From patchwork Wed Mar 9 13:16:32 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Hannes Reinecke X-Patchwork-Id: 8546211 Return-Path: X-Original-To: patchwork-linux-block@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork2.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork2.web.kernel.org (Postfix) with ESMTP id 1DC29C0553 for ; Wed, 9 Mar 2016 13:17:08 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 1C710202AE for ; Wed, 9 Mar 2016 13:17:07 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 0826720218 for ; Wed, 9 Mar 2016 13:17:06 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753640AbcCINQy (ORCPT ); Wed, 9 Mar 2016 08:16:54 -0500 Received: from mx2.suse.de ([195.135.220.15]:44231 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753642AbcCINQn (ORCPT ); Wed, 9 Mar 2016 08:16:43 -0500 X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (charybdis-ext.suse.de [195.135.220.254]) by mx2.suse.de (Postfix) with ESMTP id C2C89AC16; Wed, 9 Mar 2016 13:16:40 +0000 (UTC) From: Hannes Reinecke To: Jens Axboe Cc: Christoph Hellwig , "Martin K. Petersen" , Jeff Mahoney , linux-block@vger.kernel.org, linux-scsi@vger.kernel.org, linux-fsdevel@vger.kernel.org, Hannes Reinecke , Hannes Reinecke Subject: [PATCH 4/4] brd: thin provisioning support Date: Wed, 9 Mar 2016 14:16:32 +0100 Message-Id: <1457529392-53970-5-git-send-email-hare@suse.de> X-Mailer: git-send-email 1.8.5.6 In-Reply-To: <1457529392-53970-1-git-send-email-hare@suse.de> References: <1457529392-53970-1-git-send-email-hare@suse.de> Sender: linux-block-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-block@vger.kernel.org X-Spam-Status: No, score=-6.9 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=ham version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Implement module parameter 'rd_mem_size' to restrict the overall allocated memory size and 'rd_lowat_thresh' to add a low water mark signalling. Signed-off-by: Hannes Reinecke --- drivers/block/brd.c | 53 +++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 53 insertions(+) diff --git a/drivers/block/brd.c b/drivers/block/brd.c index cb27190..a8848c3 100644 --- a/drivers/block/brd.c +++ b/drivers/block/brd.c @@ -30,6 +30,13 @@ #define PAGE_SECTORS (1 << PAGE_SECTORS_SHIFT) /* + * Thin provisioning support + */ +static unsigned int rd_total_pages; +static int rd_mem_size = CONFIG_BLK_DEV_RAM_SIZE * CONFIG_BLK_DEV_RAM_COUNT; +static int rd_lowat_thresh = CONFIG_BLK_DEV_RAM_SIZE * CONFIG_BLK_DEV_RAM_COUNT; + +/* * Each block ramdisk device has a radix_tree brd_pages of pages that stores * the pages containing the block device's contents. A brd page's ->index is * its offset in PAGE_SIZE units. This is similar to, but in no way connected @@ -44,6 +51,12 @@ struct brd_device { struct list_head brd_list; /* + * Thin provisioning support + */ + unsigned int disk_events; + unsigned int pending_events; + + /* * Backing store of pages and lock to protect it. This is the contents * of the block device. */ @@ -91,11 +104,23 @@ static struct page *brd_insert_page(struct brd_device *brd, sector_t sector) pgoff_t idx; struct page *page; gfp_t gfp_flags; + unsigned int rd_max_pages = rd_mem_size >> (PAGE_CACHE_SHIFT - 10); + unsigned int rd_lowat_pages = rd_lowat_thresh >> (PAGE_CACHE_SHIFT - 10); page = brd_lookup_page(brd, sector); if (page) return page; + if (rd_total_pages >= rd_max_pages) + return NULL; + + if (rd_total_pages >= rd_lowat_pages && + !(brd->disk_events & DISK_EVENT_LOWAT)) { + brd->pending_events |= DISK_EVENT_LOWAT; + brd->disk_events |= DISK_EVENT_LOWAT; + disk_clear_events(brd->brd_disk, DISK_EVENT_LOWAT); + } + /* * Must use NOIO because we don't want to recurse back into the * block or filesystem layers from page reclaim. @@ -127,6 +152,7 @@ static struct page *brd_insert_page(struct brd_device *brd, sector_t sector) BUG_ON(!page); BUG_ON(page->index != idx); } + rd_total_pages++; spin_unlock(&brd->brd_lock); radix_tree_preload_end(); @@ -138,10 +164,16 @@ static void brd_free_page(struct brd_device *brd, sector_t sector) { struct page *page; pgoff_t idx; + unsigned int rd_lowat_pages = rd_lowat_thresh >> (PAGE_CACHE_SHIFT - 10); spin_lock(&brd->brd_lock); idx = sector >> PAGE_SECTORS_SHIFT; page = radix_tree_delete(&brd->brd_pages, idx); + rd_total_pages--; + if (rd_total_pages < rd_lowat_pages) { + brd->disk_events &= ~DISK_EVENT_LOWAT; + brd->pending_events &= ~DISK_EVENT_LOWAT; + } spin_unlock(&brd->brd_lock); if (page) __free_page(page); @@ -434,11 +466,22 @@ static int brd_ioctl(struct block_device *bdev, fmode_t mode, return error; } +static unsigned int brd_check_events(struct gendisk *disk, unsigned int mask) +{ + struct brd_device *brd = disk->private_data; + unsigned int pending_events = brd->pending_events & mask; + + brd->pending_events &= ~mask; + + return pending_events; +} + static const struct block_device_operations brd_fops = { .owner = THIS_MODULE, .rw_page = brd_rw_page, .ioctl = brd_ioctl, .direct_access = brd_direct_access, + .check_events = brd_check_events, }; /* @@ -456,6 +499,12 @@ static int max_part = 1; module_param(max_part, int, S_IRUGO); MODULE_PARM_DESC(max_part, "Num Minors to reserve between devices"); +module_param(rd_mem_size, int, S_IRUGO); +MODULE_PARM_DESC(rd_mem_size, "Maximal memory size in kbytes to allocate"); + +module_param(rd_lowat_thresh, int, S_IRUGO); +MODULE_PARM_DESC(rd_lowat_thresh, "Low water mark in kbytes for memory allocation"); + MODULE_LICENSE("GPL"); MODULE_ALIAS_BLOCKDEV_MAJOR(RAMDISK_MAJOR); MODULE_ALIAS("rd"); @@ -519,6 +568,8 @@ static struct brd_device *brd_alloc(int i) disk->private_data = brd; disk->queue = brd->brd_queue; disk->flags = GENHD_FL_EXT_DEVT; + disk->events = DISK_EVENT_LOWAT; + disk->async_events = DISK_EVENT_LOWAT; sprintf(disk->disk_name, "ram%d", i); set_capacity(disk, rd_size * 2); @@ -607,6 +658,8 @@ static int __init brd_init(void) if (register_blkdev(RAMDISK_MAJOR, "ramdisk")) return -EIO; + rd_total_pages = 0; + if (unlikely(!max_part)) max_part = 1;