Re: kern/56535: Memory corruption in multi-threaded Go parent process following fork() on AMD CPUs

[Patrick Welche <prlw1%welche.eu@localhost>]

The following reply was made to PR kern/56535; it has been noted by GNATS.

From: Patrick Welche <prlw1%welche.eu@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: kern/56535: Memory corruption in multi-threaded Go parent
 process following fork() on AMD CPUs
Date: Sat, 19 Aug 2023 09:23:09 +0100

 Experimentally, this fixes the reproducer in the PR:
 
 cpu0: "AMD Ryzen 7 5700X 8-Core Processor             "
 cpu0: AMD Family 19h (686-class), 3393.62 MHz
 cpu0: family 0x19 model 0x21 stepping 0x2 (id 0xa20f12)
 
 Kernel with
 
    $NetBSD: uvm_fault.c,v 1.233 2023/07/17 12:55:37 riastradh Exp $
 
 $ ./a.out & ./a.out & ./a.out  & ./a.out
 Corruption! pb, page->b = 102, 2
 Corruption! pb, page->b = 102, 102
 Corruption! pb, page->b = 102, 2
 Loop 10000...
 Corruption! pb, page->b = 102, 2
 
 Update to
 
      $NetBSD: uvm_fault.c,v 1.234 2023/08/13 23:06:07 chs Exp $
 
 and:
 
 $ ./a.out & ./a.out & ./a.out & ./a.out
 Loop 10000...
 Loop 10000...
 Loop 10000...
 Loop 10000...
 Loop 20000...
 Loop 20000...
 Loop 20000...
 Loop 20000...
 ...
 Loop 1000000...