Configurable values for initrd_high and fdt_high

Bug #1055644 reported by Mathieu Poirier
18
This bug affects 2 people
Affects Status Importance Assigned to Milestone
Linaro Image Tools
Fix Released
Critical
Paul Sokolovsky

Bug Description

In 'linaro_image_tools/media_create/boards.py', line 1898, values for 'initrd_high' and 'fdt_high' are hard coded for every board. This used to have no effect while snowball wasn't booting with a DT, which is no longer the case.

Those values are too high and need to be made configurable, probably the same way 'initrd_addr' and 'dtb_addr' are set in 'AndroidSnowballSdConfig' and 'AndroidSnowballEmmcConfig'.

Related branches

Revision history for this message
Данило Шеган (danilo) wrote :

We are going to hit backwards compatibility problems: linaro-image-tools should be able to build images even for old Linaro releases.

Since we have just rolled out 2012.09 last week, I suppose this is not critical for a release next week. If it is, please come up with good arguments and find me on IRC :)

Anyway, please do let me know about when do you require to have this and if there are other things to consider.

Changed in linaro-image-tools:
status: New → Incomplete
Revision history for this message
Данило Шеган (danilo) wrote :

And please mark the bug as 'New' once you do that so I take a look at it again. Thanks.

Revision history for this message
Patrik Ryd (patrik-ryd) wrote :

I think we need this.

Patrik Ryd (patrik-ryd)
Changed in linaro-image-tools:
status: Incomplete → New
Revision history for this message
Alexander Sack (asac) wrote :

I thought we had a per-board way of setting such configs. Isn't that the case? If so, DT never worked on our snowball, right? If so, it wouldn't risk introducing a regression if you can still opt out of using DT at cmdline.

Is that correct?

Revision history for this message
Mathieu Poirier (mathieu.poirier-deactivatedaccount) wrote :

To answer #1, this is mandatory for 12.10 and onward. As for the backward compatibility issues, I dont' think it will be a problem since uboot in older images doesn't have the intelligence to pass the DTB to ther kernel.

As for #4, we never booted snowball with a DTB.

Revision history for this message
Patrik Ryd (patrik-ryd) wrote :

We really need this problem solved since the tracking snowball does not boot without it.

We have a workaround with a boot.src from Mathieu that we use instead of the one from LAMC (documented on https://android-build.linaro.org/builds/~linaro-android/snowball-jb-gcc47-linaro-tracking-open/). It works for now, but we really need this bug fixed before we can do a release of the tracking snowball build.

Revision history for this message
Paul Sokolovsky (pfalcon) wrote :

I don't see any problem with implementing this request - we allow those values to be set in hwpack config, and if they're not set, just use as defaults the values currently hard-coded in source. That's of course assuming these changes would be done for hwpack v3, which should have all the flexibility to do this.

Changed in linaro-image-tools:
importance: Undecided → Critical
status: New → Confirmed
Fathi Boudra (fboudra)
Changed in linaro-image-tools:
milestone: none → 2012.10
Revision history for this message
Paul Sokolovsky (pfalcon) wrote :

I stand corrected. As Android doesn't use hwpacks, this is indeed a bit tangled to resolve. How do we pass those params? On command line? There can be dozen of such params, so it doesn't scale. We exactly have hwpack config to capture such info, so proper fix would be to make Android use hwpacks too (even in if highly simplified format where hwpacks consists just of hwpack config file specifying values like in question).

Changed in linaro-image-tools:
assignee: nobody → Paul Sokolovsky (pfalcon)
Changed in linaro-image-tools:
status: Confirmed → In Progress
Revision history for this message
Paul Sokolovsky (pfalcon) wrote :

Ok, implementation based on Mathieu's suggestions is posted as
https://code.launchpad.net/~pfalcon/linaro-image-tools/snowball_dts_tweaks/+merge/129419

This bug however doesn't list specific values to use (nor text at https://android-build.linaro.org/builds/~linaro-android/snowball-jb-gcc47-linaro-tracking-open/ does), so please advice about them.

Revision history for this message
Paul Sokolovsky (pfalcon) wrote :

Updated MR above with actual values and revamped tests.

Milo Casagrande (milo)
Changed in linaro-image-tools:
status: In Progress → Fix Committed
Revision history for this message
Patrik Ryd (patrik-ryd) wrote :

Tested it on both the snowball tracking open build and on the snowball stable blob build. Both boots (at least to the point that we can say that the problem is solved).

Revision history for this message
Paul Sokolovsky (pfalcon) wrote :

Thanks, closing.

Changed in linaro-image-tools:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.