Hi Stefano,
On 15 Jan 2018 19:51, "Stefano Stabellini" <sstabell...@kernel.org> wrote: On Mon, 15 Jan 2018, Jan Beulich wrote: > >>> On 13.01.18 at 07:21, <julien.gr...@linaro.org> wrote: > > On 01/12/2018 11:40 PM, Stefano Stabellini wrote: > >> handles can theoretically be NULL, check for it explicitly before > >> dereferencing it. > > > > I doubt handles could be NULL if LocateHandle succeed. This seems to be > > confirmed by the spec (Page 208 in UEFI spec 2.7). > > > > So I am not entirely convince we should add yet another check in the > > code. An ASSERT might be better. > > Indeed if there is a platform where NULL is coming back in the > success case, that platform should be named as a justification > in the commit message. Otherwise I don't see the value of this > change. Truthfully, it is mostly to silence Coverity. We can all appreciate when static analysts cannot find defects in the code. Can't you mark as such in coverity? I mean, I see limited value to always try to fix in Xen something that we can consider as a false positive. An ASSERT should satisfy it as well. Cheers,
_______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
