linux.git/lib/kunit/executor.c, branch v6.6.132

kunit: run test suites only after module initialization completes

2024-02-05T20:14:34+00:00

[ Upstream commit a1af6a2bfa0cb46d70b7df5352993e750da6c79b ]

Commit 2810c1e99867 ("kunit: Fix wild-memory-access bug in
kunit_free_suite_set()") fixed a wild-memory-access bug that could have
happened during the loading phase of test suites built and executed as
loadable modules. However, it also introduced a problematic side effect
that causes test suites modules to crash when they attempt to register
fake devices.

When a module is loaded, it traverses the MODULE_STATE_UNFORMED and
MODULE_STATE_COMING states before reaching the normal operating state
MODULE_STATE_LIVE. Finally, when the module is removed, it moves to
MODULE_STATE_GOING before being released. However, if the loading
function load_module() fails between complete_formation() and
do_init_module(), the module goes directly from MODULE_STATE_COMING to
MODULE_STATE_GOING without passing through MODULE_STATE_LIVE.

This behavior was causing kunit_module_exit() to be called without
having first executed kunit_module_init(). Since kunit_module_exit() is
responsible for freeing the memory allocated by kunit_module_init()
through kunit_filter_suites(), this behavior was resulting in a
wild-memory-access bug.

Commit 2810c1e99867 ("kunit: Fix wild-memory-access bug in
kunit_free_suite_set()") fixed this issue by running the tests when the
module is still in MODULE_STATE_COMING. However, modules in that state
are not fully initialized, lacking sysfs kobjects. Therefore, if a test
module attempts to register a fake device, it will inevitably crash.

This patch proposes a different approach to fix the original
wild-memory-access bug while restoring the normal module execution flow
by making kunit_module_exit() able to detect if kunit_module_init() has
previously initialized the tests suite set. In this way, test modules
can once again register fake devices without crashing.

This behavior is achieved by checking whether mod->kunit_suites is a
virtual or direct mapping address. If it is a virtual address, then
kunit_module_init() has allocated the suite_set in kunit_filter_suites()
using kmalloc_array(). On the contrary, if mod->kunit_suites is still
pointing to the original address that was set when looking up the
.kunit_test_suites section of the module, then the loading phase has
failed and there's no memory to be freed.

v4:
- rebased on 6.8
- noted that kunit_filter_suites() must return a virtual address
v3:
- add a comment to clarify why the start address is checked
v2:
- add include 

Fixes: 2810c1e99867 ("kunit: Fix wild-memory-access bug in kunit_free_suite_set()")
Reviewed-by: David Gow 
Tested-by: Rae Moar 
Tested-by: Richard Fitzgerald 
Reviewed-by: Javier Martinez Canillas 
Signed-off-by: Marco Pagani 
Signed-off-by: Shuah Khan 
Signed-off-by: Sasha Levin

kunit: Fix possible memory leak in kunit_filter_suites()

2023-11-20T10:59:15+00:00

[ Upstream commit 24de14c98b37ea40a7e493dfd0d93b400b6efbca ]

If the outer layer for loop is iterated more than once and it fails not
in the first iteration, the filtered_suite and filtered_suite->test_cases
allocated in the last kunit_filter_attr_tests() in last inner for loop
is leaked.

So add a new free_filtered_suite err label and free the filtered_suite
and filtered_suite->test_cases so far. And change kmalloc_array of copy
to kcalloc to Clear the copy to make the kfree safe.

Fixes: 529534e8cba3 ("kunit: Add ability to filter attributes")
Signed-off-by: Jinjie Ruan 
Reviewed-by: Rae Moar 
Reviewed-by: David Gow 
Signed-off-by: Shuah Khan 
Signed-off-by: Sasha Levin

kunit: Fix the wrong kfree of copy for kunit_filter_suites()

2023-11-20T10:59:15+00:00

[ Upstream commit e44679515a7b803cf0143dc9de3d2ecbe907f939 ]

If the outer layer for loop is iterated more than once and it fails not
in the first iteration, the copy pointer has been moved. So it should free
the original copy's backup copy_start.

Fixes: abbf73816b6f ("kunit: fix possible memory leak in kunit_filter_suites()")
Signed-off-by: Jinjie Ruan 
Reviewed-by: Rae Moar 
Reviewed-by: David Gow 
Signed-off-by: Shuah Khan 
Signed-off-by: Sasha Levin

kunit: Fix missed memory release in kunit_free_suite_set()

2023-11-20T10:59:15+00:00

[ Upstream commit a6074cf0126b0bee51ab77a15930dc24a4d5db90 ]

modprobe cpumask_kunit and rmmod cpumask_kunit, kmemleak detect
a suspected memory leak as below.

If kunit_filter_suites() in kunit_module_init() succeeds, the
suite_set.start will not be NULL and the kunit_free_suite_set() in
kunit_module_exit() should free all the memory which has not
been freed. However the test_cases in suites is left out.

unreferenced object 0xffff54ac47e83200 (size 512):
  comm "modprobe", pid 592, jiffies 4294913238 (age 1367.612s)
  hex dump (first 32 bytes):
    84 13 1a f0 d3 b6 ff ff 30 68 1a f0 d3 b6 ff ff  ........0h......
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    [<000000008dec63a2>] slab_post_alloc_hook+0xb8/0x368
    [<00000000ec280d8e>] __kmem_cache_alloc_node+0x174/0x290
    [<00000000896c7740>] __kmalloc+0x60/0x2c0
    [<000000007a50fa06>] kunit_filter_suites+0x254/0x5b8
    [<0000000078cc98e2>] kunit_module_notify+0xf4/0x240
    [<0000000033cea952>] notifier_call_chain+0x98/0x17c
    [<00000000973d05cc>] notifier_call_chain_robust+0x4c/0xa4
    [<000000005f95895f>] blocking_notifier_call_chain_robust+0x4c/0x74
    [<0000000048e36fa7>] load_module+0x1a2c/0x1c40
    [<0000000004eb8a91>] init_module_from_file+0x94/0xcc
    [<0000000037dbba28>] idempotent_init_module+0x184/0x278
    [<00000000161b75cb>] __arm64_sys_finit_module+0x68/0xa8
    [<000000006dc1669b>] invoke_syscall+0x44/0x100
    [<00000000fa87e304>] el0_svc_common.constprop.1+0x68/0xe0
    [<000000009d8ad866>] do_el0_svc+0x1c/0x28
    [<000000005b83c607>] el0_svc+0x3c/0xc4

Fixes: a127b154a8f2 ("kunit: tool: allow filtering test cases via glob")
Signed-off-by: Jinjie Ruan 
Reviewed-by: Rae Moar 
Reviewed-by: David Gow 
Signed-off-by: Shuah Khan 
Signed-off-by: Sasha Levin

kunit: Fix possible memory leak in kunit_filter_suites()

2023-09-05T18:30:06+00:00

If both filter_glob and filters are not NULL, and kunit_parse_glob_filter()
succeed, but kcalloc parsed_filters fails, the suite_glob and test_glob of
parsed kzalloc in kunit_parse_glob_filter() will be leaked.

As Rae suggested, assign -ENOMEM to *err to correctly free copy and goto
free_parsed_glob to free the suite/test_glob of parsed.

Fixes: 1c9fd080dffe ("kunit: fix uninitialized variables bug in attributes filtering")
Signed-off-by: Jinjie Ruan 
Suggested-by: Rae Moar 
Reviewed-by: David Gow 
Signed-off-by: Shuah Khan

kunit: Fix possible null-ptr-deref in kunit_parse_glob_filter()

2023-09-05T18:30:01+00:00

Inject fault while probing kunit-example-test.ko, if kzalloc fails
in kunit_parse_glob_filter(), strcpy() or strncpy() to NULL will
cause below null-ptr-deref bug. So check NULL for kzalloc() and
return int instead of void for kunit_parse_glob_filter().

 Unable to handle kernel paging request at virtual address dfff800000000000
 KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007]
 Mem abort info:
   ESR = 0x0000000096000005
   EC = 0x25: DABT (current EL), IL = 32 bits
   SET = 0, FnV = 0
   EA = 0, S1PTW = 0
   FSC = 0x05: level 1 translation fault
 Data abort info:
   ISV = 0, ISS = 0x00000005, ISS2 = 0x00000000
   CM = 0, WnR = 0, TnD = 0, TagAccess = 0
   GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
 [dfff800000000000] address between user and kernel address ranges
 Internal error: Oops: 0000000096000005 [#1] PREEMPT SMP
 Modules linked in: kunit_example_test cfg80211 rfkill 8021q garp mrp stp llc ipv6 [last unloaded: kunit_example_test]
 CPU: 4 PID: 6047 Comm: modprobe Tainted: G        W        N 6.5.0-next-20230829+ #141
 Hardware name: linux,dummy-virt (DT)
 pstate: 80000005 (Nzcv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
 pc : strncpy+0x58/0xc0
 lr : kunit_filter_suites+0x15c/0xa84
 sp : ffff800082a17420
 x29: ffff800082a17420 x28: 0000000000000000 x27: 0000000000000004
 x26: 0000000000000000 x25: ffffa847e40a5320 x24: 0000000000000001
 x23: 0000000000000000 x22: 0000000000000001 x21: dfff800000000000
 x20: 000000000000002a x19: 0000000000000000 x18: 00000000750b3b54
 x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000
 x14: 0000000000000000 x13: 34393178302f3039 x12: ffff7508fcea4ec1
 x11: 1ffff508fcea4ec0 x10: ffff7508fcea4ec0 x9 : dfff800000000000
 x8 : ffff6051b1a7f86a x7 : ffff800082a17270 x6 : 0000000000000002
 x5 : 0000000000000098 x4 : ffff028d9817b250 x3 : 0000000000000000
 x2 : 0000000000000000 x1 : ffffa847e40a5320 x0 : 0000000000000000
 Call trace:
  strncpy+0x58/0xc0
  kunit_filter_suites+0x15c/0xa84
  kunit_module_notify+0x1b0/0x3ac
  blocking_notifier_call_chain+0xc4/0x128
  do_init_module+0x250/0x594
  load_module+0x37b0/0x44b4
  init_module_from_file+0xd4/0x128
  idempotent_init_module+0x2c8/0x524
  __arm64_sys_finit_module+0xac/0x100
  invoke_syscall+0x6c/0x258
  el0_svc_common.constprop.0+0x160/0x22c
  do_el0_svc+0x44/0x5c
  el0_svc+0x38/0x78
  el0t_64_sync_handler+0x13c/0x158
  el0t_64_sync+0x190/0x194
 Code: 5400028a d343fe63 12000a62 39400034 (38f56863)
 ---[ end trace 0000000000000000 ]---
 Kernel panic - not syncing: Oops: Fatal exception
 SMP: stopping secondary CPUs
 Kernel Offset: 0x284761400000 from 0xffff800080000000
 PHYS_OFFSET: 0xfffffd7380000000
 CPU features: 0x88000203,3c020000,1000421b
 Memory Limit: none
 Rebooting in 1 seconds..

Fixes: a127b154a8f2 ("kunit: tool: allow filtering test cases via glob")
Signed-off-by: Jinjie Ruan 
Reviewed-by: Rae Moar 
Reviewed-by: David Gow 
Signed-off-by: Shuah Khan

kunit: Fix the wrong err path and add goto labels in kunit_filter_suites()

2023-09-05T18:29:55+00:00

Take the last kfree(parsed_filters) and add it to be the first. Take
the first kfree(copy) and add it to be the last. The Best practice is to
return these errors reversely.

And as David suggested, add several labels which target only the things
which actually have been allocated so far.

Fixes: 529534e8cba3 ("kunit: Add ability to filter attributes")
Fixes: abbf73816b6f ("kunit: fix possible memory leak in kunit_filter_suites()")
Signed-off-by: Jinjie Ruan 
Reviewed-by: Rae Moar 
Suggested-by: David Gow 
Reviewed-by: David Gow 
Signed-off-by: Shuah Khan

kunit: Allow kunit test modules to use test filtering

2023-08-08T19:46:18+00:00

External tools, e.g., Intel GPU tools (IGT), support execution of
individual selftests provided by kernel modules.  That could be also
applicable to kunit test modules if they provided test filtering.  But
test filtering is now possible only when kunit code is built into the
kernel.  Moreover, a filter can be specified only at boot time, then
reboot is required each time a different filter is needed.

Build the test filtering code also when kunit is configured as a module,
expose test filtering functions to other kunit source files, and use them
in kunit module notifier callback functions.  Userspace can then reload
the kunit module with a value of the filter_glob parameter tuned to a
specific kunit test module every time it wants to limit the scope of tests
executed on that module load.  Make the kunit.filter* parameters visible
in sysfs for user convenience.

v5: Refresh on tpp of attributes filtering fix
v4: Refresh on top of newly applied attributes patches and changes
    introdced by new versions of other patches submitted in series with
    this one.
v3: Fix CONFIG_GLOB, required by filtering functions, not selected when
    building as a module (lkp@intel.com).
v2: Fix new name of a structure moved to kunit namespace not updated
    across all uses (lkp@intel.com).

Signed-off-by: Janusz Krzysztofik 
Reviewed-by: David Gow 
Signed-off-by: Shuah Khan

kunit: Make 'list' action available to kunit test modules

2023-08-08T19:46:13+00:00

Results from kunit tests reported via dmesg may be interleaved with other
kernel messages.  When parsing dmesg for modular kunit results in real
time, external tools, e.g., Intel GPU tools (IGT), may want to insert
their own test name markers into dmesg at the start of each test, before
any kernel message related to that test appears there, so existing upper
level test result parsers have no doubt which test to blame for a specific
kernel message.  Unfortunately, kunit reports names of tests only at their
completion (with the exeption of a not standarized "# Subtest: "
header above a test plan of each test suite or parametrized test).

External tools could be able to insert their own "start of the test"
markers with test names included if they new those names in advance.
Test names could be learned from a list if provided by a kunit test
module.

There exists a feature of listing kunit tests without actually executing
them, but it is now limited to configurations with the kunit module built
in and covers only built-in tests, already available at boot time.
Moreover, switching from list to normal mode requires reboot.  If that
feature was also available when kunit is built as a module, userspace
could load the module with action=list parameter, load some kunit test
modules they are interested in and learn about the list of tests provided
by those modules, then unload them, reload the kunit module in normal mode
and execute the tests with their lists already known.

Extend kunit module notifier initialization callback with a processing
path for only listing the tests provided by a module if the kunit action
parameter is set to "list" or "list_attr".  For user convenience, make the
kunit.action parameter visible in sysfs.

v2: Don't use a different format, use kunit_exec_list_tests() (Rae),
  - refresh on top of new attributes patches, handle newly introduced
    kunit.action=list_attr case (Rae).

Signed-off-by: Janusz Krzysztofik 
Cc: Rae Moar 
Reviewed-by: David Gow 
Signed-off-by: Shuah Khan

kunit: Report the count of test suites in a module

2023-08-08T19:46:05+00:00

According to KTAP specification[1], results should always start from a
header that provides a TAP protocol version, followed by a test plan with
a count of items to be executed.  That pattern should be followed at each
nesting level.  In the current implementation of the top-most, i.e., test
suite level, those rules apply only for test suites built into the kernel,
executed and reported on boot.  Results submitted to dmesg from kunit test
modules loaded later are missing those top-level headers.

As a consequence, if a kunit test module provides more than one test suite
then, without the top level test plan, external tools that are parsing
dmesg for kunit test output are not able to tell how many test suites
should be expected and whether to continue parsing after complete output
from the first test suite is collected.

Submit the top-level headers also from the kunit test module notifier
initialization callback.

v3: Fix new name of a structure moved to kunit namespace not updated in
    executor_test functions (lkp@intel.com).
v2: Use kunit_exec_run_tests() (Mauro, Rae), but prevent it from
    emitting the headers when called on load of non-test modules.

[1] https://docs.kernel.org/dev-tools/ktap.html#

Signed-off-by: Janusz Krzysztofik 
Cc: Mauro Carvalho Chehab 
Cc: Rae Moar 
Reviewed-by: Rae Moar 
Reviewed-by: David Gow 
Signed-off-by: Shuah Khan