From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paolo Bonzini Subject: Re: [RFC PATCH v2 14/32] x86: mm: Provide support to use memblock when spliting large pages Date: Fri, 17 Mar 2017 12:03:31 +0100 Message-ID: <78c99889-f175-f60f-716b-34a62203418a@redhat.com> References: <148846752022.2349.13667498174822419498.stgit@brijesh-build-machine> <148846771545.2349.9373586041426414252.stgit@brijesh-build-machine> <20170310110657.hophlog2juw5hpzz@pd.tnic> <20170316182836.tyvxoeq56thtc4pd@pd.tnic> <20170317101737.icdois7sdmtutt6b@pd.tnic> <20170317105610.musvo4baokgssvye@pd.tnic> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Cc: linux-efi@vger.kernel.org, Brijesh Singh , kvm@vger.kernel.org, rkrcmar@redhat.com, matt@codeblueprint.co.uk, linux-pci@vger.kernel.org, linus.walleij@linaro.org, gary.hook@amd.com, linux-mm@kvack.org, hpa@zytor.com, cl@linux.com, tglx@linutronix.de, aarcange@redhat.com, sfr@canb.auug.org.au, mchehab@kernel.org, simon.guinot@sequanux.org, bhe@redhat.com, xemul@parallels.com, joro@8bytes.org, x86@kernel.org, peterz@infradead.org, piotr.luc@intel.com, mingo@redhat.com, msalter@redhat.com, ross.zwisler@linux.intel.com, dyoung@redhat.com, thomas.lendacky@amd.com, jroedel@suse.de, keescook@chromium.org, arnd@arndb.de, toshi.kani@hpe.com, mathieu.desnoyers@efficios.com, luto@kernel.org, bhelgaas@google.com, dan.j.williams@intel.com, andriy.shevchenko@linux.intel.com, herbert@gondor.apana.org.au, paul.gortmaker@windriver.com, devel@linuxdriverproject.org To: Borislav Petkov Return-path: In-Reply-To: <20170317105610.musvo4baokgssvye@pd.tnic> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: driverdev-devel-bounces@linuxdriverproject.org Sender: "devel" List-Id: kvm.vger.kernel.org On 17/03/2017 11:56, Borislav Petkov wrote: >> Theoretically or practically? > In the sense, it needs to be tried first to see how ugly it can get. > >> It only looks at the E820 map, doesn't it? Why does it have to do >> anything with percpu memory areas? > That's irrelevant. What we want to do is take what's in init_mm.pgd and > modify it. And use the facilities in arch/x86/mm/init_{32,64}.c because > they already know about early/late pagetable pages allocation and they > deal with the kernel pagetable anyway. If it is possible to do it in a fairly hypervisor-independent manner, I'm all for it. That is, only by looking at AMD-specified CPUID leaves and at kernel ELF sections. Paolo > And *not* teach pageattr.c about memblock because that can be misused, > as tglx pointed out on IRC.