Possibly redundant myproc call in exec?

[LIParadise]

Here in kernel/exec.c we called myproc, which we've already done just now at this line.

AFAIK since the XV6's processes are managed by the global array struct proc proc[NPROC] in the kernel, that end_op, e.t.c., may sleep and thus change the CPU on which this call is carried out should not invalidate the struct proc*.

In fact the chapter 7.4 of the XV6 book rev.3 mentions that the pointer returned by myproc is safe to use even if some interrupt causes the process to be moved to other CPU, and the XV6 exec implementation on x86 does only call myproc once.

Are there any soundness concerns for the second call to myproc here?

Best Regards

[jrmenzel]

In my fork of xv6 I only query the process once: https://github.com/jrmenzel/vimix_os/blob/184435c55764d1f59421735b0542993f40db9c26/kernel/kernel/exec.c#L82
Never had any issues.