Messages in this thread | | | Date | Wed, 17 Apr 2019 14:13:09 +0200 | From | Nicholas Mc Guire <> | Subject | Re: [PATCH V2] ARM: mvebu: at least report the kzalloc failure |
| |
On Wed, Apr 17, 2019 at 02:07:44PM +0200, Gregory CLEMENT wrote: > Hi Nicholas, > > Nicholas Mc Guire <der.herr@hofr.at> writes: > > > On Tue, Apr 16, 2019 at 03:39:57PM +0200, Andrew Lunn wrote: > >> On Tue, Apr 16, 2019 at 05:56:31AM +0200, Nicholas Mc Guire wrote: > >> > >> > Note that this will trigger a checkpatch WARNING > >> > "WARNING: Possible unnecessary 'out of memory' message" > >> > but comparing the oops with an without the one-line pr_err I would > >> > argue that it makes sense to include it: > >> > >> Hi Nicholas > >> > >> It might be worth adding this as a comment, so that newbies don't > >> submit patches removing the pr_err() because of the checkpatch > >> warning. > >> > > hmm... I think if we start doing that we would make quite a mess of > > documentation in the kernel. Also note its a warning stating "possible > > unneceessary" - so I would not see the necessity. > > > > At most I would include a note on this in the commit message so that > > anyone checking the origin would see that this is intenttional - assuming > > that people modifying code would be using git blame to locate the > > origin of any code... > > Don't bother to send a new version I don't attempt to take this > patch. As you pointed it is very unlikely that we get an error so early > during the boot for a very small amount of memory. > > If it happened then we have serious trouble and the message provided by > the BUG() call will be more than enough. > yup - its a corner case - I'm trying to filter out those cases that are actually in __init function returning void - as those cases are, it seems, are generally cases where k{m,z}allocs will not have explicit checking.
thx! hofrat
| |