Bluetooth won't work without this, and this needs to be uplifted from master to 2.2.

Comment 50

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8564120 [details]
Github pull request for platform_build on KK

NOTE: Please see https://wiki.mozilla.org/Release_Management/B2G_Landing to better understand the B2G approval process and landings.

[Approval Request Comment]
Bug caused by (feature/regressing bug #): 

  General short-coming of current update infrastructure; exposed by bug 1005934

User impact if declined: 

  Users won't be able to update system daemons, such as bluetoothd, on
  v2.2. Those programs are an integrate piece of Gecko's functionality.

Testing completed: 

  Locally generated and installed an update

Risk to taking this patch (and alternatives if risky): 

  Could be interfere with generation of update packages; some QA is advised

String or UUID changes made by this patch:

  None

Comment 51

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8564124 [details]
Github pull request for B2G

NOTE: Please see https://wiki.mozilla.org/Release_Management/B2G_Landing to better understand the B2G approval process and landings.

[Approval Request Comment]

  Please see comment #50

Comment 52

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8564125 [details]
Github pull request for gonk-misc

NOTE: Please see https://wiki.mozilla.org/Release_Management/B2G_Landing to better understand the B2G approval process and landings.

[Approval Request Comment]

  Please see comment #50

Comment 53

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8565893 [details]
Github pull request for platform_build on L

NOTE: Please see https://wiki.mozilla.org/Release_Management/B2G_Landing to better understand the B2G approval process and landings.

[Approval Request Comment]

  Please see comment #50

Comment 54

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8565896 [details]
Github pull request for platform_build on JB

NOTE: Please see https://wiki.mozilla.org/Release_Management/B2G_Landing to better understand the B2G approval process and landings.

[Approval Request Comment]

  Please see comment #50

Comment 55

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8565901 [details]
Github pull request for platform_build on ICS

NOTE: Please see https://wiki.mozilla.org/Release_Management/B2G_Landing to better understand the B2G approval process and landings.

[Approval Request Comment]

  Please see comment #50

Comment 56

[bhavana bajaj [:bajaj]]

Ni, Nhirata, to help test this patch on 3.0 before uplifting to branch.

Comment 57

[Wesly Huang (EPM)]

(In reply to Thomas Zimmermann [:tzimmermann] [:tdz] from comment #33)
> Ken, Wesley,
> 
> we will need new base images for the Flame (master + v2.2) before enabling
> bluetoothd. Can you help with that?

Hi Thomas:

sorry for late, our additional Flame contract with T2M only covers OS upgrade to 2.0 (while the original one covers 1.3 for shipping), and the v18D is so far the last SW planned for this contract. And now the next SOW with them is for L BSP porting so still not covering 2.2, thus I'm afraid we need to find other way forward.

Comment 58

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Too bad. How do we get out of that...?

Comment 59

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Gabriele,

AFAICS the FOTA scripts only support content under /system, right? Would it be possible to extend them to cover other directories and partitions as well; especially the root directory?

Comment 60

[Gabriele Svelto [:gsvelto]]

(In reply to Thomas Zimmermann [:tzimmermann] [:tdz] from comment #59)
> AFAICS the FOTA scripts only support content under /system, right? Would it
> be possible to extend them to cover other directories and partitions as
> well; especially the root directory?

By default we mount /system and /data and we generally unpack stuff only under /system. To unpack stuff directly into / we'd probably need some changes.

Comment 61

[:gerard-majax]

(In reply to Gabriele Svelto [:gsvelto] from comment #60)
> (In reply to Thomas Zimmermann [:tzimmermann] [:tdz] from comment #59)
> > AFAICS the FOTA scripts only support content under /system, right? Would it
> > be possible to extend them to cover other directories and partitions as
> > well; especially the root directory?
> 
> By default we mount /system and /data and we generally unpack stuff only
> under /system. To unpack stuff directly into / we'd probably need some
> changes.

To complement, and as explained in comment 10, we can add partitions flashing this to the recovery package, but as Gabriele replied, this needs some work.

I'm not really sure we would want to modify the boot partition like this, by mounting it in recovery mode.

Comment 62

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Moving the discussion to bug 1136638

Comment 63

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Wesly Huang from comment #57)
> (In reply to Thomas Zimmermann [:tzimmermann] [:tdz] from comment #33)
> > Ken, Wesley,
> > 
> > we will need new base images for the Flame (master + v2.2) before enabling
> > bluetoothd. Can you help with that?
> 
> Hi Thomas:
> 
> sorry for late, our additional Flame contract with T2M only covers OS
> upgrade to 2.0 (while the original one covers 1.3 for shipping), and the
> v18D is so far the last SW planned for this contract. And now the next SOW
> with them is for L BSP porting so still not covering 2.2, thus I'm afraid we
> need to find other way forward.
Wesly,
I think Thomas is talking about 'V18D_nightly' images, which can allow us to do OTA.

Comment 64

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Yes, let me clarify.

AFAICT we'd need a new base image to distribute some changes in the init scripts. I can even build that locally. The question is probably more about legally distributing such a base image to the general public (i.e., dogfooders).

Comment 65

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Thomas Zimmermann [:tzimmermann] [:tdz] from comment #64)
> Yes, let me clarify.
> 
> AFAICT we'd need a new base image to distribute some changes in the init
> scripts. I can even build that locally. The question is probably more about
> legally distributing such a base image to the general public (i.e.,
> dogfooders).

AFAICT, v18D is for v2.0 and update via official T2M FOTA server to end-users.
V18D-nightly (naming is confused) is for dogfooders who wants to update catch up the latest software.
Wesly, can you share the link so we can test with the latest V18D-nightly packages?

Comment 66

[Wesly Huang (EPM)]

(In reply to Shawn Huang [:shawnjohnjr] from comment #63)
> (In reply to Wesly Huang from comment #57)
> > (In reply to Thomas Zimmermann [:tzimmermann] [:tdz] from comment #33)
> > > Ken, Wesley,
> > > 
> > > we will need new base images for the Flame (master + v2.2) before enabling
> > > bluetoothd. Can you help with that?
> > 
> > Hi Thomas:
> > 
> > sorry for late, our additional Flame contract with T2M only covers OS
> > upgrade to 2.0 (while the original one covers 1.3 for shipping), and the
> > v18D is so far the last SW planned for this contract. And now the next SOW
> > with them is for L BSP porting so still not covering 2.2, thus I'm afraid we
> > need to find other way forward.
> Wesly,
> I think Thomas is talking about 'V18D_nightly' images, which can allow us to
> do OTA.

I see, thanks Shawn. Like just talked, there is a new v18D_nighty under testing by QA Mike now, my understanding is BT should work in that build, will share it to Thomas and you offline to double check if BT works as expected, too.(In reply to Shawn Huang [:shawnjohnjr] from comment #65)
> (In reply to Thomas Zimmermann [:tzimmermann] [:tdz] from comment #64)
> > Yes, let me clarify.
> > 
> > AFAICT we'd need a new base image to distribute some changes in the init
> > scripts. I can even build that locally. The question is probably more about
> > legally distributing such a base image to the general public (i.e.,
> > dogfooders).
> 
> AFAICT, v18D is for v2.0 and update via official T2M FOTA server to
> end-users.
> V18D-nightly (naming is confused) is for dogfooders who wants to update
> catch up the latest software.
> Wesly, can you share the link so we can test with the latest V18D-nightly
> packages?

Thanks Shawn and yes, like we just talked the new v18D_nighty under testing by QA Mike now, I will share it to Thomas and you offline to double check if BT works as expected, too.

Comment 67

[Wesly Huang (EPM)]

sorry for some dup. msg in comment#66 due to my browser misbehavior, anyway the message is the same.

Comment 68

[bhavana bajaj [:bajaj]]

nhirata, can you confirm this is good to uplift or not ?

Comment 69

[bhavana bajaj [:bajaj]]

naoki is verifying this but does not see anything blocking us from uplift yet, so I am approving this and 1134821 to not lose on more branch testing.

Comment 70

[Ryan VanderMeulen [:RyanVM]]

I made my best stab at all the relevant uplifts for v2.2. Please correct me if I messed any of it up.

(In reply to Ryan VanderMeulen [:RyanVM UTC-4] from comment #48)
> Master:
> https://github.com/mozilla-b2g/B2G/commit/
> 42dbad3bf1a2c7f8247f7e20094d7b144a31a2e6

Doesn't need uplift (all versions use the same code).

> Master:
> https://github.com/mozilla-b2g/gonk-misc/commit/
> 3905180d9e2dbe03d7ce3b3a6debe4d4adc938d1

v2.2: https://github.com/mozilla-b2g/gonk-misc/commit/6c6daa88dc3fc30df7b70837c536d868ba87673e

> Master:
> https://github.com/mozilla-b2g/platform_build/commit/
> 173b3104bfcbd23fc9dccd4b0035fc49aae3d444

v2.2: https://github.com/mozilla-b2g/platform_build/commit/05bfa210e684e07b3c884a3e6784100656eeb87f

> b2g-4.3_r2.1:
> https://github.com/mozilla-b2g/platform_build/commit/
> 4efd19d199ae52656604f794c5a77518400220fd

v2.2 is pinned to this branch where needed.

> b2g-4.4.2_r1:
> https://github.com/mozilla-b2g/platform_build/commit/
> ef937d1aca7c4cf89ecb5cc43ae8c21c2000a9db

v2.2 is pinned to this branch where needed.

> b2g-5.0.0_r6:
> https://github.com/mozilla-b2g/platform_build/commit/
> 0ca5e1456dfd417c3e4e72bb8eae02069136b99b

v2.2 is pinned to this branch where needed.

Comment 71

[Thomas Zimmermann [:tzimmermann] [:tdz]]

The patches look good to me.

Comment 72

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

Eric, could you or someone on your team take a look at this please?  I noticed that you were handling Bug 1139331 which blocked me from fully testing this.

Comment 73

[Eric Chang [:ericcc] [:echang]]

FWD to Gerry

Comment 74

[Gerry Chang [:gchang]]

After trying V18D_nightly to perform FOTA, got the same problem as bug 1155035.