Hi Dan,
Understood — I'll drop this patch. Thanks.
regards, Sun Jian
On Mon, Jan 5, 2026 at 7:55 PM Dan Carpenter dan.carpenter@linaro.org wrote:
This W=1 string truncation warnings always seems like a pointless thing. dmesg output is really only intended for developers. I don't even know how to look at the dmesg on my phone or kindle. Who cares if the last character in a really long device name is missing? I have looked at a lot of stack traces and I have never once been stymied because one character was missing in a really long device name.
And also these are 90% false positives. We just add bounds checking all of our output to prevent memory corruption and not because we think that we'll actually hit the boundaries. And the GCC is not able to analyze these in a sensible way, it's just going based on the variable types.
This patch makes the code worse (more complicated).
To recap: 1: It's warning about a non-issue. No one cares about truncated output. 2: The warning is implemented poorly. High false positive ratio. 3: The fix makes the code worse
Just fix the tool instead of making the code worse for no reason.
regards, dan carpenter