Re: [PATCH v2] selftests/mm: add simple VM_PFNMAP tests based on mmap'ing /dev/mem

[David Hildenbrand <david@redhat.com>]

On 28.05.25 12:34, Ryan Roberts wrote:
> Hi David,
> 
> 
> On 09/05/2025 16:30, David Hildenbrand wrote:
>> Let's test some basic functionality using /dev/mem. These tests will
>> implicitly cover some PAT (Page Attribute Handling) handling on x86.
>>
>> These tests will only run when /dev/mem access to the first two pages
>> in physical address space is possible and allowed; otherwise, the tests
>> are skipped.
> 
> We are seeing really horrible RAS errors with this test when run on arm64 tx2
> machine. Based solely on reviewing the code, I think the problem is that tx2
> doesn't have anything at phys address 0, so test_read_access() is trying to put
> trasactions out to a bad address on the bus.
> 
> tx2 /proc/iomem:
> 
> $ sudo cat /proc/iomem
> 30000000-37ffffff : PCI ECAM
> 38000000-3fffffff : PCI ECAM
> 40000000-5fffffff : PCI Bus 0000:00
> ...
> 
> Whereas my x86 box has some reserved memory:
> 
> $ sudo cat /proc/iomem
> 00000000-00000fff : Reserved
> 00001000-0003dfff : System RAM
> ...
> 

A quick fix would be to make this test specific to x86 (the only one I 
tested on). We should always have the lower two pages IIRC (BIOS stuff etc).

> I think perhaps the only safe way to handle this is to parse /proc/iomem for a
> region of "System RAM" that is at least 2 pages then use that for your read
> tests. This would also solve the hypothetical issue of reading something that
> has read size effects.

That sounds also plausible yes. I somehow remembered that mmap() would 
fail if "there is nothing".

> 
> I also spotted a few nits while reading the code...
> 
>>
>> On current x86-64 with PAT inside a VM, all tests pass:
>>
>> 	TAP version 13
>> 	1..6
>> 	# Starting 6 tests from 1 test cases.
>> 	#  RUN           pfnmap.madvise_disallowed ...
>> 	#            OK  pfnmap.madvise_disallowed
>> 	ok 1 pfnmap.madvise_disallowed
>> 	#  RUN           pfnmap.munmap_split ...
>> 	#            OK  pfnmap.munmap_split
>> 	ok 2 pfnmap.munmap_split
>> 	#  RUN           pfnmap.mremap_fixed ...
>> 	#            OK  pfnmap.mremap_fixed
>> 	ok 3 pfnmap.mremap_fixed
>> 	#  RUN           pfnmap.mremap_shrink ...
>> 	#            OK  pfnmap.mremap_shrink
>> 	ok 4 pfnmap.mremap_shrink
>> 	#  RUN           pfnmap.mremap_expand ...
>> 	#            OK  pfnmap.mremap_expand
>> 	ok 5 pfnmap.mremap_expand
>> 	#  RUN           pfnmap.fork ...
>> 	#            OK  pfnmap.fork
>> 	ok 6 pfnmap.fork
>> 	# PASSED: 6 / 6 tests passed.
>> 	# Totals: pass:6 fail:0 xfail:0 xpass:0 skip:0 error:0
>>
>> However, we are able to trigger:
>>
>> [   27.888251] x86/PAT: pfnmap:1790 freeing invalid memtype [mem 0x00000000-0x00000fff]
>>
 >> There are probably more things worth testing in the future, such 
as>> MAP_PRIVATE handling. But this set of tests is sufficient to cover 
most of
>> the things we will rework regarding PAT handling.
>>
>> Cc: Andrew Morton <akpm@linux-foundation.org>
>> Cc: Shuah Khan <shuah@kernel.org>
>> Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
>> Cc: Ingo Molnar <mingo@redhat.com>
>> Cc: Peter Xu <peterx@redhat.com>
>> Cc: Dev Jain <dev.jain@arm.com>
>> Signed-off-by: David Hildenbrand <david@redhat.com>
>> ---
>>
>> Hopefully I didn't miss any review feedback.
>>
>> v1 -> v2:
>> * Rewrite using kselftest_harness, which simplifies a lot of things
>> * Add to .gitignore and run_vmtests.sh
>> * Register signal handler on demand
>> * Use volatile trick to force a read (not factoring out FORCE_READ just yet)
>> * Drop mprotect() test case
>> * Add some more comments why we test certain things
>> * Use NULL for mmap() first parameter instead of 0
>> * Smaller fixes
>>
>> ---
>>   tools/testing/selftests/mm/.gitignore     |   1 +
>>   tools/testing/selftests/mm/Makefile       |   1 +
>>   tools/testing/selftests/mm/pfnmap.c       | 196 ++++++++++++++++++++++
>>   tools/testing/selftests/mm/run_vmtests.sh |   4 +
>>   4 files changed, 202 insertions(+)
>>   create mode 100644 tools/testing/selftests/mm/pfnmap.c
>>
>> diff --git a/tools/testing/selftests/mm/.gitignore b/tools/testing/selftests/mm/.gitignore
>> index 91db34941a143..824266982aa36 100644
>> --- a/tools/testing/selftests/mm/.gitignore
>> +++ b/tools/testing/selftests/mm/.gitignore
>> @@ -20,6 +20,7 @@ mremap_test
>>   on-fault-limit
>>   transhuge-stress
>>   pagemap_ioctl
>> +pfnmap
>>   *.tmp*
>>   protection_keys
>>   protection_keys_32
>> diff --git a/tools/testing/selftests/mm/Makefile b/tools/testing/selftests/mm/Makefile
>> index ad4d6043a60f0..ae6f994d3add7 100644
>> --- a/tools/testing/selftests/mm/Makefile
>> +++ b/tools/testing/selftests/mm/Makefile
>> @@ -84,6 +84,7 @@ TEST_GEN_FILES += mremap_test
>>   TEST_GEN_FILES += mseal_test
>>   TEST_GEN_FILES += on-fault-limit
>>   TEST_GEN_FILES += pagemap_ioctl
>> +TEST_GEN_FILES += pfnmap
>>   TEST_GEN_FILES += thuge-gen
>>   TEST_GEN_FILES += transhuge-stress
>>   TEST_GEN_FILES += uffd-stress
>> diff --git a/tools/testing/selftests/mm/pfnmap.c b/tools/testing/selftests/mm/pfnmap.c
>> new file mode 100644
>> index 0000000000000..8a9d19b6020c7
>> --- /dev/null
>> +++ b/tools/testing/selftests/mm/pfnmap.c
>> @@ -0,0 +1,196 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * Basic VM_PFNMAP tests relying on mmap() of '/dev/mem'
>> + *
>> + * Copyright 2025, Red Hat, Inc.
>> + *
>> + * Author(s): David Hildenbrand <david@redhat.com>
>> + */
>> +#define _GNU_SOURCE
>> +#include <stdlib.h>
>> +#include <string.h>
>> +#include <stdint.h>
>> +#include <unistd.h>
>> +#include <errno.h>
>> +#include <fcntl.h>
>> +#include <signal.h>
>> +#include <setjmp.h>
>> +#include <linux/mman.h>
>> +#include <sys/mman.h>
>> +#include <sys/wait.h>
>> +
>> +#include "../kselftest_harness.h"
>> +#include "vm_util.h"
>> +
>> +static sigjmp_buf sigjmp_buf_env;
>> +
>> +static void signal_handler(int sig)
>> +{
>> +	siglongjmp(sigjmp_buf_env, -EFAULT);
>> +}
>> +
>> +static int test_read_access(char *addr, size_t size, size_t pagesize)
>> +{
>> +	size_t offs;
>> +	int ret;
>> +
>> +	if (signal(SIGSEGV, signal_handler) == SIG_ERR)
>> +		return -EINVAL;
>> +
>> +	ret = sigsetjmp(sigjmp_buf_env, 1);
>> +	if (!ret) {
>> +		for (offs = 0; offs < size; offs += pagesize)
>> +			/* Force a read that the compiler cannot optimize out. */
>> +			*((volatile char *)(addr + offs));
>> +	}
>> +	if (signal(SIGSEGV, signal_handler) == SIG_ERR)
> 
> I think you mean:
> 	if (signal(SIGSEGV, SIG_DFL) == SIG_ERR)
> 
> Otherwise your signal_handler will remain installed, right?

Yeah, copy and past error.

> 
>> +		return -EINVAL;
>> +
>> +	return ret;
>> +}
>> +
>> +FIXTURE(pfnmap)
>> +{
>> +	size_t pagesize;
>> +	int dev_mem_fd;
>> +	char *addr1;
>> +	size_t size1;
>> +	char *addr2;
>> +	size_t size2;
>> +};
>> +
>> +FIXTURE_SETUP(pfnmap)
>> +{
>> +	self->pagesize = getpagesize();
>> +
>> +	self->dev_mem_fd = open("/dev/mem", O_RDONLY);
>> +	if (self->dev_mem_fd < 0)
>> +		SKIP(return, "Cannot open '/dev/mem'\n");
>> +
>> +	/* We'll require the first two pages throughout our tests ... */
>> +	self->size1 = self->pagesize * 2;
>> +	self->addr1 = mmap(NULL, self->size1, PROT_READ, MAP_SHARED,
>> +			   self->dev_mem_fd, 0);
>> +	if (self->addr1 == MAP_FAILED)
>> +		SKIP(return, "Cannot mmap '/dev/mem'\n");
>> +
>> +	/* ... and want to be able to read from them. */
>> +	if (test_read_access(self->addr1, self->size1, self->pagesize))
>> +		SKIP(return, "Cannot read-access mmap'ed '/dev/mem'\n");
>> +
>> +	self->size2 = 0;
>> +	self->addr2 = MAP_FAILED;
> 
> Do you need to init all the params in FIXTURE(pfnmap) to their "safe" values
> before any possible early returns? We don't want FIXTURE_TEARDOWN(pfnmap)
> getting confused.

If FIXTURE_SETUP fails, we'll exit immediately. See __TEST_F_IMPL().

Note that all tests are executed in a fork'ed child process, so quitting 
early (or not even having FIXTURE_TEARDOWN in most cases ...) does not 
really matter.

-- 
Cheers,

David / dhildenb