linux.git/include/trace/events/dma.h, branch vsnprintf

dma-mapping: trace more error paths

2024-10-29T07:54:06+00:00

It can be surprising to the user if DMA functions are only traced on
success. On failure, it can be unclear what the source of the problem
is. Fix this by tracing all functions even when they fail. Cases where
we BUG/WARN are skipped, since those should be sufficiently noisy
already.

Signed-off-by: Sean Anderson 
Reviewed-by: Steven Rostedt (Google) 
Signed-off-by: Christoph Hellwig

dma-mapping: use trace_dma_alloc for dma_alloc* instead of using trace_dma_map

2024-10-29T07:54:03+00:00

In some cases, we use trace_dma_map to trace dma_alloc* functions. This
generally follows dma_debug. However, this does not record all of the
relevant information for allocations, such as GFP flags. Create new
dma_alloc tracepoints for these functions. Note that while
dma_alloc_noncontiguous may allocate discontiguous pages (from the CPU's
point of view), the device will only see one contiguous mapping.
Therefore, we just need to trace dma_addr and size.

Signed-off-by: Sean Anderson 
Reviewed-by: Steven Rostedt (Google) 
Signed-off-by: Christoph Hellwig

dma-mapping: trace dma_alloc/free direction

2024-10-29T07:54:00+00:00

In preparation for using these tracepoints in a few more places, trace
the DMA direction as well. For coherent allocations this is always
bidirectional.

Signed-off-by: Sean Anderson 
Reviewed-by: Steven Rostedt (Google) 
Signed-off-by: Christoph Hellwig

dma-mapping: use macros to define events in a class

2024-10-29T07:53:57+00:00

Use a macro to avoid repeating the parameters and arguments for each event
in a class.

Signed-off-by: Sean Anderson 
Reviewed-by: Steven Rostedt (Google) 
Signed-off-by: Christoph Hellwig

dma-mapping: fix tracing dma_alloc/free with vmalloc'd memory

2024-10-17T15:41:29+00:00

Not all virtual addresses have physical addresses, such as if they were
vmalloc'd. Just trace the virtual address instead of trying to trace a
physical address. This aligns with the API, and is good enough to
associate dma_alloc with dma_free.

Fixes: 038eb433dc14 ("dma-mapping: add tracing for dma-mapping API calls")
Reported-by: syzbot+b4bfacdec173efaa8567@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/all/670ebde5.050a0220.d9b66.0154.GAE@google.com/
Signed-off-by: Sean Anderson 
Signed-off-by: Christoph Hellwig

dma-mapping: fix DMA API tracing for chained scatterlists

2024-09-26T18:08:12+00:00

scatterlist allocations can be chained, and thus all iterations need to
use the chain-aware iterators.  Switch the newly added tracing to use the
proper iterators so that they work with chained scatterlists.

Fixes: 038eb433dc14 ("dma-mapping: add tracing for dma-mapping API calls")
Reported-by: syzbot+95e4ef83a3024384ec7a@syzkaller.appspotmail.com
Signed-off-by: Christoph Hellwig 
Reviewed-by: Sean Anderson 
Tested-by: syzbot+95e4ef83a3024384ec7a@syzkaller.appspotmail.com

dma-mapping: add tracing for dma-mapping API calls

2024-09-10T04:47:19+00:00

When debugging drivers, it can often be useful to trace when memory gets
(un)mapped for DMA (and can be accessed by the device). Add some
tracepoints for this purpose.

Use u64 instead of phys_addr_t and dma_addr_t (and similarly %llx instead
of %pa) because libtraceevent can't handle typedefs in all cases.

Signed-off-by: Sean Anderson 
Signed-off-by: Christoph Hellwig