tools/testing/selftests: assert that anon merge cases behave as expected

Prior to the recently applied commit that permits this merge,
mprotect()'ing a faulted VMA, adjacent to an unfaulted VMA, such that the
two share characteristics would fail to merge due to what appear to be
unintended consequences of commit 965f55dea0 ("mmap: avoid merging
cloned VMAs").

Now we have fixed this bug, assert that we can indeed merge anonymous VMAs
this way.

Also assert that forked source/target VMAs are equally rejected. 
Previously, all empty target anon merges with one VMA faulted and the
other unfaulted would be rejected incorrectly, now we ensure that unforked
merge, but forked do not.

Additionally, add the new test file to the MEMORY MAPPING section in
MAINTAINERS, as these tests are explicitly memory mapping related.

Link: https://lkml.kernel.org/r/2b69330274a3b71721f7042c5eabe91143934415.1744104124.git.lorenzo.stoakes@oracle.com
Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Reviewed-by: Yeoreum Yun <yeoreum.yun@arm.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Rik van Riel <riel@surriel.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wei Yang <richard.weiyang@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

MAINTAINERS

@@ -15670,6 +15670,7 @@ F:	mm/mseal.c
 F:	mm/vma.c
 F:	mm/vma.h
 F:	mm/vma_internal.h
+F:	tools/testing/selftests/mm/merge.c
 F:	tools/testing/vma/
 
 MEMORY MAPPING - LOCKING

tools/testing/selftests/mm/.gitignore

@@ -58,3 +58,4 @@ hugetlb_dio
 pkey_sighandler_tests_32
 pkey_sighandler_tests_64
 guard-regions
+merge

tools/testing/selftests/mm/Makefile

@@ -98,6 +98,7 @@ TEST_GEN_FILES += hugetlb_madv_vs_map
 TEST_GEN_FILES += hugetlb_dio
 TEST_GEN_FILES += droppable
 TEST_GEN_FILES += guard-regions
+TEST_GEN_FILES += merge
 
 ifneq ($(ARCH),arm64)
 TEST_GEN_FILES += soft-dirty

tools/testing/selftests/mm/merge.c

@@ -0,0 +1,455 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#define _GNU_SOURCE
+#include "../kselftest_harness.h"
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <sys/mman.h>
+#include <sys/wait.h>
+#include "vm_util.h"
+
+FIXTURE(merge)
+{
+	unsigned int page_size;
+	char *carveout;
+	struct procmap_fd procmap;
+};
+
+FIXTURE_SETUP(merge)
+{
+	self->page_size = psize();
+	/* Carve out PROT_NONE region to map over. */
+	self->carveout = mmap(NULL, 12 * self->page_size, PROT_NONE,
+			      MAP_ANON | MAP_PRIVATE, -1, 0);
+	ASSERT_NE(self->carveout, MAP_FAILED);
+	/* Setup PROCMAP_QUERY interface. */
+	ASSERT_EQ(open_self_procmap(&self->procmap), 0);
+}
+
+FIXTURE_TEARDOWN(merge)
+{
+	ASSERT_EQ(munmap(self->carveout, 12 * self->page_size), 0);
+	ASSERT_EQ(close_procmap(&self->procmap), 0);
+}
+
+TEST_F(merge, mprotect_unfaulted_left)
+{
+	unsigned int page_size = self->page_size;
+	char *carveout = self->carveout;
+	struct procmap_fd *procmap = &self->procmap;
+	char *ptr;
+
+	/*
+	 * Map 10 pages of R/W memory within. MAP_NORESERVE so we don't hit
+	 * merge failure due to lack of VM_ACCOUNT flag by mistake.
+	 *
+	 * |-----------------------|
+	 * |       unfaulted       |
+	 * |-----------------------|
+	 */
+	ptr = mmap(&carveout[page_size], 10 * page_size, PROT_READ | PROT_WRITE,
+		   MAP_ANON | MAP_PRIVATE | MAP_FIXED | MAP_NORESERVE, -1, 0);
+	ASSERT_NE(ptr, MAP_FAILED);
+	/*
+	 * Now make the first 5 pages read-only, splitting the VMA:
+	 *
+	 *      RO          RW
+	 * |-----------|-----------|
+	 * | unfaulted | unfaulted |
+	 * |-----------|-----------|
+	 */
+	ASSERT_EQ(mprotect(ptr, 5 * page_size, PROT_READ), 0);
+	/*
+	 * Fault in the first of the last 5 pages so it gets an anon_vma and
+	 * thus the whole VMA becomes 'faulted':
+	 *
+	 *      RO          RW
+	 * |-----------|-----------|
+	 * | unfaulted |  faulted  |
+	 * |-----------|-----------|
+	 */
+	ptr[5 * page_size] = 'x';
+	/*
+	 * Now mprotect() the RW region read-only, we should merge (though for
+	 * ~15 years we did not! :):
+	 *
+	 *             RO
+	 * |-----------------------|
+	 * |        faulted        |
+	 * |-----------------------|
+	 */
+	ASSERT_EQ(mprotect(&ptr[5 * page_size], 5 * page_size, PROT_READ), 0);
+
+	/* Assert that the merge succeeded using PROCMAP_QUERY. */
+	ASSERT_TRUE(find_vma_procmap(procmap, ptr));
+	ASSERT_EQ(procmap->query.vma_start, (unsigned long)ptr);
+	ASSERT_EQ(procmap->query.vma_end, (unsigned long)ptr + 10 * page_size);
+}
+
+TEST_F(merge, mprotect_unfaulted_right)
+{
+	unsigned int page_size = self->page_size;
+	char *carveout = self->carveout;
+	struct procmap_fd *procmap = &self->procmap;
+	char *ptr;
+
+	/*
+	 * |-----------------------|
+	 * |       unfaulted       |
+	 * |-----------------------|
+	 */
+	ptr = mmap(&carveout[page_size], 10 * page_size, PROT_READ | PROT_WRITE,
+		   MAP_ANON | MAP_PRIVATE | MAP_FIXED | MAP_NORESERVE, -1, 0);
+	ASSERT_NE(ptr, MAP_FAILED);
+	/*
+	 * Now make the last 5 pages read-only, splitting the VMA:
+	 *
+	 *      RW          RO
+	 * |-----------|-----------|
+	 * | unfaulted | unfaulted |
+	 * |-----------|-----------|
+	 */
+	ASSERT_EQ(mprotect(&ptr[5 * page_size], 5 * page_size, PROT_READ), 0);
+	/*
+	 * Fault in the first of the first 5 pages so it gets an anon_vma and
+	 * thus the whole VMA becomes 'faulted':
+	 *
+	 *      RW          RO
+	 * |-----------|-----------|
+	 * |  faulted  | unfaulted |
+	 * |-----------|-----------|
+	 */
+	ptr[0] = 'x';
+	/*
+	 * Now mprotect() the RW region read-only, we should merge:
+	 *
+	 *             RO
+	 * |-----------------------|
+	 * |        faulted        |
+	 * |-----------------------|
+	 */
+	ASSERT_EQ(mprotect(ptr, 5 * page_size, PROT_READ), 0);
+
+	/* Assert that the merge succeeded using PROCMAP_QUERY. */
+	ASSERT_TRUE(find_vma_procmap(procmap, ptr));
+	ASSERT_EQ(procmap->query.vma_start, (unsigned long)ptr);
+	ASSERT_EQ(procmap->query.vma_end, (unsigned long)ptr + 10 * page_size);
+}
+
+TEST_F(merge, mprotect_unfaulted_both)
+{
+	unsigned int page_size = self->page_size;
+	char *carveout = self->carveout;
+	struct procmap_fd *procmap = &self->procmap;
+	char *ptr;
+
+	/*
+	 * |-----------------------|
+	 * |       unfaulted       |
+	 * |-----------------------|
+	 */
+	ptr = mmap(&carveout[2 * page_size], 9 * page_size, PROT_READ | PROT_WRITE,
+		   MAP_ANON | MAP_PRIVATE | MAP_FIXED | MAP_NORESERVE, -1, 0);
+	ASSERT_NE(ptr, MAP_FAILED);
+	/*
+	 * Now make the first and last 3 pages read-only, splitting the VMA:
+	 *
+	 *      RO          RW          RO
+	 * |-----------|-----------|-----------|
+	 * | unfaulted | unfaulted | unfaulted |
+	 * |-----------|-----------|-----------|
+	 */
+	ASSERT_EQ(mprotect(ptr, 3 * page_size, PROT_READ), 0);
+	ASSERT_EQ(mprotect(&ptr[6 * page_size], 3 * page_size, PROT_READ), 0);
+	/*
+	 * Fault in the first of the middle 3 pages so it gets an anon_vma and
+	 * thus the whole VMA becomes 'faulted':
+	 *
+	 *      RO          RW          RO
+	 * |-----------|-----------|-----------|
+	 * | unfaulted |  faulted  | unfaulted |
+	 * |-----------|-----------|-----------|
+	 */
+	ptr[3 * page_size] = 'x';
+	/*
+	 * Now mprotect() the RW region read-only, we should merge:
+	 *
+	 *             RO
+	 * |-----------------------|
+	 * |        faulted        |
+	 * |-----------------------|
+	 */
+	ASSERT_EQ(mprotect(&ptr[3 * page_size], 3 * page_size, PROT_READ), 0);
+
+	/* Assert that the merge succeeded using PROCMAP_QUERY. */
+	ASSERT_TRUE(find_vma_procmap(procmap, ptr));
+	ASSERT_EQ(procmap->query.vma_start, (unsigned long)ptr);
+	ASSERT_EQ(procmap->query.vma_end, (unsigned long)ptr + 9 * page_size);
+}
+
+TEST_F(merge, mprotect_faulted_left_unfaulted_right)
+{
+	unsigned int page_size = self->page_size;
+	char *carveout = self->carveout;
+	struct procmap_fd *procmap = &self->procmap;
+	char *ptr;
+
+	/*
+	 * |-----------------------|
+	 * |       unfaulted       |
+	 * |-----------------------|
+	 */
+	ptr = mmap(&carveout[2 * page_size], 9 * page_size, PROT_READ | PROT_WRITE,
+		   MAP_ANON | MAP_PRIVATE | MAP_FIXED | MAP_NORESERVE, -1, 0);
+	ASSERT_NE(ptr, MAP_FAILED);
+	/*
+	 * Now make the last 3 pages read-only, splitting the VMA:
+	 *
+	 *             RW               RO
+	 * |-----------------------|-----------|
+	 * |       unfaulted       | unfaulted |
+	 * |-----------------------|-----------|
+	 */
+	ASSERT_EQ(mprotect(&ptr[6 * page_size], 3 * page_size, PROT_READ), 0);
+	/*
+	 * Fault in the first of the first 6 pages so it gets an anon_vma and
+	 * thus the whole VMA becomes 'faulted':
+	 *
+	 *             RW               RO
+	 * |-----------------------|-----------|
+	 * |       unfaulted       | unfaulted |
+	 * |-----------------------|-----------|
+	 */
+	ptr[0] = 'x';
+	/*
+	 * Now make the first 3 pages read-only, splitting the VMA:
+	 *
+	 *      RO          RW          RO
+	 * |-----------|-----------|-----------|
+	 * |  faulted  |  faulted  | unfaulted |
+	 * |-----------|-----------|-----------|
+	 */
+	ASSERT_EQ(mprotect(ptr, 3 * page_size, PROT_READ), 0);
+	/*
+	 * Now mprotect() the RW region read-only, we should merge:
+	 *
+	 *             RO
+	 * |-----------------------|
+	 * |        faulted        |
+	 * |-----------------------|
+	 */
+	ASSERT_EQ(mprotect(&ptr[3 * page_size], 3 * page_size, PROT_READ), 0);
+
+	/* Assert that the merge succeeded using PROCMAP_QUERY. */
+	ASSERT_TRUE(find_vma_procmap(procmap, ptr));
+	ASSERT_EQ(procmap->query.vma_start, (unsigned long)ptr);
+	ASSERT_EQ(procmap->query.vma_end, (unsigned long)ptr + 9 * page_size);
+}
+
+TEST_F(merge, mprotect_unfaulted_left_faulted_right)
+{
+	unsigned int page_size = self->page_size;
+	char *carveout = self->carveout;
+	struct procmap_fd *procmap = &self->procmap;
+	char *ptr;
+
+	/*
+	 * |-----------------------|
+	 * |       unfaulted       |
+	 * |-----------------------|
+	 */
+	ptr = mmap(&carveout[2 * page_size], 9 * page_size, PROT_READ | PROT_WRITE,
+		   MAP_ANON | MAP_PRIVATE | MAP_FIXED | MAP_NORESERVE, -1, 0);
+	ASSERT_NE(ptr, MAP_FAILED);
+	/*
+	 * Now make the first 3 pages read-only, splitting the VMA:
+	 *
+	 *      RO                RW
+	 * |-----------|-----------------------|
+	 * | unfaulted |       unfaulted       |
+	 * |-----------|-----------------------|
+	 */
+	ASSERT_EQ(mprotect(ptr, 3 * page_size, PROT_READ), 0);
+	/*
+	 * Fault in the first of the last 6 pages so it gets an anon_vma and
+	 * thus the whole VMA becomes 'faulted':
+	 *
+	 *      RO                RW
+	 * |-----------|-----------------------|
+	 * | unfaulted |        faulted        |
+	 * |-----------|-----------------------|
+	 */
+	ptr[3 * page_size] = 'x';
+	/*
+	 * Now make the last 3 pages read-only, splitting the VMA:
+	 *
+	 *      RO          RW          RO
+	 * |-----------|-----------|-----------|
+	 * | unfaulted |  faulted  |  faulted  |
+	 * |-----------|-----------|-----------|
+	 */
+	ASSERT_EQ(mprotect(&ptr[6 * page_size], 3 * page_size, PROT_READ), 0);
+	/*
+	 * Now mprotect() the RW region read-only, we should merge:
+	 *
+	 *             RO
+	 * |-----------------------|
+	 * |        faulted        |
+	 * |-----------------------|
+	 */
+	ASSERT_EQ(mprotect(&ptr[3 * page_size], 3 * page_size, PROT_READ), 0);
+
+	/* Assert that the merge succeeded using PROCMAP_QUERY. */
+	ASSERT_TRUE(find_vma_procmap(procmap, ptr));
+	ASSERT_EQ(procmap->query.vma_start, (unsigned long)ptr);
+	ASSERT_EQ(procmap->query.vma_end, (unsigned long)ptr + 9 * page_size);
+}
+
+TEST_F(merge, forked_target_vma)
+{
+	unsigned int page_size = self->page_size;
+	char *carveout = self->carveout;
+	struct procmap_fd *procmap = &self->procmap;
+	pid_t pid;
+	char *ptr, *ptr2;
+	int i;
+
+	/*
+	 * |-----------|
+	 * | unfaulted |
+	 * |-----------|
+	 */
+	ptr = mmap(&carveout[page_size], 5 * page_size, PROT_READ | PROT_WRITE,
+		   MAP_ANON | MAP_PRIVATE | MAP_FIXED, -1, 0);
+	ASSERT_NE(ptr, MAP_FAILED);
+
+	/*
+	 * Fault in process.
+	 *
+	 * |-----------|
+	 * |  faulted  |
+	 * |-----------|
+	 */
+	ptr[0] = 'x';
+
+	pid = fork();
+	ASSERT_NE(pid, -1);
+
+	if (pid != 0) {
+		wait(NULL);
+		return;
+	}
+
+	/* Child process below: */
+
+	/* Reopen for child. */
+	ASSERT_EQ(close_procmap(&self->procmap), 0);
+	ASSERT_EQ(open_self_procmap(&self->procmap), 0);
+
+	/* unCOWing everything does not cause the AVC to go away. */
+	for (i = 0; i < 5 * page_size; i += page_size)
+		ptr[i] = 'x';
+
+	/*
+	 * Map in adjacent VMA in child.
+	 *
+	 *     forked
+	 * |-----------|-----------|
+	 * |  faulted  | unfaulted |
+	 * |-----------|-----------|
+	 *      ptr         ptr2
+	 */
+	ptr2 = mmap(&ptr[5 * page_size], 5 * page_size, PROT_READ | PROT_WRITE,
+		   MAP_ANON | MAP_PRIVATE | MAP_FIXED, -1, 0);
+	ASSERT_NE(ptr2, MAP_FAILED);
+
+	/* Make sure not merged. */
+	ASSERT_TRUE(find_vma_procmap(procmap, ptr));
+	ASSERT_EQ(procmap->query.vma_start, (unsigned long)ptr);
+	ASSERT_EQ(procmap->query.vma_end, (unsigned long)ptr + 5 * page_size);
+}
+
+TEST_F(merge, forked_source_vma)
+{
+	unsigned int page_size = self->page_size;
+	char *carveout = self->carveout;
+	struct procmap_fd *procmap = &self->procmap;
+	pid_t pid;
+	char *ptr, *ptr2;
+	int i;
+
+	/*
+	 * |-----------|------------|
+	 * | unfaulted | <unmapped> |
+	 * |-----------|------------|
+	 */
+	ptr = mmap(&carveout[page_size], 5 * page_size, PROT_READ | PROT_WRITE,
+		   MAP_ANON | MAP_PRIVATE | MAP_FIXED | MAP_NORESERVE, -1, 0);
+	ASSERT_NE(ptr, MAP_FAILED);
+
+	/*
+	 * Fault in process.
+	 *
+	 * |-----------|------------|
+	 * |  faulted  | <unmapped> |
+	 * |-----------|------------|
+	 */
+	ptr[0] = 'x';
+
+	pid = fork();
+	ASSERT_NE(pid, -1);
+
+	if (pid != 0) {
+		wait(NULL);
+		return;
+	}
+
+	/* Child process below: */
+
+	/* Reopen for child. */
+	ASSERT_EQ(close_procmap(&self->procmap), 0);
+	ASSERT_EQ(open_self_procmap(&self->procmap), 0);
+
+	/* unCOWing everything does not cause the AVC to go away. */
+	for (i = 0; i < 5 * page_size; i += page_size)
+		ptr[i] = 'x';
+
+	/*
+	 * Map in adjacent VMA in child, ptr2 after ptr, but incompatible.
+	 *
+	 *   forked RW      RWX
+	 * |-----------|-----------|
+	 * |  faulted  | unfaulted |
+	 * |-----------|-----------|
+	 *      ptr        ptr2
+	 */
+	ptr2 = mmap(&carveout[6 * page_size], 5 * page_size, PROT_READ | PROT_WRITE | PROT_EXEC,
+		   MAP_ANON | MAP_PRIVATE | MAP_FIXED | MAP_NORESERVE, -1, 0);
+	ASSERT_NE(ptr2, MAP_FAILED);
+
+	/* Make sure not merged. */
+	ASSERT_TRUE(find_vma_procmap(procmap, ptr2));
+	ASSERT_EQ(procmap->query.vma_start, (unsigned long)ptr2);
+	ASSERT_EQ(procmap->query.vma_end, (unsigned long)ptr2 + 5 * page_size);
+
+	/*
+	 * Now mprotect forked region to RWX so it becomes the source for the
+	 * merge to unfaulted region:
+	 *
+	 *  forked RWX      RWX
+	 * |-----------|-----------|
+	 * |  faulted  | unfaulted |
+	 * |-----------|-----------|
+	 *      ptr         ptr2
+	 *
+	 * This should NOT result in a merge, as ptr was forked.
+	 */
+	ASSERT_EQ(mprotect(ptr, 5 * page_size, PROT_READ | PROT_WRITE | PROT_EXEC), 0);
+	/* Again, make sure not merged. */
+	ASSERT_TRUE(find_vma_procmap(procmap, ptr2));
+	ASSERT_EQ(procmap->query.vma_start, (unsigned long)ptr2);
+	ASSERT_EQ(procmap->query.vma_end, (unsigned long)ptr2 + 5 * page_size);
+}
+
+TEST_HARNESS_MAIN

tools/testing/selftests/mm/run_vmtests.sh

@@ -79,6 +79,8 @@ separated by spaces:
 	test prctl(PR_SET_MDWE, ...)
 - page_frag
 	test handling of page fragment allocation and freeing
+- vma_merge
+	test VMA merge cases behave as expected
 
 example: ./run_vmtests.sh -t "hmm mmap ksm"
 EOF
@@ -421,6 +423,8 @@ CATEGORY="madv_guard" run_test ./guard-regions
 # MADV_POPULATE_READ and MADV_POPULATE_WRITE tests
 CATEGORY="madv_populate" run_test ./madv_populate
 
+CATEGORY="vma_merge" run_test ./merge
+
 if [ -x ./memfd_secret ]
 then
 (echo 0 > /proc/sys/kernel/yama/ptrace_scope 2>&1) | tap_prefix