I propose to implement PSA_Panic () as a macro (like the classic C assert feature) and not as a plain C function.
The benefit is that the macro allows you to insert also information like __FILE__ and __LINE__ which helps during the development phase of projects (i.e. when TF-M is incorrectly used by the non-secure side. __FILE__ and __LINE__ is useful also when no source code is available (at the debug stage) as it allows support to provide hints for the root cause of the issue.
You may have different variants of the PSA_Panic macro, i.e. a debug and release version.
Reinhard
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
Ken,
Today there are multiple different implementation of event logging used in the industry, but in a nutshell all use a similar concept.
You may start reading http://www.keil.com/pack/doc/compiler/EventRecorder/html/er_overview.html which is the implementation used currently in MDK. It uses a buffer that is read by the debugger. The buffer size is configurable and a faster debugger requires typically smaller buffer sizes. Also it introduces an XML file that describes the messages - and this description is used also by other tools like Percepio.
You may remap the annotation hooks also to classic printf.
So overall the concept gives you more flexibility and introducing it at a later stage into a project usually creates legacy effects.
For example we had in MDK middleware initially printf type annotations and introduced later events. As users are reluctant to change technology during the project lifecycle we end up to maintain both the old printf and the new event recorder annotations. This could have been avoided by using the right (flexible) concept from the beginning.
Reinhard
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
Hi,
Let's continue to discuss this topic.
I agree with this, the CI is just a tool, we should not only rely on it.
Thanks Gyorgy to point out the quality policy, we need to think about them when we discuss the version control.
In the current status, there is only one master branch for most of the development work in TF-M. Of course, we can add a warning to say that the master is constantly changing, there may be some problems in building or test running, and let the user use the release tag. But it is not convenience and impossible if they are upstreaming patches. And even to TF-M developers, we had met several times of the master branch is broken.
There is another way maybe help to solve this problem, we can define a merge policy like this: The patches cannot be merged to master branch if the patch is not based on the TOP HEAD. But this is very difficult to follow because there may be several maintainers could and need to merged patches at the same time. They need to align with each other to confirm their patches are on the TOP. But now, we often meet this case, while one patch rebases to TOP and waiting for the CI result, another patch is merged into master. The patch needs to rebase again.
I do not want to give more examples here.
Of course, it is a big change to involve a new branch, there must be many documents that need to be updated, and some policies need to be changed. And even this needs someone to maintain the alignment with 2 branches. But I think it is more useful and helpful for all users.
Thanks,
Edison
-----Original Message-----
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of Gyorgy Szing via TF-M
Sent: Friday, December 13, 2019 10:01 PM
To: Minos Galanakis <Minos.Galanakis(a)arm.com>; Soby Mathew <Soby.Mathew(a)arm.com>; tf-m(a)lists.trustedfirmware.org
Cc: nd <nd(a)arm.com>
Subject: Re: [TF-M] Create another branch for feature development
Hi,
Hi,
I agree, the CI shall not dictate how we use the version control system. It shall adapt.
Regarding your suggestions, I think the main problem is we are mixing stuff, this time quality with version control. Before we make decisions we shall understand where we are.
The current quality policy is that we only make releases for communication purposes. To give a clean interface for tf-m users and to allow planning their work. Releases allow them to execute their tf-m integration process less frequently. Only for each release or specific releases and not for each commit. The current quality policy identifies a single quality level only, and says any patch we publish is "golden quality", it matches the highest quality standard we can achieve (with sane constraints). Also to make our life easy we decided to use the master branch to hold these patches.
At the same time we use the master branch for development. Any change we make is made against master. This means each pull request and thus each review targets master. For review purposes the best is to have a chain of small modifications, otherwise the review content becomes too large to follow.
The TF-A "branching strategy" tries to address this issue by introducing an integration branch used for development. This allows master to be more release specific.
I suggest to take the following approach (details to be discussed):
- introduce more quality levels i.e.:
- none: content of a topic branch, or content pushed to review.
- bronze: content passed code review and patch specific testing.
- silver: content passed a more though daily testing.
- gold: a release. A pack of source-code, feature state document (release notes), reviewed documentation (user manual, reference manual), test evidence, documentation of test efforts to allow repeatability. The version control system can be used to store content, and to provide identification info (i.e. tagging), but most likely the release will need other kind of storage to be used (i.e. documentation).
- platina: reaching extra quality level trough passing PSA or some FUSA qualification. Or we may simply use extra release for this.
Naming the quality levels allows us to have a cleaner definition of what can be expected at a specific level (set of quality measures, i.e. static analysis, code review, test configuration). It would also allow us cleaner communication and to find how we use the version system for quality purposes.
I also expect this discussion to help defining how the version system is used for development purposes.
The current state works ok, but is a sort of naturally grown. We might have reached the point when more pragmatic approach may be needed.
/George
-----Original Message-----
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of Minos Galanakis via TF-M
Sent: 13 December 2019 12:23
To: Edison Ai (Arm Technology China) via TF-M <tf-m(a)lists.trustedfirmware.org>; Soby Mathew <Soby.Mathew(a)arm.com>
Cc: nd <nd(a)arm.com>
Subject: Re: [TF-M] Create another branch for feature development
Hi all.
My personal comments on this.
I would like to point out that the CI is a tool, not the core project. I do not believe we should be changing our development strategy based on what the tool is doing. We should instead adjust the tool to fit our requirements.
* No patches should be/ are merged to master when CI fails. If master breaks it should most commonly be because of something we are not testing for. Using an integration branch would not change that.
* As a developer I find it more convoluted to work with projects who use different integration strategies. The most common assumption in open source projects is that you have a master branch which is the bleeding edge, but can contain untested bugs, and the release immutable git tags for versioning. Using branch merges as versioning is a design for the pull request model which is not quite compatible with Gerrit.
* Most of the CI downtime has nothing to do with the merge strategy, they are more of a chicken and the egg philosophical problem. If your patch or branch introduced a change which affects the tests outputs, how will you test it if the CI expects the old output? An integration branch would not solve the merge freeze periods, would just affect a different branch from master.
* I believe feature branches are quite useful, since changes to master do not disrupt the development flow of a big change, and even though they will require some maintenance to re-sync before the final patch , it will be handled by an engineer who knows the feature, and full understands the regression vectors.
If I were to suggest some changes for stability purposes, I would start smaller:
* Update documentation to instruct users to check out from release tags, warning then that master is constantly changing.
* Adjust the CI to detect an Allow-CI flag from every branch. That way developers can test any patch from any feature branch. The logic for that is already present in the code, but requires Gerrit to be configured accordingly.
* Add an undo process. This would be the only case for an integration branch. All patches are merged to a temporary branch, after confirming they have passed testing individually. On the once per day nightly test, the group of different patches, will be tested against an extensive job, in models and hardware, and only if successful it will fast forward master to that state.
Regards
Minos
________________________________
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> on behalf of Edison Ai (Arm Technology China) via TF-M <tf-m(a)lists.trustedfirmware.org>
Sent: 13 December 2019 08:55
To: Soby Mathew <Soby.Mathew(a)arm.com>; 'tf-m(a)lists.trustedfirmware.org' <tf-m(a)lists.trustedfirmware.org>
Cc: nd <nd(a)arm.com>
Subject: Re: [TF-M] Create another branch for feature development
Hi Soby,
Thanks for your detail description.
> Integration is a temporary merge branch to merge several patches and run the CI against. Usually once CI passes, the master will be fast forwarded to integration within a day.
> This helps us to test integration of patches and detect any failure before master is updated. This means the master will pass CI at any given merge point.
I think it's a good method like this so that we can double confirm the "master" branch is stable.
And this also can fix one case even the CI can work normally: one patch is ready to merge, and it is not based on the latest HEAD, but there is no conflict. We can merge the patch directly and let gerrit do rebase by itself. But we cannot confirm the CI test can pass.
Any comment for this from others?
For multiple feature branches, I think we can stop to discuss about it now until we have some strong demands for it. It is indeed a big change for us now.
Thanks,
Edison
-----Original Message-----
From: Soby Mathew <Soby.Mathew(a)arm.com>
Sent: Friday, December 13, 2019 5:14 AM
To: Edison Ai (Arm Technology China) <Edison.Ai(a)arm.com>; 'tf-m(a)lists.trustedfirmware.org' <tf-m(a)lists.trustedfirmware.org>
Cc: nd <nd(a)arm.com>
Subject: Re: [TF-M] Create another branch for feature development
On 11/12/2019 09:05, Edison Ai (Arm Technology China) via TF-M wrote:
> Hi Gyorgy,
>
> Thanks to point it out. I agree with you that it will be better if we can align these two projects in this. I had a quick check the branches from TF-A: https://git.trustedfirmware.org/TF-A/trusted-firmware-a.git/.
> There are three branches in TF-A:
> - "integration" branch, should be used for new features.
> - "master" branch, which is behind of "integration" branch. But I am nor sure what is the strategy to update it.
Hi Edison,
Integration is a temporary merge branch to merge several patches and run the CI against. Usually once CI passes, the master will be fast forwarded to integration within a day.
This helps us to test integration of patches and detect any failure before master is updated. This means the master will pass CI at any given merge point.
> - "topics/epic_beta0_spmd", I thinks it should like a feature branch for big feature.
> @Soby Mathew Could you help to share more information about it? Thanks very much.
We usually do not have feature branches in TF-A. The topics/epic_beta0_spmd is a prototyping branch where we wanted to share code with collaborators outside TF-A. The patches on this branch are not visible in gerrit review and no patches in gerrit review will be merged to this branch. Once the prototyping is complete, then patches on this branch will be reworked and pushed to gerrit review and finally get merged to integration and this branch will be deleted.
Our experience have been, long running development branches are generally a maintenance overhead. Merging these development branches before a release may also be risky as some of the changes may have unknown interactions and may become difficult to resolve.
The "topic" in gerrit review is effectively a branch. For example, this
review:
https://review.trustedfirmware.org/#/q/topic:od/debugfs+(status:open+OR+sta…
is a set of patches on topic "od/debugfs" and can be treated as development branch. This branch is alive as long as patches are not merged.
We need to understand the motivations for the change. Broken CI is an argument but development branches will only exacerbate that problem since we don't know the stability of each of those branches. Also merge conflict will not reduce due to development branches. Its just delaying the merge conflict to a later point.
There may be other reasons, but generally it is better to merge sensible patches (+2ed) within a feature even before the feature is complete as it will reduce merge conflicts (we have to ensure testing/build coverage for the patch). These are my thoughts on this.
Best Regards
Soby Mathew
--
TF-M mailing list
TF-M(a)lists.trustedfirmware.org
https://lists.trustedfirmware.org/mailman/listinfo/tf-m
--
TF-M mailing list
TF-M(a)lists.trustedfirmware.org
https://lists.trustedfirmware.org/mailman/listinfo/tf-m
--
TF-M mailing list
TF-M(a)lists.trustedfirmware.org
https://lists.trustedfirmware.org/mailman/listinfo/tf-m
Hi Tamas,
IAR provides C libraries in various configurations, depending on
optimization levels, target types, needed features.
Normally the linker selects the appropriate libraries depending on
target, optimization, semihosting etc, but it can be told to not do
that, in case there is a need to provide non-standard libraries.
Cheers,
/Thomas
Den 2020-01-09 kl. 09:41, skrev Tamas Ban via TF-M:
>
> Hi Thomas,
>
> Just for my understanding:
>
> * Does IAR provide a C std. lib as part of IAR toolchain package?
> * How the std C lib linked to the image? Does user provide an
> explicit flag which std. C lib to linked to the image?
>
> Tamas
>
> *From:*TF-M <tf-m-bounces(a)lists.trustedfirmware.org> *On Behalf Of
> *Thomas Törnblom via TF-M
> *Sent:* 08 January 2020 12:45
> *To:* tf-m(a)lists.trustedfirmware.org
> *Subject:* Re: [TF-M] [Request For Comments] apply "-fno-builtin" as
> default compiler flags
>
> The IAR toolchain does not produce any special "builtin" calls and
> thus does not have any flag similar to "-fno-builtin".
>
> /Thomas
>
> Den 2020-01-08 kl. 03:53, skrev Ken Liu via TF-M:
>
> Hi,
>
> �
>
> As TF-M needs runtime APIs so we are creating the Secure Partition
> runtime library, code is ready but we have not forwarded all
> necessary runtime APIs to the version TF-M implemented, this was
> caused by the toolchain optimization for built-in APIs, such as:
>
> �
>
> - Forward printf(%s) to puts if there is only one string parameter.
>
> - ARMCLANG would forward memxxx API into an optimized variant.
>
> �
>
> With the '-fno-builtin' flags set in the toolchain, this
> optimization would be disabled so that user just implement the
> same name built-in to replace the toolchain version.
>
> �
>
> Please help to check these point before applying '-fno-builtin'
> and provide your feedback:
>
> �
>
> - Could toolchains out of ARMCLANG and GNUARM have a similar flag?
>
> - Would it affect your project setting and how does it affect?
>
> �
>
> Please help to feedback. I will keep this thread open for ~1 week
> and let's get a conclusion after this.
>
> �
>
> Thanks!
>
> �
>
> /Ken
>
>
>
> --
>
> *Thomas T�rnblom*, /Product Engineer/
> IAR Systems AB
> Box 23051, Strandbodgatan 1
> SE-750 23 Uppsala, SWEDEN
> Mobile: +46 76 180 17 80 Fax: +46 18 16 78 01
> E-mail: thomas.tornblom(a)iar.com
> <mailto:thomas.tornblom@iar.com>Website: www.iar.com <http://www.iar.com>
> Twitter: www.twitter.com/iarsystems <http://www.twitter.com/iarsystems>
>
> IMPORTANT NOTICE: The contents of this email and any attachments are
> confidential and may also be privileged. If you are not the intended
> recipient, please notify the sender immediately and do not disclose
> the contents to any other person, use it for any purpose, or store or
> copy the information in any medium. Thank you. IMPORTANT NOTICE: The
> contents of this email and any attachments are confidential and may
> also be privileged. If you are not the intended recipient, please
> notify the sender immediately and do not disclose the contents to any
> other person, use it for any purpose, or store or copy the information
> in any medium. Thank you.
>
--
*Thomas Törnblom*, /Product Engineer/
IAR Systems AB
Box 23051, Strandbodgatan 1
SE-750 23 Uppsala, SWEDEN
Mobile: +46 76 180 17 80 Fax: +46 18 16 78 01
E-mail: thomas.tornblom(a)iar.com <mailto:thomas.tornblom@iar.com>
Website: www.iar.com <http://www.iar.com>
Twitter: www.twitter.com/iarsystems <http://www.twitter.com/iarsystems>
Dear All,
The next session of the Technical Forum is planned in 2 weeks, Thursday, January 23rd at 7:00-8:00 UTC.
Please treat this email and an early invitation for agenda topic collection. Any questions, proposals, concerns are all valid points for our open discussion so do not hesitate to share it.
A big or complicated topics are worth to preliminary discussed over a mailing list.
Best regards,
Anton Komlev
Hi,
I assume the main purpose of isolation would be protect the code been seen by the AppRoT. Let's check with the FF author for detailed answers.
The building instructions now is just create separate libraries and finally combine them together - since vendors can create Secure Partitions, these modularized building can't be avoided.
/Ken
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of Reinhard Keil via TF-M
Sent: Thursday, January 9, 2020 4:00 PM
To: tf-m(a)lists.trustedfirmware.org
Subject: [TF-M] Code Protection between secure services
I suggest we review the requirement of code isolation on the secure side.
R/W data and R/O data should definitely be isolated, but code isolation has implications:
* Code cannot be share between services (i.e. no linker optimization to reduce memory footprint)
* Sharing library code
* Overall the build instructions of the system are more complicated
* Adding device specific driver code (i.e. to crypto) can become tricky
Reinhard
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
Hi Thomas,
Just for my understanding:
* Does IAR provide a C std. lib as part of IAR toolchain package?
* How the std C lib linked to the image? Does user provide an explicit flag which std. C lib to linked to the image?
Tamas
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of Thomas Törnblom via TF-M
Sent: 08 January 2020 12:45
To: tf-m(a)lists.trustedfirmware.org
Subject: Re: [TF-M] [Request For Comments] apply "-fno-builtin" as default compiler flags
The IAR toolchain does not produce any special "builtin" calls and thus does not have any flag similar to "-fno-builtin".
/Thomas
Den 2020-01-08 kl. 03:53, skrev Ken Liu via TF-M:
Hi,
�
As TF-M needs runtime APIs so we are creating the Secure Partition runtime library, code is ready but we have not forwarded all necessary runtime APIs to the version TF-M implemented, this was caused by the toolchain optimization for built-in APIs, such as:
�
- Forward printf(%s) to puts if there is only one string parameter.
- ARMCLANG would forward memxxx API into an optimized variant.
�
With the '-fno-builtin' flags set in the toolchain, this optimization would be disabled so that user just implement the same name built-in to replace the toolchain version.
�
Please help to check these point before applying '-fno-builtin' and provide your feedback:
�
- Could toolchains out of ARMCLANG and GNUARM have a similar flag?
- Would it affect your project setting and how does it affect?
�
Please help to feedback. I will keep this thread open for ~1 week and let's get a conclusion after this.
�
Thanks!
�
/Ken
--
Thomas T�rnblom, Product Engineer
IAR Systems AB
Box 23051, Strandbodgatan 1
SE-750 23 Uppsala, SWEDEN
Mobile: +46 76 180 17 80 Fax: +46 18 16 78 01
E-mail: thomas.tornblom(a)iar.com<mailto:thomas.tornblom@iar.com> Website: www.iar.com<http://www.iar.com>
Twitter: www.twitter.com/iarsystems<http://www.twitter.com/iarsystems>
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
Hi,
We have a prototype audit logging service and one of the purposes is to provide the event logging.
The reason we keep printf first is that it is so direct and some projects reference it, for example, booting or services. Eventually, we need can investigate the possibility of forwarding all printf into event logging.
One question is that there needs to be a buffer for collecting the events so it has limited capacity?
/Ken
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of Reinhard Keil via TF-M
Sent: Thursday, January 9, 2020 3:42 PM
To: tf-m(a)lists.trustedfirmware.org
Subject: [TF-M] The logging mechanism change in TF-M
I recommend instead of using printf with plain text strings, to change the concept to event annotations.
This is already implemented in several RTOS systems (i.e. FreeRTOS or RTX).
Event annotations give you more flexibility:
* can be mapped to memory buffer, printf output, or analysers like Event Recorder in MDK or Percepio
* annotations are described in XML file using event groups. Can be mapped to test automation systems and reduce overhead
* no direct mapping to a peripheral. UART may be not available on all systems.
* overall less overhead in the system (as printf is expensive). Can remain in production code when it just goes to memory buffer
Let me know if you want to have further pointers or information.
Reinhard
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
I suggest we review the requirement of code isolation on the secure side.
R/W data and R/O data should definitely be isolated, but code isolation has implications:
* Code cannot be share between services (i.e. no linker optimization to reduce memory footprint)
* Sharing library code
* Overall the build instructions of the system are more complicated
* Adding device specific driver code (i.e. to crypto) can become tricky
Reinhard
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
I recommend instead of using printf with plain text strings, to change the concept to event annotations.
This is already implemented in several RTOS systems (i.e. FreeRTOS or RTX).
Event annotations give you more flexibility:
* can be mapped to memory buffer, printf output, or analysers like Event Recorder in MDK or Percepio
* annotations are described in XML file using event groups. Can be mapped to test automation systems and reduce overhead
* no direct mapping to a peripheral. UART may be not available on all systems.
* overall less overhead in the system (as printf is expensive). Can remain in production code when it just goes to memory buffer
Let me know if you want to have further pointers or information.
Reinhard
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
