From: Catalin Marinas Subject: Re: [PATCH] arm64: mm: check virtual addr in virt_to_page() if CONFIG_DEBUG_VIRTUAL=y Date: Tue, 16 Apr 2019 12:30:58 +0100 Message-ID: <20190416113057.GA28994@arrakis.emea.arm.com> References: <1555349796-9456-1-git-send-email-miles.chen@mediatek.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <1555349796-9456-1-git-send-email-miles.chen@mediatek.com> Sender: linux-kernel-owner@vger.kernel.org To: Miles Chen Cc: Will Deacon , linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, linux-mediatek@lists.infradead.org, wsd_upstream@mediatek.com, Ard Biesheuvel List-Id: linux-mediatek@lists.infradead.org On Tue, Apr 16, 2019 at 01:36:36AM +0800, Miles Chen wrote: > This change uses the original virt_to_page() (the one with __pa()) to > check the given virtual address if CONFIG_DEBUG_VIRTUAL=y. > > Recently, I worked on a bug: a driver passes a symbol address to > dma_map_single() and the virt_to_page() (called by dma_map_single()) > does not work for non-linear addresses after commit 9f2875912dac > ("arm64: mm: restrict virt_to_page() to the linear mapping"). > > I tried to trap the bug by enabling CONFIG_DEBUG_VIRTUAL but it > did not work - bacause the commit removes the __pa() from > virt_to_page() but CONFIG_DEBUG_VIRTUAL checks the virtual address > in __pa()/__virt_to_phys(). > > A simple solution is to use the original virt_to_page() > (the one with__pa()) if CONFIG_DEBUG_VIRTUAL=y. > > Signed-off-by: Miles Chen > Cc: Ard Biesheuvel > --- > arch/arm64/include/asm/memory.h | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/arm64/include/asm/memory.h b/arch/arm64/include/asm/memory.h > index 290195168bb3..2cb8248fa2c8 100644 > --- a/arch/arm64/include/asm/memory.h > +++ b/arch/arm64/include/asm/memory.h > @@ -302,7 +302,7 @@ static inline void *phys_to_virt(phys_addr_t x) > */ > #define ARCH_PFN_OFFSET ((unsigned long)PHYS_PFN_OFFSET) > > -#ifndef CONFIG_SPARSEMEM_VMEMMAP > +#if !defined(CONFIG_SPARSEMEM_VMEMMAP) || defined(CONFIG_DEBUG_VIRTUAL) > #define virt_to_page(kaddr) pfn_to_page(__pa(kaddr) >> PAGE_SHIFT) > #define _virt_addr_valid(kaddr) pfn_valid(__pa(kaddr) >> PAGE_SHIFT) > #else IIUC, this shouldn't change the behaviour of virt_addr_valid(). The patch looks fine to me. Acked-by: Catalin Marinas