Re: [PATCH] nvme: Fix nvme_setup_command metadata trace event for cdw10

[Keith Busch]

On Wed, Jul 06, 2022 at 04:16:38PM +0800, John Garry wrote:
> For x86_64 allmodconfig I get this warning:
> 
> In function ‘fortify_memcpy_chk’,
>     inlined from ‘perf_trace_nvme_setup_cmd’ at drivers/nvme/host/./trace.h:47:1:
> ./include/linux/fortify-string.h:352:4: error: call to ‘__read_overflow2_field’ declared with attribute warning: detected read beyond size of field (2nd parameter); maybe use struct_group()? [-Werror]
>     __read_overflow2_field(q_size_field, size);
>     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> In function ‘fortify_memcpy_chk’,
>     inlined from ‘trace_event_raw_event_nvme_setup_cmd’ at drivers/nvme/host/./trace.h:47:1:
> ./include/linux/fortify-string.h:352:4: error: call to ‘__read_overflow2_field’ declared with attribute warning: detected read beyond size of field (2nd parameter); maybe use struct_group()? [-Werror]
>     __read_overflow2_field(q_size_field, size);
>     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> cdw10 metadata is 24 bytes, and we try to copy size of cdw10 metadata from
> nvme_command.common.cdw10 into that cdw10 metadata, but
> nvme_command.common.cdw10 is only 4 bytes in size.
> 
> Fix by making the trace metadata size as 4 bytes.
> 
> I find that this warning started first appearing from commit f68f2ff91512
> ("fortify: Detect struct member overflows in memcpy() at compile-time").

Did you test what the trace looks like afte this? We're losing valuable trace
data here. The field is supposed to get CDW's 10 - 15, so that's 24 bytes. I
don't know why it cares that the address of the field being read is only 4
bytes; we want everything that comes after it too.