This feedback applies to all patches in this series. On Thu, May 20, 2021, Joe Richey wrote: > From: Joe Richey <joerichey@xxxxxxxxxx> > > A previous patch Heh, I think it goes without saying that the code was introduced by a previous patch, unless you've invented a time machine, in which case we should talk... > [1] used the BIT() macro to define the > KVM_DIRTY_GFN_F_* constants in KVM's UAPI header. > > This macro is defined in the kernel but not in the UAPI headers. > > [1] https://patchwork.kernel.org/patch/11854393 Linking to the patch isn't helpful/desirable in this case because it doesn't provide any info about when the commit actually landed in the kernel. And depending on the whims of the maintainer, what was posted may not exactly match the code that was commited. What you want is a Fixes: tag that points at the offending commit. The Fixes: tag will also get the fix picked up for stable kernels, though in KVM we often explicitly add "Cc: stable@xxxxxxxxxxxxxxx" (though IIRC tglx prefers not to have the explicit Cc). Anyways, the changelog can simply be something like: Replace BIT() in KVM's UAPI header with an open coded equivalent. BIT() is not defined in the UAPI headers and its usage may cause userspace build errors. Fixes: fb04a1eddb1a ("KVM: X86: Implement ring-based dirty memory tracking") Signed-off-by: Joe Richey <joerichey@xxxxxxxxxx>
