linux-stable.git/kernel/trace/trace.h, branch v6.10.7

tracing: Have format file honor EVENT_FILE_FL_FREED

2024-08-14T13:34:33+00:00

commit b1560408692cd0ab0370cfbe9deb03ce97ab3f6d upstream.

When eventfs was introduced, special care had to be done to coordinate the
freeing of the file meta data with the files that are exposed to user
space. The file meta data would have a ref count that is set when the file
is created and would be decremented and freed after the last user that
opened the file closed it. When the file meta data was to be freed, it
would set a flag (EVENT_FILE_FL_FREED) to denote that the file is freed,
and any new references made (like new opens or reads) would fail as it is
marked freed. This allowed other meta data to be freed after this flag was
set (under the event_mutex).

All the files that were dynamically created in the events directory had a
pointer to the file meta data and would call event_release() when the last
reference to the user space file was closed. This would be the time that it
is safe to free the file meta data.

A shortcut was made for the "format" file. It's i_private would point to
the "call" entry directly and not point to the file's meta data. This is
because all format files are the same for the same "call", so it was
thought there was no reason to differentiate them.  The other files
maintain state (like the "enable", "trigger", etc). But this meant if the
file were to disappear, the "format" file would be unaware of it.

This caused a race that could be trigger via the user_events test (that
would create dynamic events and free them), and running a loop that would
read the user_events format files:

In one console run:

 # cd tools/testing/selftests/user_events
 # while true; do ./ftrace_test; done

And in another console run:

 # cd /sys/kernel/tracing/
 # while true; do cat events/user_events/__test_event/format; done 2>/dev/null

With KASAN memory checking, it would trigger a use-after-free bug report
(which was a real bug). This was because the format file was not checking
the file's meta data flag "EVENT_FILE_FL_FREED", so it would access the
event that the file meta data pointed to after the event was freed.

After inspection, there are other locations that were found to not check
the EVENT_FILE_FL_FREED flag when accessing the trace_event_file. Add a
new helper function: event_file_file() that will make sure that the
event_mutex is held, and will return NULL if the trace_event_file has the
EVENT_FILE_FL_FREED flag set. Have the first reference of the struct file
pointer use event_file_file() and check for NULL. Later uses can still use
the event_file_data() helper function if the event_mutex is still held and
was not released since the event_file_file() call.

Link: https://lore.kernel.org/all/20240719204701.1605950-1-minipli@grsecurity.net/

Cc: stable@vger.kernel.org
Cc: Masami Hiramatsu 
Cc: Mathieu Desnoyers   
Cc: Ajay Kaher 
Cc: Ilkka Naulapää    
Cc: Linus Torvalds 
Cc: Al   Viro 
Cc: Dan Carpenter   
Cc: Beau Belgrave 
Cc: Florian Fainelli  
Cc: Alexey Makhalov    
Cc: Vasavi Sirnapalli    
Link: https://lore.kernel.org/20240730110657.3b69d3c1@gandalf.local.home
Fixes: b63db58e2fa5d ("eventfs/tracing: Add callback for release of an eventfs_inode")
Reported-by: Mathias Krause 
Tested-by: Mathias Krause 
Signed-off-by: Steven Rostedt (Google) 
Signed-off-by: Greg Kroah-Hartman

tracing: Allow user-space mapping of the ring-buffer

2024-05-13T22:09:56+00:00

Currently, user-space extracts data from the ring-buffer via splice,
which is handy for storage or network sharing. However, due to splice
limitations, it is imposible to do real-time analysis without a copy.

A solution for that problem is to let the user-space map the ring-buffer
directly.

The mapping is exposed via the per-CPU file trace_pipe_raw. The first
element of the mapping is the meta-page. It is followed by each
subbuffer constituting the ring-buffer, ordered by their unique page ID:

  * Meta-page -- include/uapi/linux/trace_mmap.h for a description
  * Subbuf ID 0
  * Subbuf ID 1
     ...

It is therefore easy to translate a subbuf ID into an offset in the
mapping:

  reader_id = meta->reader->id;
  reader_offset = meta->meta_page_size + reader_id * meta->subbuf_size;

When new data is available, the mapper must call a newly introduced ioctl:
TRACE_MMAP_IOCTL_GET_READER. This will update the Meta-page reader ID to
point to the next reader containing unread data.

Mapping will prevent snapshot and buffer size modifications.

Link: https://lore.kernel.org/linux-trace-kernel/20240510140435.3550353-4-vdonnefort@google.com

CC: 
Signed-off-by: Vincent Donnefort 
Signed-off-by: Steven Rostedt (Google)

tracing: Add snapshot refcount

2024-03-18T14:12:47+00:00

When a ring-buffer is memory mapped by user-space, no trace or
ring-buffer swap is possible. This means the snapshot feature is
mutually exclusive with the memory mapping. Having a refcount on
snapshot users will help to know if a mapping is possible or not.

Instead of relying on the global trace_types_lock, a new spinlock is
introduced to serialize accesses to trace_array->snapshot. This intends
to allow access to that variable in a context where the mmap lock is
already held.

Link: https://lore.kernel.org/linux-trace-kernel/20240220202310.2489614-4-vdonnefort@google.com

Signed-off-by: Vincent Donnefort 
Signed-off-by: Steven Rostedt (Google)

tracing: Move saved_cmdline code into trace_sched_switch.c

2024-03-17T11:58:53+00:00

The code that handles saved_cmdlines is split between the trace.c file and
the trace_sched_switch.c. There's some history to this. The
trace_sched_switch.c was originally created to handle the sched_switch
tracer that was deprecated due to sched_switch trace event making it
obsolete. But that file did not get deleted as it had some code to help
with saved_cmdlines. But trace.c has grown tremendously since then. Just
move all the saved_cmdlines code into trace_sched_switch.c as that's the
only reason that file still exists, and trace.c has gotten too big.

No functional changes.

Link: https://lore.kernel.org/linux-trace-kernel/20240220140703.497966629@goodmis.org

Cc: Masami Hiramatsu 
Cc: Mark Rutland 
Cc: Mathieu Desnoyers 
Cc: Andrew Morton 
Cc: Tim Chen 
Cc: Vincent Donnefort 
Cc: Sven Schnelle 
Cc: Mete Durlu 
Signed-off-by: Steven Rostedt (Google)

ring-buffer: Page size per ring buffer

2023-12-20T12:54:55+00:00

Currently the size of one sub buffer page is global for all buffers and
it is hard coded to one system page. In order to introduce configurable
ring buffer sub page size, the internal logic should be refactored to
work with sub page size per ring buffer.

Link: https://lore.kernel.org/linux-trace-devel/20211213094825.61876-3-tz.stoyanov@gmail.com
Link: https://lore.kernel.org/linux-trace-kernel/20231219185628.009147038@goodmis.org

Cc: Masami Hiramatsu 
Cc: Mark Rutland 
Cc: Mathieu Desnoyers 
Cc: Andrew Morton 
Cc: Vincent Donnefort 
Cc: Kent Overstreet 
Signed-off-by: Tzvetomir Stoyanov (VMware) 
Signed-off-by: Steven Rostedt (Google)

tracing: Allow creating instances with specified system events

2023-12-19T04:14:16+00:00

A trace instance may only need to enable specific events. As the eventfs
directory of an instance currently creates all events which adds overhead,
allow internal instances to be created with just the events in systems
that they care about. This currently only deals with systems and not
individual events, but this should bring down the overhead of creating
instances for specific use cases quite bit.

The trace_array_get_by_name() now has another parameter "systems". This
parameter is a const string pointer of a comma/space separated list of
event systems that should be created by the trace_array. (Note if the
trace_array already exists, this parameter is ignored).

The list of systems is saved and if a module is loaded, its events will
not be added unless the system for those events also match the systems
string.

Link: https://lore.kernel.org/linux-trace-kernel/20231213093701.03fddec0@gandalf.local.home

Cc: Masami Hiramatsu 
Cc: Mark Rutland 
Cc: Mathieu Desnoyers 
Cc: Sean Paul 
Cc: Arun Easi   
Cc: Daniel Wagner 
Tested-by: Dmytro Maluka 
Signed-off-by: Steven Rostedt (Google)

tracing: Fix uaf issue when open the hist or hist_debug file

2023-12-14T04:29:59+00:00

KASAN report following issue. The root cause is when opening 'hist'
file of an instance and accessing 'trace_event_file' in hist_show(),
but 'trace_event_file' has been freed due to the instance being removed.
'hist_debug' file has the same problem. To fix it, call
tracing_{open,release}_file_tr() in file_operations callback to have
the ref count and avoid 'trace_event_file' being freed.

  BUG: KASAN: slab-use-after-free in hist_show+0x11e0/0x1278
  Read of size 8 at addr ffff242541e336b8 by task head/190

  CPU: 4 PID: 190 Comm: head Not tainted 6.7.0-rc5-g26aff849438c #133
  Hardware name: linux,dummy-virt (DT)
  Call trace:
   dump_backtrace+0x98/0xf8
   show_stack+0x1c/0x30
   dump_stack_lvl+0x44/0x58
   print_report+0xf0/0x5a0
   kasan_report+0x80/0xc0
   __asan_report_load8_noabort+0x1c/0x28
   hist_show+0x11e0/0x1278
   seq_read_iter+0x344/0xd78
   seq_read+0x128/0x1c0
   vfs_read+0x198/0x6c8
   ksys_read+0xf4/0x1e0
   __arm64_sys_read+0x70/0xa8
   invoke_syscall+0x70/0x260
   el0_svc_common.constprop.0+0xb0/0x280
   do_el0_svc+0x44/0x60
   el0_svc+0x34/0x68
   el0t_64_sync_handler+0xb8/0xc0
   el0t_64_sync+0x168/0x170

  Allocated by task 188:
   kasan_save_stack+0x28/0x50
   kasan_set_track+0x28/0x38
   kasan_save_alloc_info+0x20/0x30
   __kasan_slab_alloc+0x6c/0x80
   kmem_cache_alloc+0x15c/0x4a8
   trace_create_new_event+0x84/0x348
   __trace_add_new_event+0x18/0x88
   event_trace_add_tracer+0xc4/0x1a0
   trace_array_create_dir+0x6c/0x100
   trace_array_create+0x2e8/0x568
   instance_mkdir+0x48/0x80
   tracefs_syscall_mkdir+0x90/0xe8
   vfs_mkdir+0x3c4/0x610
   do_mkdirat+0x144/0x200
   __arm64_sys_mkdirat+0x8c/0xc0
   invoke_syscall+0x70/0x260
   el0_svc_common.constprop.0+0xb0/0x280
   do_el0_svc+0x44/0x60
   el0_svc+0x34/0x68
   el0t_64_sync_handler+0xb8/0xc0
   el0t_64_sync+0x168/0x170

  Freed by task 191:
   kasan_save_stack+0x28/0x50
   kasan_set_track+0x28/0x38
   kasan_save_free_info+0x34/0x58
   __kasan_slab_free+0xe4/0x158
   kmem_cache_free+0x19c/0x508
   event_file_put+0xa0/0x120
   remove_event_file_dir+0x180/0x320
   event_trace_del_tracer+0xb0/0x180
   __remove_instance+0x224/0x508
   instance_rmdir+0x44/0x78
   tracefs_syscall_rmdir+0xbc/0x140
   vfs_rmdir+0x1cc/0x4c8
   do_rmdir+0x220/0x2b8
   __arm64_sys_unlinkat+0xc0/0x100
   invoke_syscall+0x70/0x260
   el0_svc_common.constprop.0+0xb0/0x280
   do_el0_svc+0x44/0x60
   el0_svc+0x34/0x68
   el0t_64_sync_handler+0xb8/0xc0
   el0t_64_sync+0x168/0x170

Link: https://lore.kernel.org/linux-trace-kernel/20231214012153.676155-1-zhengyejian1@huawei.com

Suggested-by: Steven Rostedt 
Signed-off-by: Zheng Yejian 
Signed-off-by: Steven Rostedt (Google)

tracing: Have trace_event_file have ref counters

2023-11-02T03:44:44+00:00

The following can crash the kernel:

 # cd /sys/kernel/tracing
 # echo 'p:sched schedule' > kprobe_events
 # exec 5>>events/kprobes/sched/enable
 # > kprobe_events
 # exec 5>&-

The above commands:

 1. Change directory to the tracefs directory
 2. Create a kprobe event (doesn't matter what one)
 3. Open bash file descriptor 5 on the enable file of the kprobe event
 4. Delete the kprobe event (removes the files too)
 5. Close the bash file descriptor 5

The above causes a crash!

 BUG: kernel NULL pointer dereference, address: 0000000000000028
 #PF: supervisor read access in kernel mode
 #PF: error_code(0x0000) - not-present page
 PGD 0 P4D 0
 Oops: 0000 [#1] PREEMPT SMP PTI
 CPU: 6 PID: 877 Comm: bash Not tainted 6.5.0-rc4-test-00008-g2c6b6b1029d4-dirty #186
 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
 RIP: 0010:tracing_release_file_tr+0xc/0x50

What happens here is that the kprobe event creates a trace_event_file
"file" descriptor that represents the file in tracefs to the event. It
maintains state of the event (is it enabled for the given instance?).
Opening the "enable" file gets a reference to the event "file" descriptor
via the open file descriptor. When the kprobe event is deleted, the file is
also deleted from the tracefs system which also frees the event "file"
descriptor.

But as the tracefs file is still opened by user space, it will not be
totally removed until the final dput() is called on it. But this is not
true with the event "file" descriptor that is already freed. If the user
does a write to or simply closes the file descriptor it will reference the
event "file" descriptor that was just freed, causing a use-after-free bug.

To solve this, add a ref count to the event "file" descriptor as well as a
new flag called "FREED". The "file" will not be freed until the last
reference is released. But the FREE flag will be set when the event is
removed to prevent any more modifications to that event from happening,
even if there's still a reference to the event "file" descriptor.

Link: https://lore.kernel.org/linux-trace-kernel/20231031000031.1e705592@gandalf.local.home/
Link: https://lore.kernel.org/linux-trace-kernel/20231031122453.7a48b923@gandalf.local.home

Cc: stable@vger.kernel.org
Cc: Mark Rutland 
Fixes: f5ca233e2e66d ("tracing: Increase trace array ref count on enable and filter files")
Reported-by: Beau Belgrave 
Tested-by: Beau Belgrave 
Reviewed-by: Masami Hiramatsu (Google) 
Signed-off-by: Steven Rostedt (Google)

eventfs: Remove eventfs_file and just use eventfs_inode

2023-10-04T21:11:50+00:00

Instead of having a descriptor for every file represented in the eventfs
directory, only have the directory itself represented. Change the API to
send in a list of entries that represent all the files in the directory
(but not other directories). The entry list contains a name and a callback
function that will be used to create the files when they are accessed.

struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry *parent,
						const struct eventfs_entry *entries,
						int size, void *data);

is used for the top level eventfs directory, and returns an eventfs_inode
that will be used by:

struct eventfs_inode *eventfs_create_dir(const char *name, struct eventfs_inode *parent,
					 const struct eventfs_entry *entries,
					 int size, void *data);

where both of the above take an array of struct eventfs_entry entries for
every file that is in the directory.

The entries are defined by:

typedef int (*eventfs_callback)(const char *name, umode_t *mode, void **data,
				const struct file_operations **fops);

struct eventfs_entry {
	const char			*name;
	eventfs_callback		callback;
};

Where the name is the name of the file and the callback gets called when
the file is being created. The callback passes in the name (in case the
same callback is used for multiple files), a pointer to the mode, data and
fops. The data will be pointing to the data that was passed in
eventfs_create_dir() or eventfs_create_events_dir() but may be overridden
to point to something else, as it will be used to point to the
inode->i_private that is created. The information passed back from the
callback is used to create the dentry/inode.

If the callback fills the data and the file should be created, it must
return a positive number. On zero or negative, the file is ignored.

This logic may also be used as a prototype to convert entire pseudo file
systems into just-in-time allocation.

The "show_events_dentry" file has been updated to show the directories,
and any files they have.

With just the eventfs_file allocations:

 Before after deltas for meminfo (in kB):

   MemFree:		-14360
   MemAvailable:	-14260
   Buffers:		40
   Cached:		24
   Active:		44
   Inactive:		48
   Inactive(anon):	28
   Active(file):	44
   Inactive(file):	20
   Dirty:		-4
   AnonPages:		28
   Mapped:		4
   KReclaimable:	132
   Slab:		1604
   SReclaimable:	132
   SUnreclaim:		1472
   Committed_AS:	12

 Before after deltas for slabinfo:

   :			[ *  = ]

   ext4_inode_cache	27		[* 1184 = 31968 ]
   extent_status	102		[*   40 = 4080 ]
   tracefs_inode_cache	144		[*  656 = 94464 ]
   buffer_head		39		[*  104 = 4056 ]
   shmem_inode_cache	49		[*  800 = 39200 ]
   filp			-53		[*  256 = -13568 ]
   dentry		251		[*  192 = 48192 ]
   lsm_file_cache	277		[*   32 = 8864 ]
   vm_area_struct	-14		[*  184 = -2576 ]
   trace_event_file	1748		[*   88 = 153824 ]
   kmalloc-1k		35		[* 1024 = 35840 ]
   kmalloc-256		49		[*  256 = 12544 ]
   kmalloc-192		-28		[*  192 = -5376 ]
   kmalloc-128		-30		[*  128 = -3840 ]
   kmalloc-96		10581		[*   96 = 1015776 ]
   kmalloc-64		3056		[*   64 = 195584 ]
   kmalloc-32		1291		[*   32 = 41312 ]
   kmalloc-16		2310		[*   16 = 36960 ]
   kmalloc-8		9216		[*    8 = 73728 ]

 Free memory dropped by 14,360 kB
 Available memory dropped by 14,260 kB
 Total slab additions in size: 1,771,032 bytes

With this change:

 Before after deltas for meminfo (in kB):

   MemFree:		-12084
   MemAvailable:	-11976
   Buffers:		32
   Cached:		32
   Active:		72
   Inactive:		168
   Inactive(anon):	176
   Active(file):	72
   Inactive(file):	-8
   Dirty:		24
   AnonPages:		196
   Mapped:		8
   KReclaimable:	148
   Slab:		836
   SReclaimable:	148
   SUnreclaim:		688
   Committed_AS:	324

 Before after deltas for slabinfo:

   :			[ *  = ]

   tracefs_inode_cache	144		[* 656 = 94464 ]
   shmem_inode_cache	-23		[* 800 = -18400 ]
   filp			-92		[* 256 = -23552 ]
   dentry		179		[* 192 = 34368 ]
   lsm_file_cache	-3		[* 32 = -96 ]
   vm_area_struct	-13		[* 184 = -2392 ]
   trace_event_file	1748		[* 88 = 153824 ]
   kmalloc-1k		-49		[* 1024 = -50176 ]
   kmalloc-256		-27		[* 256 = -6912 ]
   kmalloc-128		1864		[* 128 = 238592 ]
   kmalloc-64		4685		[* 64 = 299840 ]
   kmalloc-32		-72		[* 32 = -2304 ]
   kmalloc-16		256		[* 16 = 4096 ]
   total = 721352

 Free memory dropped by 12,084 kB
 Available memory dropped by 11,976 kB
 Total slab additions in size:  721,352 bytes

That's over 2 MB in savings per instance for free and available memory,
and over 1 MB in savings per instance of slab memory.

Link: https://lore.kernel.org/linux-trace-kernel/20231003184059.4924468e@gandalf.local.home
Link: https://lore.kernel.org/linux-trace-kernel/20231004165007.43d79161@gandalf.local.home

Cc: Masami Hiramatsu 
Cc: Mark Rutland 
Cc: Andrew Morton 
Cc: Ajay Kaher 
Signed-off-by: Steven Rostedt (Google)

tracing: Expand all ring buffers individually

2023-10-03T23:02:06+00:00

The ring buffer of global_trace is set to the minimum size in
order to save memory on boot up and then it will be expand when
some trace feature enabled.

However currently operations under an instance can also cause
global_trace ring buffer being expanded, and the expanded memory
would be wasted if global_trace then not being used.

See following case, we enable 'sched_switch' event in instance 'A', then
ring buffer of global_trace is unexpectedly expanded to be 1410KB, also
the '(expanded: 1408)' from 'buffer_size_kb' of instance is confusing.

  # cd /sys/kernel/tracing
  # mkdir instances/A
  # cat buffer_size_kb
  7 (expanded: 1408)
  # cat instances/A/buffer_size_kb
  1410 (expanded: 1408)
  # echo sched:sched_switch > instances/A/set_event
  # cat buffer_size_kb
  1410
  # cat instances/A/buffer_size_kb
  1410

To fix it, we can:
  - Make 'ring_buffer_expanded' as a member of 'struct trace_array';
  - Make 'ring_buffer_expanded' of instance is defaultly true,
    global_trace is defaultly false;
  - In order not to expose 'global_trace' outside of file
    'kernel/trace/trace.c', introduce trace_set_ring_buffer_expanded()
    to set 'ring_buffer_expanded' as 'true';
  - Pass the expected trace_array to tracing_update_buffers().

Link: https://lore.kernel.org/linux-trace-kernel/20230906091837.3998020-1-zhengyejian1@huawei.com

Signed-off-by: Zheng Yejian 
Signed-off-by: Steven Rostedt (Google)