linux-kernel - Re: [PATCH v2 2/2] selftests/cgroup: add test for zswap incompressible pages

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <3d7e7e82-594c-4387-8dbd-2b78e888ead4@gmail.com>
Date: Sun, 8 Feb 2026 10:49:02 -0800
From: JP Kobryn <inwardvessel@...il.com>
To: SeongJae Park <sj@...nel.org>, Jiayuan Chen <jiayuan.chen@...ux.dev>
Cc: linux-mm@...ck.org, Jiayuan Chen <jiayuan.chen@...pee.com>,
 Tejun Heo <tj@...nel.org>, Johannes Weiner <hannes@...xchg.org>,
 Michal Koutný <mkoutny@...e.com>,
 Jonathan Corbet <corbet@....net>, Michal Hocko <mhocko@...nel.org>,
 Roman Gushchin <roman.gushchin@...ux.dev>,
 Shakeel Butt <shakeel.butt@...ux.dev>, Muchun Song <muchun.song@...ux.dev>,
 Andrew Morton <akpm@...ux-foundation.org>,
 Yosry Ahmed <yosry.ahmed@...ux.dev>, Nhat Pham <nphamcs@...il.com>,
 Chengming Zhou <chengming.zhou@...ux.dev>, Shuah Khan <shuah@...nel.org>,
 cgroups@...r.kernel.org, linux-doc@...r.kernel.org,
 linux-kernel@...r.kernel.org, linux-kselftest@...r.kernel.org
Subject: Re: [PATCH v2 2/2] selftests/cgroup: add test for zswap
 incompressible pages

On 2/6/26 5:35 PM, SeongJae Park wrote:
> On Fri,  6 Feb 2026 15:22:16 +0800 Jiayuan Chen <jiayuan.chen@...ux.dev> wrote:
> 
>> From: Jiayuan Chen <jiayuan.chen@...pee.com>
[...]
>> diff --git a/tools/testing/selftests/cgroup/test_zswap.c b/tools/testing/selftests/cgroup/test_zswap.c
>> index 64ebc3f3f203..8cb8a131357d 100644
>> --- a/tools/testing/selftests/cgroup/test_zswap.c
>> +++ b/tools/testing/selftests/cgroup/test_zswap.c
>> @@ -5,6 +5,7 @@
>>   #include <unistd.h>
>>   #include <stdio.h>
>>   #include <signal.h>
>> +#include <fcntl.h>
>>   #include <sys/sysinfo.h>
>>   #include <string.h>
>>   #include <sys/wait.h>
>> @@ -574,6 +575,100 @@ static int test_no_kmem_bypass(const char *root)
>>   	return ret;
>>   }
>>   
>> +static int allocate_random_and_wait(const char *cgroup, void *arg)
>> +{
>> +	size_t size = (size_t)arg;
>> +	char *mem;
>> +	int fd;
>> +	ssize_t n;
>> +
>> +	mem = malloc(size);
>> +	if (!mem)
>> +		return -1;
>> +
>> +	/* Fill with random data from /dev/urandom - incompressible */
>> +	fd = open("/dev/urandom", O_RDONLY);
>> +	if (fd < 0) {
>> +		free(mem);
>> +		return -1;
>> +	}
>> +
>> +	for (size_t i = 0; i < size; ) {
>> +		n = read(fd, mem + i, size - i);
>> +		if (n <= 0)
>> +			break;
>> +		i += n;
>> +	}
>> +	close(fd);
>> +
>> +	/* Touch all pages to ensure they're faulted in */
>> +	for (size_t i = 0; i < size; i += 4096)
> 
> Nit.  I show test_zswapin() is using PAGE_SIZE.  Maybe the above code can also
> use it?
> 
>> +		mem[i] = mem[i];
>> +
>> +	/* Keep memory alive for parent to reclaim and check stats */
>> +	pause();
>> +	free(mem);
>> +	return 0;
>> +}
>> +
>> +static long get_zswap_incomp(const char *cgroup)
>> +{
>> +	return cg_read_key_long(cgroup, "memory.stat", "zswap_incomp ");
>> +}
>> +
>> +/*
>> + * Test that incompressible pages (random data) are tracked by zswap_incomp.
>> + *
>> + * Since incompressible pages stored in zswap are charged at full PAGE_SIZE
>> + * (no memory savings), we cannot rely on memory.max pressure to push them
>> + * into zswap. Instead, we allocate random data within memory.max, then use
>> + * memory.reclaim to proactively push pages into zswap while checking the stat
>> + * before the child exits (zswap_incomp is a gauge that decreases on free).
>> + */
>> +static int test_zswap_incompressible(const char *root)
>> +{
>> +	int ret = KSFT_FAIL;
>> +	char *test_group;
>> +	long zswap_incomp;
>> +	pid_t child_pid;
>> +	int child_status;
>> +
>> +	test_group = cg_name(root, "zswap_incompressible_test");
>> +	if (!test_group)
>> +		goto out;
>> +	if (cg_create(test_group))
>> +		goto out;
>> +	if (cg_write(test_group, "memory.max", "32M"))
>> +		goto out;
>> +
>> +	child_pid = cg_run_nowait(test_group, allocate_random_and_wait,
>> +				  (void *)MB(4));
>> +	if (child_pid < 0)
>> +		goto out;
>> +
>> +	/* Wait for child to finish allocating */
>> +	usleep(500000);
> 
> We might be better to revisit here in future to avoid racy test results.  But
> this seems good enough for now.

How about using some form of synchronization like an eventfd? The parent
can wait here for the child to write the event and avoid the race with
the arbitrary sleep.