From: Paul Mackerras Date: Mon, 29 Apr 2019 02:05:55 +0000 Subject: Re: [PATCH v8 05/20] KVM: PPC: Book3S HV: Remove pmd_is_leaf() Message-Id: <20190429020555.GB11154@blackberry> List-Id: References: <20190403141627.11664-1-steven.price@arm.com> <20190403141627.11664-6-steven.price@arm.com> In-Reply-To: <20190403141627.11664-6-steven.price@arm.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Steven Price Cc: linux-mm@kvack.org, Andy Lutomirski , Ard Biesheuvel , Arnd Bergmann , Borislav Petkov , Catalin Marinas , Dave Hansen , Ingo Molnar , James Morse , =?iso-8859-1?B?Suly9G1l?= Glisse , Peter Zijlstra , Thomas Gleixner , Will Deacon , x86@kernel.org, "H. Peter Anvin" , linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, Mark Rutland , "Liang, Kan" , Andrew Morton , Benjamin Herrenschmidt , Michael Ellerman , kvm-ppc@vger.kernel.org, linuxppc-dev@lists.ozlabs.org On Wed, Apr 03, 2019 at 03:16:12PM +0100, Steven Price wrote: > Since pmd_large() is now always available, pmd_is_leaf() is redundant. > Replace all uses with calls to pmd_large(). NAK. I don't want to do this, because pmd_is_leaf() is purely about the guest page tables (the "partition-scoped" radix tree which specifies the guest physical to host physical translation), not about anything to do with the Linux process page tables. The guest page tables have the same format as the Linux process page tables, but they are managed separately. If it makes things clearer, I could rename it to "guest_pmd_is_leaf()" or something similar. Paul.