On Mon, 2020-11-23 at 09:33 +1100, Finn Thain wrote:
On Sun, 22 Nov 2020, Joe Perches wrote:
But provably correct conversions IMO _should_ be done and IMO churn considerations should generally have less importance.
[]
Moreover, the patch review workload for skilled humans is being generated by the automation, which is completely backwards: the machine is supposed to be helping.
Which is why the provably correct matters.
coccinelle transforms can be, but are not necessarily, provably correct.
The _show transforms done via the sysfs_emit_dev.cocci script are correct as in commit aa838896d87a ("drivers core: Use sysfs_emit and sysfs_emit_at for show(device *...) functions")
Worthwhile? A different question, but I think yes as it reduces the overall question space of the existing other sprintf overrun possibilities.