On Tue, 17 Jul 2012, Pekka Enberg wrote: > Well, even SLUB checks for !name in mainline so that's definitely > worth including unconditionally. Furthermore, the size related checks > certainly make sense and I don't see any harm in having them as well. There is a WARN_ON() there and then it returns NULL!!! Crazy. Causes a NULL pointer dereference later in the caller? > As for "in_interrupt()", I really don't see the point in keeping that > around. We could push it down to mm/slab.c in "__kmem_cache_create()" > if we wanted to. Ok we could do that but I guess we are in the discussion of how much checking should be done for a production kernel. I think these checks are way out of hand. We cannot afford to consistently check parameters to all kernel functions in production. We will only do these checks in a select manner if these values could result in serious difficult to debug problems. The checks in slab look like debugging code that someone needed for a specific debugging scenario. I can understand that we would keep that in for development but not for production. Maybe I am a bit biased but my prod kernels need to have minimal memory footprint due to excessive code size causing regressions. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>