From mboxrd@z Thu Jan 1 00:00:00 1970 From: mark.rutland@arm.com (Mark Rutland) Date: Mon, 25 Jul 2016 15:22:42 +0100 Subject: [PATCH 1/2] arm64: mm: make create_mapping_late() non-allocating In-Reply-To: <3728BEC2-B1C8-40A5-9E0C-E27D56149BD4@linaro.org> References: <1469208745-6693-1-git-send-email-ard.biesheuvel@linaro.org> <1469208745-6693-2-git-send-email-ard.biesheuvel@linaro.org> <20160725134550.GA14465@leverpostej> <3728BEC2-B1C8-40A5-9E0C-E27D56149BD4@linaro.org> Message-ID: <20160725142241.GC14465@leverpostej> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Mon, Jul 25, 2016 at 04:10:50PM +0200, Ard Biesheuvel wrote: > > On 25 jul. 2016, at 15:46, Mark Rutland wrote: > > How about we drop the __init marker from create_mapping_noalloc(), and > > update the callers in mark_rodata_ro() to use that, so that we can drop > > create_mapping_late() entirely. > > > > Other than the __init marker and name, create_mapping_late() and > > create_mapping_noalloc() would be identical after this change, and the > > naming of create_mapping_noalloc() does better describe what we're > > trying to achieve in mark_rodata_ro(). > > Not entirely. The _late() variant should map down to pages if the > first mapping was mapped down to pages as well. The _noalloc() variant > may end attempting to merge regions Ah, I was comparing against mainline rather than the arm64 for-next/core branch with the recent rework. Sorry for the noise. Given that, FWIW: Acked-by: Mark Rutland Mark.