task_fork.c: Fix vfork for BUILD_KERNEL

[pussuw]

Summary

Two things need to be done when vfork'ing:

• Must attach to parent's address environment (the addrenv is shared)
• Must allocate a kernel stack (where would the register context go otherwise)

Note that this code assumes the address environment is shared, since we don't support fork() which would clone the address environment instead.

Impact

Fix vfork() for BUILD_KERNEL

Testing

rv-virt:knsh64 + ostest pass

[lupyuen]

Sorry @pussuw I think we might have a problem:

(1) knsh32 (on macOS) hangs at OSTest vfork: knsh32 Log / Build Script

$ git reset --hard 6047a9fe146ad4b4e8b2dc7bffe92d9075db429f
HEAD is now at 6047a9fe14 task_fork.c: Fix vfork for BUILD_KERNEL
...
NuttShell (NSH) NuttX-12.6.0-RC1
nsh> uname -a
NuttX 12.6.0-RC1 6047a9fe14 Aug  2 2024 09:41:15 risc-v rv-virt

nsh> ostest
user_main: vfork() test
[Hangs here]

(2) knsh64 (on macOS) loops forever at OSTest vfork with an Assertion Failure: knsh64 Log / Build Script

user_main: vfork() test
_assert: Current Version: NuttX  12.6.0-RC1 6047a9fe14 Aug  2 2024 09:37:38 risc-v
_assert: Assertion failed : at file:
_assert: Current Version: NuttX  12.6.0-RC1 6047a9fe14 Aug  2 2024 09:37:38 risc-v

Any idea why this happens? Thanks!

Update: I see that the vfork test has been recently enabled: apache/nuttx-apps#2457

(FYI: I discovered this through the Daily Regression Test for knsh32 and Milk-V Duo S)

[pussuw]

@lupyuen maybe there is a dependency to #12812

Hopefully I'll get it merged today. Sorry for the regression and the inconvenience.

[lupyuen]

@pussuw Cool no worries thanks! :-)