From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-3.8 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS,URIBL_BLOCKED autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 506DDC67863 for ; Fri, 19 Oct 2018 01:58:40 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id E5AB22087A for ; Fri, 19 Oct 2018 01:58:39 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org E5AB22087A Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=linux-foundation.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726678AbeJSKCa (ORCPT ); Fri, 19 Oct 2018 06:02:30 -0400 Received: from mail.linuxfoundation.org ([140.211.169.12]:37970 "EHLO mail.linuxfoundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726506AbeJSKC3 (ORCPT ); Fri, 19 Oct 2018 06:02:29 -0400 Received: from localhost.localdomain (c-24-4-154-175.hsd1.ca.comcast.net [24.4.154.175]) by mail.linuxfoundation.org (Postfix) with ESMTPSA id 03082B8E; Fri, 19 Oct 2018 01:58:35 +0000 (UTC) Date: Thu, 18 Oct 2018 18:58:34 -0700 From: Andrew Morton To: Mike Rapoport Cc: Richard Henderson , Ivan Kokshaysky , Michal Hocko , linux-alpha , linux-mm , lkml Subject: Re: [RESEND PATCH v2] alpha: switch to NO_BOOTMEM Message-Id: <20181018185834.a3925ae72b9d7bbf43b1b8a3@linux-foundation.org> In-Reply-To: <1535952894-10967-1-git-send-email-rppt@linux.vnet.ibm.com> References: <1535952894-10967-1-git-send-email-rppt@linux.vnet.ibm.com> X-Mailer: Sylpheed 3.5.1 (GTK+ 2.24.31; x86_64-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org No reviews or acks for this one yet? From: Mike Rapoport Subject: alpha: switch to NO_BOOTMEM Replace bootmem allocator with memblock and enable use of NO_BOOTMEM like on most other architectures. Alpha gets the description of the physical memory from the firmware as an array of memory clusters. Each cluster that is not reserved by the firmware is added to memblock.memory. Once the memblock.memory is set up, we reserve the kernel and initrd pages with memblock reserve. Since we don't need the bootmem bitmap anymore, the code that finds an appropriate place is removed. The conversion does not take care of NUMA support which is marked broken for more than 10 years now. Link: http://lkml.kernel.org/r/1535952894-10967-1-git-send-email-rppt@linux.vnet.ibm.com Signed-off-by: Mike Rapoport Cc: Richard Henderson Cc: Ivan Kokshaysky Cc: Michal Hocko Signed-off-by: Andrew Morton --- --- a/arch/alpha/Kconfig~alpha-switch-to-no_bootmem +++ a/arch/alpha/Kconfig @@ -31,6 +31,8 @@ config ALPHA select ODD_RT_SIGACTION select OLD_SIGSUSPEND select CPU_NO_EFFICIENT_FFS if !ALPHA_EV67 + select HAVE_MEMBLOCK + select NO_BOOTMEM help The Alpha is a 64-bit general-purpose processor designed and marketed by the Digital Equipment Corporation of blessed memory, --- a/arch/alpha/kernel/core_irongate.c~alpha-switch-to-no_bootmem +++ a/arch/alpha/kernel/core_irongate.c @@ -21,6 +21,7 @@ #include #include #include +#include #include #include @@ -241,8 +242,7 @@ albacore_init_arch(void) size / 1024); } #endif - reserve_bootmem_node(NODE_DATA(0), pci_mem, memtop - - pci_mem, BOOTMEM_DEFAULT); + memblock_reserve(pci_mem, memtop - pci_mem); printk("irongate_init_arch: temporarily reserving " "region %08lx-%08lx for PCI\n", pci_mem, memtop - 1); } --- a/arch/alpha/kernel/setup.c~alpha-switch-to-no_bootmem +++ a/arch/alpha/kernel/setup.c @@ -30,6 +30,7 @@ #include #include #include +#include #include #include #include @@ -312,9 +313,7 @@ setup_memory(void *kernel_end) { struct memclust_struct * cluster; struct memdesc_struct * memdesc; - unsigned long start_kernel_pfn, end_kernel_pfn; - unsigned long bootmap_size, bootmap_pages, bootmap_start; - unsigned long start, end; + unsigned long kernel_size; unsigned long i; /* Find free clusters, and init and free the bootmem accordingly. */ @@ -322,6 +321,8 @@ setup_memory(void *kernel_end) (hwrpb->mddt_offset + (unsigned long) hwrpb); for_each_mem_cluster(memdesc, cluster, i) { + unsigned long end; + printk("memcluster %lu, usage %01lx, start %8lu, end %8lu\n", i, cluster->usage, cluster->start_pfn, cluster->start_pfn + cluster->numpages); @@ -335,6 +336,9 @@ setup_memory(void *kernel_end) end = cluster->start_pfn + cluster->numpages; if (end > max_low_pfn) max_low_pfn = end; + + memblock_add(PFN_PHYS(cluster->start_pfn), + cluster->numpages << PAGE_SHIFT); } /* @@ -363,87 +367,9 @@ setup_memory(void *kernel_end) max_low_pfn = mem_size_limit; } - /* Find the bounds of kernel memory. */ - start_kernel_pfn = PFN_DOWN(KERNEL_START_PHYS); - end_kernel_pfn = PFN_UP(virt_to_phys(kernel_end)); - bootmap_start = -1; - - try_again: - if (max_low_pfn <= end_kernel_pfn) - panic("not enough memory to boot"); - - /* We need to know how many physically contiguous pages - we'll need for the bootmap. */ - bootmap_pages = bootmem_bootmap_pages(max_low_pfn); - - /* Now find a good region where to allocate the bootmap. */ - for_each_mem_cluster(memdesc, cluster, i) { - if (cluster->usage & 3) - continue; - - start = cluster->start_pfn; - end = start + cluster->numpages; - if (start >= max_low_pfn) - continue; - if (end > max_low_pfn) - end = max_low_pfn; - if (start < start_kernel_pfn) { - if (end > end_kernel_pfn - && end - end_kernel_pfn >= bootmap_pages) { - bootmap_start = end_kernel_pfn; - break; - } else if (end > start_kernel_pfn) - end = start_kernel_pfn; - } else if (start < end_kernel_pfn) - start = end_kernel_pfn; - if (end - start >= bootmap_pages) { - bootmap_start = start; - break; - } - } - - if (bootmap_start == ~0UL) { - max_low_pfn >>= 1; - goto try_again; - } - - /* Allocate the bootmap and mark the whole MM as reserved. */ - bootmap_size = init_bootmem(bootmap_start, max_low_pfn); - - /* Mark the free regions. */ - for_each_mem_cluster(memdesc, cluster, i) { - if (cluster->usage & 3) - continue; - - start = cluster->start_pfn; - end = cluster->start_pfn + cluster->numpages; - if (start >= max_low_pfn) - continue; - if (end > max_low_pfn) - end = max_low_pfn; - if (start < start_kernel_pfn) { - if (end > end_kernel_pfn) { - free_bootmem(PFN_PHYS(start), - (PFN_PHYS(start_kernel_pfn) - - PFN_PHYS(start))); - printk("freeing pages %ld:%ld\n", - start, start_kernel_pfn); - start = end_kernel_pfn; - } else if (end > start_kernel_pfn) - end = start_kernel_pfn; - } else if (start < end_kernel_pfn) - start = end_kernel_pfn; - if (start >= end) - continue; - - free_bootmem(PFN_PHYS(start), PFN_PHYS(end) - PFN_PHYS(start)); - printk("freeing pages %ld:%ld\n", start, end); - } - - /* Reserve the bootmap memory. */ - reserve_bootmem(PFN_PHYS(bootmap_start), bootmap_size, - BOOTMEM_DEFAULT); - printk("reserving pages %ld:%ld\n", bootmap_start, bootmap_start+PFN_UP(bootmap_size)); + /* Reserve the kernel memory. */ + kernel_size = virt_to_phys(kernel_end) - KERNEL_START_PHYS; + memblock_reserve(KERNEL_START_PHYS, kernel_size); #ifdef CONFIG_BLK_DEV_INITRD initrd_start = INITRD_START; @@ -459,8 +385,8 @@ setup_memory(void *kernel_end) initrd_end, phys_to_virt(PFN_PHYS(max_low_pfn))); } else { - reserve_bootmem(virt_to_phys((void *)initrd_start), - INITRD_SIZE, BOOTMEM_DEFAULT); + memblock_reserve(virt_to_phys((void *)initrd_start), + INITRD_SIZE); } } #endif /* CONFIG_BLK_DEV_INITRD */ --- a/arch/alpha/mm/numa.c~alpha-switch-to-no_bootmem +++ a/arch/alpha/mm/numa.c @@ -11,6 +11,7 @@ #include #include #include +#include #include #include #include @@ -59,12 +60,10 @@ setup_memory_node(int nid, void *kernel_ struct memclust_struct * cluster; struct memdesc_struct * memdesc; unsigned long start_kernel_pfn, end_kernel_pfn; - unsigned long bootmap_size, bootmap_pages, bootmap_start; unsigned long start, end; unsigned long node_pfn_start, node_pfn_end; unsigned long node_min_pfn, node_max_pfn; int i; - unsigned long node_datasz = PFN_UP(sizeof(pg_data_t)); int show_init = 0; /* Find the bounds of current node */ @@ -134,24 +133,14 @@ setup_memory_node(int nid, void *kernel_ /* Cute trick to make sure our local node data is on local memory */ node_data[nid] = (pg_data_t *)(__va(node_min_pfn << PAGE_SHIFT)); #endif - /* Quasi-mark the pg_data_t as in-use */ - node_min_pfn += node_datasz; - if (node_min_pfn >= node_max_pfn) { - printk(" not enough mem to reserve NODE_DATA"); - return; - } - NODE_DATA(nid)->bdata = &bootmem_node_data[nid]; - printk(" Detected node memory: start %8lu, end %8lu\n", node_min_pfn, node_max_pfn); DBGDCONT(" DISCONTIG: node_data[%d] is at 0x%p\n", nid, NODE_DATA(nid)); - DBGDCONT(" DISCONTIG: NODE_DATA(%d)->bdata is at 0x%p\n", nid, NODE_DATA(nid)->bdata); /* Find the bounds of kernel memory. */ start_kernel_pfn = PFN_DOWN(KERNEL_START_PHYS); end_kernel_pfn = PFN_UP(virt_to_phys(kernel_end)); - bootmap_start = -1; if (!nid && (node_max_pfn < end_kernel_pfn || node_min_pfn > start_kernel_pfn)) panic("kernel loaded out of ram"); @@ -161,89 +150,11 @@ setup_memory_node(int nid, void *kernel_ has much larger alignment than 8Mb, so it's safe. */ node_min_pfn &= ~((1UL << (MAX_ORDER-1))-1); - /* We need to know how many physically contiguous pages - we'll need for the bootmap. */ - bootmap_pages = bootmem_bootmap_pages(node_max_pfn-node_min_pfn); - - /* Now find a good region where to allocate the bootmap. */ - for_each_mem_cluster(memdesc, cluster, i) { - if (cluster->usage & 3) - continue; - - start = cluster->start_pfn; - end = start + cluster->numpages; - - if (start >= node_max_pfn || end <= node_min_pfn) - continue; - - if (end > node_max_pfn) - end = node_max_pfn; - if (start < node_min_pfn) - start = node_min_pfn; - - if (start < start_kernel_pfn) { - if (end > end_kernel_pfn - && end - end_kernel_pfn >= bootmap_pages) { - bootmap_start = end_kernel_pfn; - break; - } else if (end > start_kernel_pfn) - end = start_kernel_pfn; - } else if (start < end_kernel_pfn) - start = end_kernel_pfn; - if (end - start >= bootmap_pages) { - bootmap_start = start; - break; - } - } - - if (bootmap_start == -1) - panic("couldn't find a contiguous place for the bootmap"); - - /* Allocate the bootmap and mark the whole MM as reserved. */ - bootmap_size = init_bootmem_node(NODE_DATA(nid), bootmap_start, - node_min_pfn, node_max_pfn); - DBGDCONT(" bootmap_start %lu, bootmap_size %lu, bootmap_pages %lu\n", - bootmap_start, bootmap_size, bootmap_pages); - - /* Mark the free regions. */ - for_each_mem_cluster(memdesc, cluster, i) { - if (cluster->usage & 3) - continue; - - start = cluster->start_pfn; - end = cluster->start_pfn + cluster->numpages; - - if (start >= node_max_pfn || end <= node_min_pfn) - continue; + memblock_add(PFN_PHYS(node_min_pfn), + (node_max_pfn - node_min_pfn) << PAGE_SHIFT); - if (end > node_max_pfn) - end = node_max_pfn; - if (start < node_min_pfn) - start = node_min_pfn; - - if (start < start_kernel_pfn) { - if (end > end_kernel_pfn) { - free_bootmem_node(NODE_DATA(nid), PFN_PHYS(start), - (PFN_PHYS(start_kernel_pfn) - - PFN_PHYS(start))); - printk(" freeing pages %ld:%ld\n", - start, start_kernel_pfn); - start = end_kernel_pfn; - } else if (end > start_kernel_pfn) - end = start_kernel_pfn; - } else if (start < end_kernel_pfn) - start = end_kernel_pfn; - if (start >= end) - continue; - - free_bootmem_node(NODE_DATA(nid), PFN_PHYS(start), PFN_PHYS(end) - PFN_PHYS(start)); - printk(" freeing pages %ld:%ld\n", start, end); - } - - /* Reserve the bootmap memory. */ - reserve_bootmem_node(NODE_DATA(nid), PFN_PHYS(bootmap_start), - bootmap_size, BOOTMEM_DEFAULT); - printk(" reserving pages %ld:%ld\n", bootmap_start, bootmap_start+PFN_UP(bootmap_size)); + NODE_DATA(nid)->node_start_pfn = node_min_pfn; + NODE_DATA(nid)->node_present_pages = node_max_pfn - node_min_pfn; node_set_online(nid); } @@ -251,6 +162,7 @@ setup_memory_node(int nid, void *kernel_ void __init setup_memory(void *kernel_end) { + unsigned long kernel_size; int nid; show_mem_layout(); @@ -262,6 +174,9 @@ setup_memory(void *kernel_end) for (nid = 0; nid < MAX_NUMNODES; nid++) setup_memory_node(nid, kernel_end); + kernel_size = virt_to_phys(kernel_end) - KERNEL_START_PHYS; + memblock_reserve(KERNEL_START_PHYS, kernel_size); + #ifdef CONFIG_BLK_DEV_INITRD initrd_start = INITRD_START; if (initrd_start) { @@ -279,9 +194,8 @@ setup_memory(void *kernel_end) phys_to_virt(PFN_PHYS(max_low_pfn))); } else { nid = kvaddr_to_nid(initrd_start); - reserve_bootmem_node(NODE_DATA(nid), - virt_to_phys((void *)initrd_start), - INITRD_SIZE, BOOTMEM_DEFAULT); + memblock_reserve(virt_to_phys((void *)initrd_start), + INITRD_SIZE); } } #endif /* CONFIG_BLK_DEV_INITRD */ @@ -303,9 +217,8 @@ void __init paging_init(void) dma_local_pfn = virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT; for_each_online_node(nid) { - bootmem_data_t *bdata = &bootmem_node_data[nid]; - unsigned long start_pfn = bdata->node_min_pfn; - unsigned long end_pfn = bdata->node_low_pfn; + unsigned long start_pfn = NODE_DATA(nid)->node_start_pfn; + unsigned long end_pfn = start_pfn + NODE_DATA(nid)->node_present_pages; if (dma_local_pfn >= end_pfn - start_pfn) zones_size[ZONE_DMA] = end_pfn - start_pfn; _