linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Lu Fengqi <lufq.fnst@cn.fujitsu.com>
To: Qu Wenruo <quwenruo.btrfs@gmx.com>
Cc: <linux-btrfs@vger.kernel.org>, <dsterba@suse.cz>
Subject: Re: [PATCH 1/2] btrfs-progs: mkfs: Enhance minimal device size calculation to fix mkfs failure on small file
Date: Thu, 12 Oct 2017 17:23:14 +0800	[thread overview]
Message-ID: <20171012092314.GB495@fnst.localdomain> (raw)
In-Reply-To: <20171012072858.19005-1-quwenruo.btrfs@gmx.com>

On Thu, Oct 12, 2017 at 12:28:57AM -0700, Qu Wenruo wrote:
>Since commit c11e36a29e84 ("Btrfs-progs: Do not force mixed block group
>creation unless '-M' option is specified"), mkfs no longer use mixed
>block group unless specified manually.
>
>This breaks the minimal device size calculation, which only considered
>mixed block group use case.
>
>This patch enhances minimal device size calculation for mkfs, by using
>different minimal stripe length (calculated from code) for different
>profiles, and use them to calculate minimal device size.
>
>Reported-by: Wesley Aptekar-Cassels <W.Aptekar@gmail.com>
>Fixes: c11e36a29e84 ("Btrfs-progs: Do not force mixed block group creation unless '-M' option is specified")
>Signed-off-by: Qu Wenruo <quwenruo.btrfs@gmx.com>
>---
> mkfs/common.c | 74 +++++++++++++++++++++++++++++++++++++++++++++++++++--------
> mkfs/common.h |  7 +++---
> mkfs/main.c   | 12 ++++++----
> 3 files changed, 77 insertions(+), 16 deletions(-)
>
>diff --git a/mkfs/common.c b/mkfs/common.c
>index c9ce10d46a9d..64461daf3917 100644
>--- a/mkfs/common.c
>+++ b/mkfs/common.c
>@@ -438,12 +438,6 @@ out:
> 	return ret;
> }
> 
>-u64 btrfs_min_dev_size(u32 nodesize)
>-{
>-	return 2 * (BTRFS_MKFS_SYSTEM_GROUP_SIZE +
>-		    btrfs_min_global_blk_rsv_size(nodesize));
>-}
>-
> /*
>  * Btrfs minimum size calculation is complicated, it should include at least:
>  * 1. system group size
>@@ -454,11 +448,71 @@ u64 btrfs_min_dev_size(u32 nodesize)
>  * To avoid the overkill calculation, (system group + global block rsv) * 2
>  * for *EACH* device should be good enough.
>  */
>-u64 btrfs_min_global_blk_rsv_size(u32 nodesize)
>+static u64 btrfs_min_global_blk_rsv_size(u32 nodesize)
> {
> 	return (u64)nodesize << 10;
> }
> 
>+u64 btrfs_min_dev_size(u32 nodesize, int mixed, u64 meta_profile,
>+		       u64 data_profile)
>+{
>+	u64 reserved = 0;
>+	u64 meta_size;
>+	u64 data_size;
>+
>+	if (mixed)
>+		return 2 * (BTRFS_MKFS_SYSTEM_GROUP_SIZE +
>+			    btrfs_min_global_blk_rsv_size(nodesize));
>+
>+	/*
>+	 * minimal size calculation is complex due to several factors:
>+	 * 1) Temporary chunk resue
>+	 *    If specified chunk profile is SINGLE, we can resue
>+	 *    temporary chunks, no need to alloc new chunks.
>+	 *
>+	 * 2) Different minimal chunk size for different profile
>+	 *    For initial sys chunk, chunk size is fixed to 4M.
>+	 *    For single profile, minimal chunk size is 8M for all.
>+	 *    For other profiles, minimal chunk and stripe size
>+	 *    differs from 8M to 64M.
>+	 *
>+	 * To calculate it a little easier, here we assume we don't
>+	 * reuse any temporary chunk, and calculate the size all
>+	 * by ourselves.
>+	 *
>+	 * Temporary chunks sizes are always fixed:
>+	 * One initial sys chunk, one SINGLE meta, and one SINGLE data.
>+	 * The latter two are all 8M, accroding to @calc_size of
>+	 * btrfs_alloc_chunk().
>+	 */
>+	reserved += BTRFS_MKFS_SYSTEM_GROUP_SIZE + SZ_8M * 2;
>+
>+	/*
>+	 * For real chunks, we need to refer different sizes:
>+	 * For SINGLE, it's still fixed to 8M (@calc_size).
>+	 * For other profiles, refer to max(@min_stripe_size, @calc_size).
>+	 *
>+	 * And use the stripe size to calculate its physical used space.
>+	 */
>+	if (meta_profile & BTRFS_BLOCK_GROUP_PROFILE_MASK)
>+		meta_size = SZ_8M + SZ_32M;
>+	else
>+		meta_size = SZ_8M + SZ_8M;
>+	/* Only metadata put 2 stripes into one disk */
>+	if (meta_profile & BTRFS_BLOCK_GROUP_DUP)
>+		meta_size *= 2;
>+	reserved += meta_size;
>+
>+	if (data_profile & BTRFS_BLOCK_GROUP_PROFILE_MASK)
>+		data_size = SZ_64M;
>+	else
>+		data_size = SZ_8M;
>+	if (data_profile & BTRFS_BLOCK_GROUP_DUP)
>+		data_size *= 2;
>+	reserved += data_size;
>+	return reserved;
>+}
>+
> #define isoctal(c)	(((c) & ~7) == '0')
> 
> static inline void translate(char *f, char *t)
>@@ -698,7 +752,8 @@ int is_vol_small(const char *file)
> 	}
> }
> 
>-int test_minimum_size(const char *file, u32 nodesize)
>+int test_minimum_size(const char *file, u32 nodesize, int mixed,
>+		      u64 meta_profile, u64 data_profile)
> {
> 	int fd;
> 	struct stat statbuf;
>@@ -710,7 +765,8 @@ int test_minimum_size(const char *file, u32 nodesize)
> 		close(fd);
> 		return -errno;
> 	}
>-	if (btrfs_device_size(fd, &statbuf) < btrfs_min_dev_size(nodesize)) {
>+	if (btrfs_device_size(fd, &statbuf) <
>+	    btrfs_min_dev_size(nodesize, mixed, meta_profile, data_profile)) {
> 		close(fd);
> 		return 1;
> 	}
>diff --git a/mkfs/common.h b/mkfs/common.h
>index dee0ea9740e0..0d1aa5c17c03 100644
>--- a/mkfs/common.h
>+++ b/mkfs/common.h
>@@ -66,9 +66,10 @@ struct btrfs_mkfs_config {
> };
> 
> int make_btrfs(int fd, struct btrfs_mkfs_config *cfg);
>-u64 btrfs_min_dev_size(u32 nodesize);
>-u64 btrfs_min_global_blk_rsv_size(u32 nodesize);
>-int test_minimum_size(const char *file, u32 nodesize);
>+u64 btrfs_min_dev_size(u32 nodesize, int mixed, u64 meta_profile,
>+		       u64 data_profile);
>+int test_minimum_size(const char *file, u32 nodesize, int mixed,
>+		      u64 meta_profile, u64 data_profile);
> int is_vol_small(const char *file);
> int test_num_disk_vs_raid(u64 metadata_profile, u64 data_profile,
> 	u64 dev_cnt, int mixed, int ssd);
>diff --git a/mkfs/main.c b/mkfs/main.c
>index 1b4cabc1ef90..527c508f64a5 100644
>--- a/mkfs/main.c
>+++ b/mkfs/main.c
>@@ -1436,6 +1436,7 @@ int main(int argc, char **argv)
> 	u64 num_of_meta_chunks = 0;
> 	u64 size_of_data = 0;
> 	u64 source_dir_size = 0;
>+	u64 min_dev_size;
> 	int dev_cnt = 0;
> 	int saved_optind;
> 	char fs_uuid[BTRFS_UUID_UNPARSED_SIZE] = { 0 };
>@@ -1646,19 +1647,22 @@ int main(int argc, char **argv)
> 		goto error;
> 	}
> 
>+	min_dev_size = btrfs_min_dev_size(nodesize, mixed, metadata_profile,
>+					  data_profile);
> 	/* Check device/block_count after the nodesize is determined */
>-	if (block_count && block_count < btrfs_min_dev_size(nodesize)) {
>+	if (block_count && block_count < min_dev_size) {
> 		error("size %llu is too small to make a usable filesystem",
> 			block_count);
> 		error("minimum size for btrfs filesystem is %llu",
>-			btrfs_min_dev_size(nodesize));
>+			min_dev_size);
> 		goto error;
> 	}
> 	for (i = saved_optind; i < saved_optind + dev_cnt; i++) {
> 		char *path;
> 
> 		path = argv[i];
>-		ret = test_minimum_size(path, nodesize);
>+		ret = test_minimum_size(path, nodesize, mixed, metadata_profile,
>+					data_profile);

As mentioned in the offline discussion, I think test_minimum_size don't
need so many parameters, min_dev_size is enough.

The rest looks good for me.

Reviewed-by: Lu Fengqi <lufq.fnst@cn.fujitsu.com>

-- 
Thanks,
Lu

> 		if (ret < 0) {
> 			error("failed to check size for %s: %s",
> 				path, strerror(-ret));
>@@ -1668,7 +1672,7 @@ int main(int argc, char **argv)
> 			error("'%s' is too small to make a usable filesystem",
> 				path);
> 			error("minimum size for each btrfs device is %llu",
>-				btrfs_min_dev_size(nodesize));
>+				min_dev_size);
> 			goto error;
> 		}
> 	}
>-- 
>2.14.2
>
>--
>To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
>



      parent reply	other threads:[~2017-10-12  9:23 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-12  7:28 [PATCH 1/2] btrfs-progs: mkfs: Enhance minimal device size calculation to fix mkfs failure on small file Qu Wenruo
2017-10-12  7:28 ` [PATCH 2/2] btrfs-progs: test/mkfs: Test if the minimal device size is valid Qu Wenruo
2017-10-12  8:24   ` Lu Fengqi
2017-10-12  8:41     ` Qu Wenruo
2017-10-12  8:49   ` [PATCH v2 " Qu Wenruo
2017-10-12  9:23 ` Lu Fengqi [this message]

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=20171012092314.GB495@fnst.localdomain \
    --to=lufq.fnst@cn.fujitsu.com \
    --cc=dsterba@suse.cz \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=quwenruo.btrfs@gmx.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: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).