From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Daney Subject: Re: [PATCH v12 4/5] arm64, numa: Add NUMA support for arm64 platforms. Date: Mon, 29 Feb 2016 15:42:58 -0800 Message-ID: <56D4D782.3010503@gmail.com> References: <1456192703-2274-1-git-send-email-ddaney.cavm@gmail.com> <1456192703-2274-5-git-send-email-ddaney.cavm@gmail.com> <20160229173446.GG31343@rric.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20160229173446.GG31343-vWBEXY7mpu582hYKe6nXyg@public.gmane.org> Sender: linux-efi-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Robert Richter Cc: Will Deacon , linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org, Rob Herring , Pawel Moll , Mark Rutland , Ian Campbell , Kumar Gala , devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Ard Biesheuvel , Frank Rowand , Grant Likely , Catalin Marinas , Matt Fleming , linux-efi-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Ganapatrao Kulkarni , Robert Richter , linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, David Daney List-Id: devicetree@vger.kernel.org On 02/29/2016 09:34 AM, Robert Richter wrote: > On 22.02.16 17:58:22, David Daney wrote: >> From: Ganapatrao Kulkarni >> >> Attempt to get the memory and CPU NUMA node via of_numa. If that >> fails, default the dummy NUMA node and map all memory and CPUs to node >> 0. >> >> Tested-by: Shannon Zhao >> Reviewed-by: Robert Richter >> Signed-off-by: Ganapatrao Kulkarni >> Signed-off-by: David Daney >> --- >> arch/arm64/Kconfig | 26 +++ >> arch/arm64/include/asm/mmzone.h | 12 ++ >> arch/arm64/include/asm/numa.h | 45 +++++ >> arch/arm64/include/asm/topology.h | 10 + >> arch/arm64/kernel/pci.c | 10 + >> arch/arm64/kernel/setup.c | 4 + >> arch/arm64/kernel/smp.c | 4 + >> arch/arm64/mm/Makefile | 1 + >> arch/arm64/mm/init.c | 34 +++- >> arch/arm64/mm/mmu.c | 1 + >> arch/arm64/mm/numa.c | 403 ++++++++++++++++++++++++++++++++++++++ >> 11 files changed, 545 insertions(+), 5 deletions(-) >> create mode 100644 arch/arm64/include/asm/mmzone.h >> create mode 100644 arch/arm64/include/asm/numa.h >> create mode 100644 arch/arm64/mm/numa.c > >> +static int __init numa_init(int (*init_func)(void)) >> +{ >> + int ret; >> + >> + nodes_clear(numa_nodes_parsed); >> + nodes_clear(node_possible_map); >> + nodes_clear(node_online_map); >> + numa_free_distance(); ^^^^^^^^^^^^^ Cleanup for any previous numa_alloc_distance() >> + >> + ret = numa_alloc_distance(); >> + if (ret < 0) >> + return ret; > > If you move this before the remaining initializers, you will need to > clean this up on error. Yes, we do this. See above. > So better move it back after > numa_register_nodes() as it was in v10. This should work since > distances are used not earlier than numa is enabled. I moved it here for a reason. The init_func (of_numa_init() in this case) makes callbacks that use the numa_distance object. We need to allocate it before using it. Allocating it after calling the init_func() is too late. > > -Robert > >> + >> + ret = init_func(); >> + if (ret < 0) >> + return ret; >> + >> + if (nodes_empty(numa_nodes_parsed)) >> + return -EINVAL; >> + >> + ret = numa_register_nodes(); >> + if (ret < 0) >> + return ret; >> + >> + setup_node_to_cpumask_map(); >> + >> + /* init boot processor */ >> + cpu_to_node_map[0] = 0; >> + map_cpu_to_node(0, 0); >> + >> + return 0; >> +}