STM32F207ZGTx in tools/arm_pack_manager/index.json not populated correctly

[JanneKiiskila]

Description of defect

See originating PR; #11707

The Keil CMSIS-pack file has this for the

        <!-- *************************  Device 'STM32F207ZG'  ***************************** -->
        <device Dname="STM32F207ZG">
          <memory id="IROM1"                               start="0x08000000" size="0x00100000" startup="1" default="1"/>
          <memory id="IRAM1"                               start="0x20000000" size="0x00020000" init   ="0" default="1"/>
          <algorithm name="CMSIS/Flash/STM32F2xx_1024.FLM" start="0x08000000" size="0x00100000"             default="1"/>
          <algorithm name="CMSIS/Flash/STM32F2xx_OPT.FLM"  start="0x1FFFC000" size="0x00000010"             default="0"/>
          <algorithm name="CMSIS/Flash/STM32F2xx_OTP.FLM"  start="0x1FFF7800" size="0x00000210"             default="0"/>
          <feature type="ADC"           n="24"      m="12"/>
          <feature type="IOs"           n="114"/>
          <variant Dvariant="STM32F207ZGTx"> <feature type="QFP" n="144"/> </variant>
        </device>

The current index.json under tools/arm_pack_manager has this information:

        "name": "STM32F207ZGTx", 
        "processor": {
            "Symmetric": {
                "core": "CortexM3", 
                "fpu": "None", 
                "mpu": "Present", 
                "units": 1
            }
        }, 
        "sectors": [
            [
                134217728, 
                16384
            ], 
            [
                134283264, 
                65536
            ], 
            [
                134348800, 
                131072
            ], 
            [
                536836096, 
                528
            ], 
            [
                536854528, 
                16
            ]
        ], 

So, the information seems to be essentially there, but it is not parsed out somehow / generated to the output.

Output is:

    "name": "STM32F207ZGTx", 
    "processor": {
        "Symmetric": {
            "core": "CortexM3", 
            "fpu": "None", 
            "mpu": "Present", 
            "units": 1
        }
    }, 
    "sectors": null, 

This causes failures in PR checks with "STM32F207ZG sectors not written out (sectors : null)".

@theotherjimmy analyzed root cause to be that CPM wrapper in Mbed OS is not populating the fields correctly (see pyocd/cmsis-pack-manager#126).

Target(s) affected by this defect ?

STM32F207ZGTx

Toolchain(s) (name and version) displaying this defect ?

mbed cli v1.10.1

What version of Mbed-os are you using (tag or sha) ?

5.14.0

What version(s) of tools are you using. List all that apply (E.g. mbed-cli)

How is this defect reproduced ?

Clone some repo, which pulls in mbed-os, too.

cd mbed-os
python project.py -m STM32F7xx --update-packs

analyze the produced index.json file in tools/arm_pack_manager.

[JanneKiiskila]

@ARMmbed/mbed-os-tools

[mark-edgeworth]

What should the sectors entry look like for the example above?

[JanneKiiskila]

The ones we have in the currently merged index.json in Mbed OS master (and previous release branches) has it correctly. Well, the issue itself also says what it is currently.

Should be this;

"sectors": [
            [
                134217728, 
                16384
            ], 
            [
                134283264, 
                65536
            ], 
            [
                134348800, 
                131072
            ], 
            [
                536836096, 
                528
            ], 
            [
                536854528, 
                16
            ]
        ], 

but we get ´sectors:** null´.

[ciarmcom]

Internal Jira reference: https://jira.arm.com/browse/IOTOSM-1463

[JanneKiiskila]

We have now the same problems also with STM32WB and STM32L4 series.

[jeromecoutant]

We got now 4 referenced issues waiting for this one...

@ARMmbed/mbed-os-tools please have a look

@MarceloSalazar
@bulislaw

[bulislaw]

@jeromecoutant it was escalated to a blocker internally, @ARMmbed/mbed-os-tools should look at it asap.

[0xc0170]

@jeromecoutant it was escalated to a blocker internally, @ARMmbed/mbed-os-tools should look at it asap.

Any update with this one?

[jeromecoutant]

@bulislaw

it was escalated to a blocker internally, @ARMmbed/mbed-os-tools should look at it asap.

Any update?

[JanneKiiskila]

Seems to me that most of CMSIS-pack files do not even have the required information to populate it necessarily (at least if the sector info should be the flash erase sector size information).

[bulislaw]

@ARMmbed/mbed-os-tools any updates on this?

[ciarmcom]

We closed this issue because it has been inactive for quite some time and we believe it to be low priority. If you think that the priority should be higher, then please reopen with your justification for increasing the priority.