storage: fix potential memory corruption and check return values

[michalpasztamobica]

Description

The change in l. 2484 of lfs.c is not really a fix, as we can't see anything wrong with that line, but it will silence the Coverity tool warning (thanks, @JuhPuur ).

The two other changes are true code improvements.

Pull request type

[x] Fix
[ ] Refactor
[ ] Target update
[ ] Functionality change
[ ] Docs update
[ ] Test update
[ ] Breaking change

Reviewers

@SeppoTakalo
@VeijoPesonen
@geky

Release Notes

[ciarmcom]

@michalpasztamobica, thank you for your changes.
@VeijoPesonen @SeppoTakalo @geky @ARMmbed/mbed-os-storage @ARMmbed/mbed-os-maintainers please review.

[kjbracey]

What was the coverity warning for 2484?

[kjbracey]

Style barf (backwards and silly boolean comparison instead of just !), but I guess it's consistent with other code nearby.

[michalpasztamobica]

Coverity has this nice section which I used to just keep the code consistent:

Examples where return value from this function is checked
--
A1 example_checked | SPIFBlockDevice.cpp:982 
B1 example_checked | SPIFBlockDevice.cpp:182 
C1 example_checked | SPIFBlockDevice.cpp:410 
D1 example_checked | SPIFBlockDevice.cpp:324

[michalpasztamobica]

What was the coverity warning for 2484?

Link to original warning:

incr: Incrementing cwd@dim2. The value of cwd@dim2 is now 1.
incr: Incrementing cwd@dim2. The value of cwd@dim2 is now 2.
const: At condition cwd@dim2 < 2U, the value of cwd@dim2 must be equal to 2.
dead_error_condition: The condition cwd@dim2 < 2U cannot be true.    	
CID 267716 (#1 of 1): Logically dead code (DEADCODE)
dead_error_line: Execution cannot reach this statement: cwd.d.tail[cwd@dim2] = 0UL;.

[kjbracey]

Hmm, that's just nonsense, isn't it? It's generating a warning about some sort of comparison it's generated internally while inspecting an initialiser. Weird. Still, modified version is neater.

[michalpasztamobica]

@kjbracey-arm , I agree and I also consulted three other colleagues who had the same opinion. @JuhPuur came up with this work around we have under review here.

I think it would make sense to report this to Coverity developers for further clarification. Maybe we are missing something here? @OPpuolitaival, would you advise us if this reasonable and how to go about this?

[geky]

Change is a noop for littlefs 👍

I couldn't make heads or tails of the Coverity warning. The modified version is neater but can only be used in initialization blocks. There's a lot of pair updates in the code where you can use this shortcut which is why pair[0]=1; pair[1]=1; is used often.

[0xc0170]

Ci started meanwhile reviews needed (storage team) (I triggered 2 jobs, not certain how but one I aborted just now)

[mbed-ci]

Test run: SUCCESS

Summary: 11 of 11 test jobs passed
Build number : 2
Build artifacts