[LTP] [LTP PATCH v3 1/1] madvise11: Add test for memory allocation / Soft-offlining possible race

Richard Palethorpe rpalethorpe@suse.de
Mon Feb 27 11:16:04 CET 2023


Hello,

william.roche@oracle.com writes:

> From: William Roche <william.roche@oracle.com>
>
> Test a possible race condition between mmap() allocating memory and
> madvise() used to Soft-offline an unrelated memory page.
> As fixed with the following kernel commit:
> d4ae9916ea29 mm: soft-offline: close the race against page allocation
>
> If everything works, restore all poisoned pages created by this test.
>
> Reviewed-by: Richard Palethorpe <rpalethorpe@suse.de>
> Reviewed-by: Petr Vorel <pvorel@suse.cz>
> Signed-off-by: William Roche <william.roche@oracle.com>

Merged, thanks!

I think this test is well worth the effort.

This took an extra week to get merged because we were at a workshop.

> +
> +static int sigbus_received;

I added volatile in an extra patch to this. Usually I would recommend
tst_atomic* instead, but it is protected by a mutex anyway.

-- 
Thank you,
Richard.


More information about the ltp mailing list