From mboxrd@z Thu Jan 1 00:00:00 1970 From: takahiro.akashi@linaro.org (AKASHI Takahiro) Date: Thu, 7 Jul 2016 10:17:47 +0900 Subject: [PATCH v21 1/8] arm64: kdump: reserve memory for crash dump kernel In-Reply-To: <2543175.KeBui093Sr@hactar> References: <20160706075226.27609-1-takahiro.akashi@linaro.org> <20160706075226.27609-2-takahiro.akashi@linaro.org> <2543175.KeBui093Sr@hactar> Message-ID: <20160707011745.GX20774@linaro.org> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Jul 06, 2016 at 03:11:45PM -0300, Thiago Jung Bauermann wrote: > Am Mittwoch, 06 Juli 2016, 16:52:19 schrieb AKASHI Takahiro: > > On the startup of primary kernel, the memory region used by crash dump > > kernel must be specified by "crashkernel=" kernel parameter. > > reserve_crashkernel() will allocate and reserve the region for later use. > > > > User space tools, like kexec-tools, will be able to find that region > > marked as "Crash kernel" in /proc/iomem. > > On powerpc, userspace tools get everything from the device tree (exposed to > userspace in /proc/device-tree/), not /proc/iomem. I started my kdump patches with arm (not arm64) implementation, and so /proc/iomem is used for consistency. Since my kexec-tools (for arm64) have to access /proc/iomem to identify usable memory regions on *UEFI* systems anyway, I didn't think it was odd. (Please note that UEFI won't expose memory regions information via a device tree, but via ACPI table.) > In the case of the crashkernel reserved region, that information is in > /chosen/linux,crashkernel-base and /chosen/linux,crashkernel-size. > > Either way is fine I think. I'm just mentioning this for reference in case > you want the ARM implementation to be closer to another arch which is also > based on the device tree. Adding those properties to a device tree is quite easy (and it won't conflict with /proc/iomem), so I'd like to defer to arm64 maintainers. Thanks, -Takahiro AKASHI > -- > []'s > Thiago Jung Bauermann > IBM Linux Technology Center >