From: "Leizhen (ThunderTown)" <thunder.leizhen@huawei.com> To: Will Deacon <will.deacon@arm.com> Cc: Catalin Marinas <catalin.marinas@arm.com>, linux-arm-kernel <linux-arm-kernel@lists.infradead.org>, linux-kernel <linux-kernel@vger.kernel.org>, Rob Herring <robh+dt@kernel.org>, "Frank Rowand" <frowand.list@gmail.com>, devicetree <devicetree@vger.kernel.org>, Zefan Li <lizefan@huawei.com>, Xinwei Hu <huxinwei@huawei.com>, Tianhong Ding <dingtianhong@huawei.com>, Hanjun Guo <guohanjun@huawei.com> Subject: Re: [PATCH v7 11/14] arm64/numa: support HAVE_MEMORYLESS_NODES Date: Sat, 27 Aug 2016 19:05:17 +0800 [thread overview] Message-ID: <57C173ED.60501@huawei.com> (raw) In-Reply-To: <20160826154356.GJ30302@arm.com> On 2016/8/26 23:43, Will Deacon wrote: > On Wed, Aug 24, 2016 at 03:44:50PM +0800, Zhen Lei wrote: >> Some numa nodes may have no memory. For example: >> 1. cpu0 on node0 >> 2. cpu1 on node1 >> 3. device0 access the momory from node0 and node1 take the same time. >> >> So, we can not simply classify device0 to node0 or node1, but we can >> define a node2 which distances to node0 and node1 are the same. >> >> Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com> >> --- >> arch/arm64/Kconfig | 4 ++++ >> arch/arm64/kernel/smp.c | 1 + >> arch/arm64/mm/numa.c | 43 +++++++++++++++++++++++++++++++++++++++++-- >> 3 files changed, 46 insertions(+), 2 deletions(-) >> >> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig >> index 2815af6..3a2b6ed 100644 >> --- a/arch/arm64/Kconfig >> +++ b/arch/arm64/Kconfig >> @@ -611,6 +611,10 @@ config NEED_PER_CPU_EMBED_FIRST_CHUNK >> def_bool y >> depends on NUMA >> >> +config HAVE_MEMORYLESS_NODES >> + def_bool y >> + depends on NUMA >> + >> source kernel/Kconfig.preempt >> source kernel/Kconfig.hz >> >> diff --git a/arch/arm64/kernel/smp.c b/arch/arm64/kernel/smp.c >> index d93d433..4879085 100644 >> --- a/arch/arm64/kernel/smp.c >> +++ b/arch/arm64/kernel/smp.c >> @@ -619,6 +619,7 @@ static void __init of_parse_and_init_cpus(void) >> } >> >> bootcpu_valid = true; >> + early_map_cpu_to_node(0, of_node_to_nid(dn)); > > This seems unrelated? I will get off my work soon. Maybe I need put it into patch 12. > >> /* >> * cpu_logical_map has already been >> diff --git a/arch/arm64/mm/numa.c b/arch/arm64/mm/numa.c >> index 6853db7..114180f 100644 >> --- a/arch/arm64/mm/numa.c >> +++ b/arch/arm64/mm/numa.c >> @@ -129,6 +129,14 @@ void __init early_map_cpu_to_node(unsigned int cpu, int nid) >> nid = 0; >> >> cpu_to_node_map[cpu] = nid; >> + >> + /* >> + * We should set the numa node of cpu0 as soon as possible, because it >> + * has already been set up online before. cpu_to_node(0) will soon be >> + * called. >> + */ >> + if (!cpu) >> + set_cpu_numa_node(cpu, nid); > > Likewise. > >> } >> >> #ifdef CONFIG_HAVE_SETUP_PER_CPU_AREA >> @@ -211,6 +219,35 @@ int __init numa_add_memblk(int nid, u64 start, u64 end) >> return ret; >> } >> >> +static u64 __init alloc_node_data_from_nearest_node(int nid, const size_t size) >> +{ >> + int i, best_nid, distance; >> + u64 pa; >> + DECLARE_BITMAP(nodes_map, MAX_NUMNODES); >> + >> + bitmap_zero(nodes_map, MAX_NUMNODES); >> + bitmap_set(nodes_map, nid, 1); >> + >> +find_nearest_node: >> + best_nid = NUMA_NO_NODE; >> + distance = INT_MAX; >> + >> + for_each_clear_bit(i, nodes_map, MAX_NUMNODES) >> + if (numa_distance[nid][i] < distance) { >> + best_nid = i; >> + distance = numa_distance[nid][i]; >> + } >> + >> + pa = memblock_alloc_nid(size, SMP_CACHE_BYTES, best_nid); >> + if (!pa) { >> + BUG_ON(best_nid == NUMA_NO_NODE); >> + bitmap_set(nodes_map, best_nid, 1); >> + goto find_nearest_node; >> + } >> + >> + return pa; >> +} >> + >> /** >> * Initialize NODE_DATA for a node on the local memory >> */ >> @@ -224,7 +261,9 @@ static void __init setup_node_data(int nid, u64 start_pfn, u64 end_pfn) >> pr_info("Initmem setup node %d [mem %#010Lx-%#010Lx]\n", >> nid, start_pfn << PAGE_SHIFT, (end_pfn << PAGE_SHIFT) - 1); >> >> - nd_pa = memblock_alloc_try_nid(nd_size, SMP_CACHE_BYTES, nid); >> + nd_pa = memblock_alloc_nid(nd_size, SMP_CACHE_BYTES, nid); >> + if (!nd_pa) >> + nd_pa = alloc_node_data_from_nearest_node(nid, nd_size); > > Why not add memblock_alloc_near_nid to the core code, and make it do > what you need there? I'm thinking about it next week. But some ARCHs like X86/IA64 have their own implementation. > > Will > > . >
WARNING: multiple messages have this Message-ID (diff)
From: thunder.leizhen@huawei.com (Leizhen (ThunderTown)) To: linux-arm-kernel@lists.infradead.org Subject: [PATCH v7 11/14] arm64/numa: support HAVE_MEMORYLESS_NODES Date: Sat, 27 Aug 2016 19:05:17 +0800 [thread overview] Message-ID: <57C173ED.60501@huawei.com> (raw) In-Reply-To: <20160826154356.GJ30302@arm.com> On 2016/8/26 23:43, Will Deacon wrote: > On Wed, Aug 24, 2016 at 03:44:50PM +0800, Zhen Lei wrote: >> Some numa nodes may have no memory. For example: >> 1. cpu0 on node0 >> 2. cpu1 on node1 >> 3. device0 access the momory from node0 and node1 take the same time. >> >> So, we can not simply classify device0 to node0 or node1, but we can >> define a node2 which distances to node0 and node1 are the same. >> >> Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com> >> --- >> arch/arm64/Kconfig | 4 ++++ >> arch/arm64/kernel/smp.c | 1 + >> arch/arm64/mm/numa.c | 43 +++++++++++++++++++++++++++++++++++++++++-- >> 3 files changed, 46 insertions(+), 2 deletions(-) >> >> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig >> index 2815af6..3a2b6ed 100644 >> --- a/arch/arm64/Kconfig >> +++ b/arch/arm64/Kconfig >> @@ -611,6 +611,10 @@ config NEED_PER_CPU_EMBED_FIRST_CHUNK >> def_bool y >> depends on NUMA >> >> +config HAVE_MEMORYLESS_NODES >> + def_bool y >> + depends on NUMA >> + >> source kernel/Kconfig.preempt >> source kernel/Kconfig.hz >> >> diff --git a/arch/arm64/kernel/smp.c b/arch/arm64/kernel/smp.c >> index d93d433..4879085 100644 >> --- a/arch/arm64/kernel/smp.c >> +++ b/arch/arm64/kernel/smp.c >> @@ -619,6 +619,7 @@ static void __init of_parse_and_init_cpus(void) >> } >> >> bootcpu_valid = true; >> + early_map_cpu_to_node(0, of_node_to_nid(dn)); > > This seems unrelated? I will get off my work soon. Maybe I need put it into patch 12. > >> /* >> * cpu_logical_map has already been >> diff --git a/arch/arm64/mm/numa.c b/arch/arm64/mm/numa.c >> index 6853db7..114180f 100644 >> --- a/arch/arm64/mm/numa.c >> +++ b/arch/arm64/mm/numa.c >> @@ -129,6 +129,14 @@ void __init early_map_cpu_to_node(unsigned int cpu, int nid) >> nid = 0; >> >> cpu_to_node_map[cpu] = nid; >> + >> + /* >> + * We should set the numa node of cpu0 as soon as possible, because it >> + * has already been set up online before. cpu_to_node(0) will soon be >> + * called. >> + */ >> + if (!cpu) >> + set_cpu_numa_node(cpu, nid); > > Likewise. > >> } >> >> #ifdef CONFIG_HAVE_SETUP_PER_CPU_AREA >> @@ -211,6 +219,35 @@ int __init numa_add_memblk(int nid, u64 start, u64 end) >> return ret; >> } >> >> +static u64 __init alloc_node_data_from_nearest_node(int nid, const size_t size) >> +{ >> + int i, best_nid, distance; >> + u64 pa; >> + DECLARE_BITMAP(nodes_map, MAX_NUMNODES); >> + >> + bitmap_zero(nodes_map, MAX_NUMNODES); >> + bitmap_set(nodes_map, nid, 1); >> + >> +find_nearest_node: >> + best_nid = NUMA_NO_NODE; >> + distance = INT_MAX; >> + >> + for_each_clear_bit(i, nodes_map, MAX_NUMNODES) >> + if (numa_distance[nid][i] < distance) { >> + best_nid = i; >> + distance = numa_distance[nid][i]; >> + } >> + >> + pa = memblock_alloc_nid(size, SMP_CACHE_BYTES, best_nid); >> + if (!pa) { >> + BUG_ON(best_nid == NUMA_NO_NODE); >> + bitmap_set(nodes_map, best_nid, 1); >> + goto find_nearest_node; >> + } >> + >> + return pa; >> +} >> + >> /** >> * Initialize NODE_DATA for a node on the local memory >> */ >> @@ -224,7 +261,9 @@ static void __init setup_node_data(int nid, u64 start_pfn, u64 end_pfn) >> pr_info("Initmem setup node %d [mem %#010Lx-%#010Lx]\n", >> nid, start_pfn << PAGE_SHIFT, (end_pfn << PAGE_SHIFT) - 1); >> >> - nd_pa = memblock_alloc_try_nid(nd_size, SMP_CACHE_BYTES, nid); >> + nd_pa = memblock_alloc_nid(nd_size, SMP_CACHE_BYTES, nid); >> + if (!nd_pa) >> + nd_pa = alloc_node_data_from_nearest_node(nid, nd_size); > > Why not add memblock_alloc_near_nid to the core code, and make it do > what you need there? I'm thinking about it next week. But some ARCHs like X86/IA64 have their own implementation. > > Will > > . >
next prev parent reply other threads:[~2016-08-27 11:06 UTC|newest] Thread overview: 92+ messages / expand[flat|nested] mbox.gz Atom feed top 2016-08-24 7:44 [PATCH v7 00/14] fix some type infos and bugs for arm64/of numa Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-24 7:44 ` [PATCH v7 01/14] of/numa: remove a duplicated pr_debug information Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-24 7:44 ` [PATCH v7 02/14] of/numa: fix a memory@ node can only contains one memory block Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-24 7:44 ` [PATCH v7 03/14] arm64/numa: add nid check for " Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-26 12:39 ` Will Deacon 2016-08-26 12:39 ` Will Deacon 2016-08-26 12:39 ` Will Deacon 2016-08-27 8:02 ` Leizhen (ThunderTown) 2016-08-27 8:02 ` Leizhen (ThunderTown) 2016-08-24 7:44 ` [PATCH v7 04/14] of/numa: remove a duplicated warning Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-24 7:44 ` [PATCH v7 05/14] arm64/numa: avoid inconsistent information to be printed Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-26 12:47 ` Will Deacon 2016-08-26 12:47 ` Will Deacon 2016-08-26 12:47 ` Will Deacon 2016-08-27 8:54 ` Leizhen (ThunderTown) 2016-08-27 8:54 ` Leizhen (ThunderTown) 2016-08-27 8:54 ` Leizhen (ThunderTown) 2016-08-30 17:51 ` Will Deacon 2016-08-30 17:51 ` Will Deacon 2016-08-30 17:51 ` Will Deacon 2016-08-31 2:29 ` Leizhen (ThunderTown) 2016-08-31 2:29 ` Leizhen (ThunderTown) 2016-08-31 2:29 ` Leizhen (ThunderTown) 2016-08-24 7:44 ` [PATCH v7 06/14] of_numa: Use of_get_next_parent to simplify code Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-24 7:44 ` [PATCH v7 07/14] of_numa: Use pr_fmt() Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-24 7:44 ` [PATCH v7 08/14] arm64: numa: " Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-26 12:54 ` Will Deacon 2016-08-26 12:54 ` Will Deacon 2016-08-26 12:54 ` Will Deacon 2016-08-27 9:14 ` Leizhen (ThunderTown) 2016-08-27 9:14 ` Leizhen (ThunderTown) 2016-08-27 9:14 ` Leizhen (ThunderTown) 2016-08-24 7:44 ` [PATCH v7 09/14] arm64/numa: support HAVE_SETUP_PER_CPU_AREA Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-26 13:28 ` Will Deacon 2016-08-26 13:28 ` Will Deacon 2016-08-26 13:28 ` Will Deacon 2016-08-27 10:06 ` Leizhen (ThunderTown) 2016-08-27 10:06 ` Leizhen (ThunderTown) 2016-08-27 10:06 ` Leizhen (ThunderTown) 2016-08-24 7:44 ` [PATCH v7 10/14] arm64/numa: define numa_distance as array to simplify code Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-26 15:29 ` Will Deacon 2016-08-26 15:29 ` Will Deacon 2016-08-27 10:29 ` Leizhen (ThunderTown) 2016-08-27 10:29 ` Leizhen (ThunderTown) 2016-08-27 10:29 ` Leizhen (ThunderTown) 2016-08-24 7:44 ` [PATCH v7 11/14] arm64/numa: support HAVE_MEMORYLESS_NODES Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-26 15:43 ` Will Deacon 2016-08-26 15:43 ` Will Deacon 2016-08-27 11:05 ` Leizhen (ThunderTown) [this message] 2016-08-27 11:05 ` Leizhen (ThunderTown) 2016-08-29 3:15 ` Leizhen (ThunderTown) 2016-08-29 3:15 ` Leizhen (ThunderTown) 2016-08-29 3:15 ` Leizhen (ThunderTown) 2016-08-24 7:44 ` [PATCH v7 12/14] arm64/numa: remove the limitation that cpu0 must bind to node0 Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-26 15:49 ` Will Deacon 2016-08-26 15:49 ` Will Deacon 2016-08-26 15:49 ` Will Deacon 2016-08-29 6:55 ` Leizhen (ThunderTown) 2016-08-29 6:55 ` Leizhen (ThunderTown) 2016-08-29 6:55 ` Leizhen (ThunderTown) 2016-08-24 7:44 ` [PATCH v7 13/14] of/numa: remove the constraint on the distances of node pairs Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-24 7:44 ` [PATCH v7 14/14] Documentation: " Zhen Lei 2016-08-24 7:44 ` Zhen Lei 2016-08-26 15:35 ` Will Deacon 2016-08-26 15:35 ` Will Deacon 2016-08-27 10:44 ` Leizhen (ThunderTown) 2016-08-27 10:44 ` Leizhen (ThunderTown) 2016-08-27 10:44 ` Leizhen (ThunderTown) 2016-08-30 17:55 ` Will Deacon 2016-08-30 17:55 ` Will Deacon 2016-08-30 17:55 ` Will Deacon 2016-08-31 2:46 ` Leizhen (ThunderTown) 2016-08-31 2:46 ` Leizhen (ThunderTown) 2016-08-31 2:46 ` Leizhen (ThunderTown)
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=57C173ED.60501@huawei.com \ --to=thunder.leizhen@huawei.com \ --cc=catalin.marinas@arm.com \ --cc=devicetree@vger.kernel.org \ --cc=dingtianhong@huawei.com \ --cc=frowand.list@gmail.com \ --cc=guohanjun@huawei.com \ --cc=huxinwei@huawei.com \ --cc=linux-arm-kernel@lists.infradead.org \ --cc=linux-kernel@vger.kernel.org \ --cc=lizefan@huawei.com \ --cc=robh+dt@kernel.org \ --cc=will.deacon@arm.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.