Linux Btrfs filesystem development
 help / color / mirror / Atom feed
From: Qu Wenruo <wqu@suse.com>
To: Li Zhang <zhanglikernel@gmail.com>, linux-btrfs@vger.kernel.org
Subject: Re: [PATCH V2] Make btrfs_prepare_device parallel during mkfs.btrfs
Date: Tue, 30 Aug 2022 05:55:44 +0800	[thread overview]
Message-ID: <6e4f65b7-2482-866b-fbfc-ea136910ccb7@suse.com> (raw)
In-Reply-To: <1661697000-18809-1-git-send-email-zhanglikernel@gmail.com>



On 2022/8/28 22:30, Li Zhang wrote:
> [enhancement]
> When a disk is formatted as btrfs, it calls
> btrfs_prepare_device for each device, which takes too much time.
> 
> [implementation]
> Put each btrfs_prepare_device into a thread,
> wait for the first thread to complete to mkfs.btrfs,
> and wait for other threads to complete before adding
> other devices to the file system.
> 
> [test]
> Using the btrfs-progs test case mkfs-tests, mkfs.btrfs works fine.
> 
> But I don't have an actual zoed device,
> so I don't know how much time it saves, If you guys
> have a way to test it, please let me know.
> 
> Signed-off-by: Li Zhang <zhanglikernel@gmail.com>

Reviewed-by: Qu Wenruo <wqu@suse.com>

Thanks,
Qu
> ---
> Issue: 496
> 
> V1:
> * Put btrfs_prepare_device into threads and make them parallel
> 
> V2:
> * Set the 4 variables used by btrfs_prepare_device as global variables.
> * Use pthread_mutex to ensure error messages are not messed up.
> * Correct the error message
> * Wait for all threads to exit in a loop
> 
>   mkfs/main.c | 132 +++++++++++++++++++++++++++++++++++++++++++-----------------
>   1 file changed, 95 insertions(+), 37 deletions(-)
> 
> diff --git a/mkfs/main.c b/mkfs/main.c
> index ce096d3..b111f12 100644
> --- a/mkfs/main.c
> +++ b/mkfs/main.c
> @@ -31,6 +31,7 @@
>   #include <uuid/uuid.h>
>   #include <ctype.h>
>   #include <blkid/blkid.h>
> +#include <pthread.h>
>   #include "kernel-shared/ctree.h"
>   #include "kernel-shared/disk-io.h"
>   #include "kernel-shared/free-space-tree.h"
> @@ -60,6 +61,20 @@ struct mkfs_allocation {
>   	u64 system;
>   };
>   
> +static bool zero_end;
> +static bool discard;
> +static bool zoned;
> +static int oflags;
> +
> +static pthread_mutex_t prepare_mutex;
> +
> +struct prepare_device_progress {
> +	char *file;
> +	u64 dev_block_count;
> +	u64 block_count;
> +	int ret;
> +};
> +
>   static int create_metadata_block_groups(struct btrfs_root *root, bool mixed,
>   				struct mkfs_allocation *allocation)
>   {
> @@ -969,6 +984,30 @@ fail:
>   	return ret;
>   }
>   
> +static void *prepare_one_dev(void *ctx)
> +{
> +	struct prepare_device_progress *prepare_ctx = ctx;
> +	int fd;
> +
> +	fd = open(prepare_ctx->file, oflags);
> +	if (fd < 0) {
> +		pthread_mutex_lock(&prepare_mutex);
> +		error("unable to open %s: %m", prepare_ctx->file);
> +		pthread_mutex_unlock(&prepare_mutex);
> +		prepare_ctx->ret = fd;
> +		return NULL;
> +	}
> +	prepare_ctx->ret = btrfs_prepare_device(fd,
> +		prepare_ctx->file, &prepare_ctx->dev_block_count,
> +		prepare_ctx->block_count,
> +		(bconf.verbose ? PREP_DEVICE_VERBOSE : 0) |
> +		(zero_end ? PREP_DEVICE_ZERO_END : 0) |
> +		(discard ? PREP_DEVICE_DISCARD : 0) |
> +		(zoned ? PREP_DEVICE_ZONED : 0));
> +	close(fd);
> +	return NULL;
> +}
> +
>   int BOX_MAIN(mkfs)(int argc, char **argv)
>   {
>   	char *file;
> @@ -984,7 +1023,7 @@ int BOX_MAIN(mkfs)(int argc, char **argv)
>   	u32 nodesize = 0;
>   	u32 sectorsize = 0;
>   	u32 stripesize = 4096;
> -	bool zero_end = true;
> +	zero_end = true;
>   	int fd = -1;
>   	int ret = 0;
>   	int close_ret;
> @@ -993,11 +1032,10 @@ int BOX_MAIN(mkfs)(int argc, char **argv)
>   	bool nodesize_forced = false;
>   	bool data_profile_opt = false;
>   	bool metadata_profile_opt = false;
> -	bool discard = true;
> +	discard = true;
>   	bool ssd = false;
> -	bool zoned = false;
> +	zoned = false;
>   	bool force_overwrite = false;
> -	int oflags;
>   	char *source_dir = NULL;
>   	bool source_dir_set = false;
>   	bool shrink_rootdir = false;
> @@ -1006,6 +1044,8 @@ int BOX_MAIN(mkfs)(int argc, char **argv)
>   	u64 shrink_size;
>   	int dev_cnt = 0;
>   	int saved_optind;
> +	pthread_t *t_prepare = NULL;
> +	struct prepare_device_progress *prepare_ctx = NULL;
>   	char fs_uuid[BTRFS_UUID_UNPARSED_SIZE] = { 0 };
>   	u64 features = BTRFS_MKFS_DEFAULT_FEATURES;
>   	u64 runtime_features = BTRFS_MKFS_DEFAULT_RUNTIME_FEATURES;
> @@ -1428,29 +1468,49 @@ int BOX_MAIN(mkfs)(int argc, char **argv)
>   		goto error;
>   	}
>   
> -	dev_cnt--;
> +	t_prepare = calloc(dev_cnt, sizeof(*t_prepare));
> +	prepare_ctx = calloc(dev_cnt, sizeof(*prepare_ctx));
> +
> +	if (!t_prepare || !prepare_ctx) {
> +		error("unable to alloc thread for preparing dev");
> +		goto error;
> +	}
>   
> +	pthread_mutex_init(&prepare_mutex, NULL);
> +	zero_end = zero_end;
> +	discard = discard;
> +	zoned = zoned;
>   	oflags = O_RDWR;
> -	if (zoned && zoned_model(file) == ZONED_HOST_MANAGED)
> -		oflags |= O_DIRECT;
> +	for (i = 0; i < dev_cnt; i++) {
> +		if (zoned && zoned_model(argv[optind + i - 1]) ==
> +			ZONED_HOST_MANAGED) {
> +			oflags |= O_DIRECT;
> +			break;
> +		}
> +	}
> +	for (i = 0; i < dev_cnt; i++) {
> +		prepare_ctx[i].file = argv[optind + i - 1];
> +		prepare_ctx[i].block_count = block_count;
> +		prepare_ctx[i].dev_block_count = block_count;
> +		ret = pthread_create(&t_prepare[i], NULL,
> +			prepare_one_dev, &prepare_ctx[i]);
> +		if (ret) {
> +			error("create thread for prepare devices failed, errno:%d", ret);
> +			goto error;
> +		}
> +	}
> +	for (i = 0; i < dev_cnt; i++)
> +		pthread_join(t_prepare[i], NULL);
> +	ret = prepare_ctx[0].ret;
>   
> -	/*
> -	 * Open without O_EXCL so that the problem should not occur by the
> -	 * following operation in kernel:
> -	 * (btrfs_register_one_device() fails if O_EXCL is on)
> -	 */
> -	fd = open(file, oflags);
> -	if (fd < 0) {
> -		error("unable to open %s: %m", file);
> +	if (ret) {
> +		error("unable prepare device:%s.\n", prepare_ctx[0].file);
>   		goto error;
>   	}
> -	ret = btrfs_prepare_device(fd, file, &dev_block_count, block_count,
> -			(zero_end ? PREP_DEVICE_ZERO_END : 0) |
> -			(discard ? PREP_DEVICE_DISCARD : 0) |
> -			(bconf.verbose ? PREP_DEVICE_VERBOSE : 0) |
> -			(zoned ? PREP_DEVICE_ZONED : 0));
> -	if (ret)
> -		goto error;
> +
> +	dev_cnt--;
> +	fd = open(file, oflags);
> +	dev_block_count = prepare_ctx[0].dev_block_count;
>   	if (block_count && block_count > dev_block_count) {
>   		error("%s is smaller than requested size, expected %llu, found %llu",
>   		      file, (unsigned long long)block_count,
> @@ -1459,7 +1519,7 @@ int BOX_MAIN(mkfs)(int argc, char **argv)
>   	}
>   
>   	/* To create the first block group and chunk 0 in make_btrfs */
> -	system_group_size = zoned ?  zone_size(file) : BTRFS_MKFS_SYSTEM_GROUP_SIZE;
> +	system_group_size = zoned ? zone_size(file) : BTRFS_MKFS_SYSTEM_GROUP_SIZE;
>   	if (dev_block_count < system_group_size) {
>   		error("device is too small to make filesystem, must be at least %llu",
>   				(unsigned long long)system_group_size);
> @@ -1558,14 +1618,10 @@ int BOX_MAIN(mkfs)(int argc, char **argv)
>   		goto raid_groups;
>   
>   	while (dev_cnt-- > 0) {
> +		int dev_index = argc - saved_optind - dev_cnt - 1;
>   		file = argv[optind++];
>   
> -		/*
> -		 * open without O_EXCL so that the problem should not
> -		 * occur by the following processing.
> -		 * (btrfs_register_one_device() fails if O_EXCL is on)
> -		 */
> -		fd = open(file, O_RDWR);
> +		fd = open(file, oflags);
>   		if (fd < 0) {
>   			error("unable to open %s: %m", file);
>   			goto error;
> @@ -1578,13 +1634,11 @@ int BOX_MAIN(mkfs)(int argc, char **argv)
>   			close(fd);
>   			continue;
>   		}
> -		ret = btrfs_prepare_device(fd, file, &dev_block_count,
> -				block_count,
> -				(bconf.verbose ? PREP_DEVICE_VERBOSE : 0) |
> -				(zero_end ? PREP_DEVICE_ZERO_END : 0) |
> -				(discard ? PREP_DEVICE_DISCARD : 0) |
> -				(zoned ? PREP_DEVICE_ZONED : 0));
> -		if (ret) {
> +		dev_block_count = prepare_ctx[dev_index]
> +			.dev_block_count;
> +
> +		if (prepare_ctx[dev_index].ret) {
> +			error("unable prepare device:%s.\n", prepare_ctx[dev_index].file);
>   			goto error;
>   		}
>   
> @@ -1763,12 +1817,16 @@ out:
>   
>   	btrfs_close_all_devices();
>   	free(label);
> -
> +	free(t_prepare);
> +	free(prepare_ctx);
>   	return !!ret;
> +
>   error:
>   	if (fd > 0)
>   		close(fd);
>   
> +	free(t_prepare);
> +	free(prepare_ctx);
>   	free(label);
>   	exit(1);
>   success:

  reply	other threads:[~2022-08-29 21:56 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-08-28 14:30 [PATCH V2] Make btrfs_prepare_device parallel during mkfs.btrfs Li Zhang
2022-08-29 21:55 ` Qu Wenruo [this message]
2022-08-30 17:30 ` David Sterba

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=6e4f65b7-2482-866b-fbfc-ea136910ccb7@suse.com \
    --to=wqu@suse.com \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=zhanglikernel@gmail.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