aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/pwm/pwm-stm32.c
diff options
context:
space:
mode:
authorSteven Rostedt <[email protected]>2025-03-13 11:16:18 +0000
committerSteven Rostedt (Google) <[email protected]>2025-03-28 12:39:56 +0000
commit5dbeb56bb9589e1051f6af6877cd375f3a901afb (patch)
treeb50177030a4e3c55042c88319b6c83434b71c577 /drivers/pwm/pwm-stm32.c
parenttracing: Fix a compilation error without CONFIG_MODULES (diff)
downloadkernel-5dbeb56bb9589e1051f6af6877cd375f3a901afb.tar.gz
kernel-5dbeb56bb9589e1051f6af6877cd375f3a901afb.zip
tracing: Initialize scratch_size to zero to prevent UB
In allocate_trace_buffer() the following code: buf->buffer = ring_buffer_alloc_range(size, rb_flags, 0, tr->range_addr_start, tr->range_addr_size, struct_size(tscratch, entries, 128)); tscratch = ring_buffer_meta_scratch(buf->buffer, &scratch_size); setup_trace_scratch(tr, tscratch, scratch_size); Has undefined behavior if ring_buffer_alloc_range() fails because "scratch_size" is not initialize. If the allocation fails, then buf->buffer will be NULL. The ring_buffer_meta_scratch() will return NULL immediately if it is passed a NULL buffer and it will not update scratch_size. Then setup_trace_scratch() will return immediately if tscratch is NULL. Although there's no real issue here, but it is considered undefined behavior to pass an uninitialized variable to a function as input, and UBSan may complain about it. Just initialize scratch_size to zero to make the code defined behavior and a little more robust. Link: https://lore.kernel.org/all/[email protected]/ Reported-by: Dan Carpenter <[email protected]> Signed-off-by: Steven Rostedt (Google) <[email protected]>
Diffstat (limited to 'drivers/pwm/pwm-stm32.c')
0 files changed, 0 insertions, 0 deletions