diff mbox

[06/27] xfs_scrub: create an abstraction for a block device

Message ID 151520352472.2027.8375038586822233334.stgit@magnolia (mailing list archive)
State Superseded
Headers show

Commit Message

Darrick J. Wong Jan. 6, 2018, 1:52 a.m. UTC 
From: Darrick J. Wong <darrick.wong@oracle.com>

Create an abstraction to handle all of our low level disk operations.
We'll eventually use it to bind to a fs mount point and block device.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 scrub/Makefile |    2 +
 scrub/disk.c   |  164 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 scrub/disk.h   |   39 +++++++++++++
 3 files changed, 205 insertions(+)
 create mode 100644 scrub/disk.c
 create mode 100644 scrub/disk.h



--
To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Eric Sandeen Jan. 11, 2018, 11:24 p.m. UTC | #1 
On 1/5/18 7:52 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>

...

> +/*
> + * Disk Abstraction
> + *
> + * These routines help us to discover the geometry of a block device,
> + * estimate the amount of concurrent IOs that we can send to it, and
> + * abstract the process of performing read verification of disk blocks.
> + */
> +
> +/* Figure out how many disk heads are available. */
> +static unsigned int
> +__disk_heads(
> +	struct disk		*disk)
> +{
> +	int			iomin;
> +	int			ioopt;
> +	unsigned short		rot;
> +	int			error;
> +
> +	/* If it's not a block device, throw all the CPUs at it. */
> +	if (!S_ISBLK(disk->d_sb.st_mode))
> +		return nproc;
> +
> +	/* Non-rotational device?  Throw all the CPUs. */
> +	rot = 1;
> +	error = ioctl(disk->d_fd, BLKROTATIONAL, &rot);
> +	if (error == 0 && rot == 0)
> +		return nproc;

I needed

+#ifndef BLKROTATIONAL
+#define BLKROTATIONAL _IO(0x12,126)
+#endif

to make this compile on my not /that/ ancient (?) rhel6 box ;)

-Eric
--
To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Darrick J. Wong Jan. 11, 2018, 11:59 p.m. UTC | #2 
On Thu, Jan 11, 2018 at 05:24:58PM -0600, Eric Sandeen wrote:
> On 1/5/18 7:52 PM, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> ...
> 
> > +/*
> > + * Disk Abstraction
> > + *
> > + * These routines help us to discover the geometry of a block device,
> > + * estimate the amount of concurrent IOs that we can send to it, and
> > + * abstract the process of performing read verification of disk blocks.
> > + */
> > +
> > +/* Figure out how many disk heads are available. */
> > +static unsigned int
> > +__disk_heads(
> > +	struct disk		*disk)
> > +{
> > +	int			iomin;
> > +	int			ioopt;
> > +	unsigned short		rot;
> > +	int			error;
> > +
> > +	/* If it's not a block device, throw all the CPUs at it. */
> > +	if (!S_ISBLK(disk->d_sb.st_mode))
> > +		return nproc;
> > +
> > +	/* Non-rotational device?  Throw all the CPUs. */
> > +	rot = 1;
> > +	error = ioctl(disk->d_fd, BLKROTATIONAL, &rot);
> > +	if (error == 0 && rot == 0)
> > +		return nproc;
> 
> I needed
> 
> +#ifndef BLKROTATIONAL
> +#define BLKROTATIONAL _IO(0x12,126)
> +#endif
> 
> to make this compile on my not /that/ ancient (?) rhel6 box ;)

Hmm... well, since I don't see backporting xfs kernel scrub to 2.6.32
maybe xfsprogs' build system should just turn off xfs_scrub on old
systems?

In any case, I #ifdef BLKROTATIONAL'd out the entire clause.

--D

> -Eric
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Eric Sandeen Jan. 12, 2018, 12:04 a.m. UTC | #3 
On 1/11/18 5:59 PM, Darrick J. Wong wrote:
> On Thu, Jan 11, 2018 at 05:24:58PM -0600, Eric Sandeen wrote:
...

>>> +	/* Non-rotational device?  Throw all the CPUs. */
>>> +	rot = 1;
>>> +	error = ioctl(disk->d_fd, BLKROTATIONAL, &rot);
>>> +	if (error == 0 && rot == 0)
>>> +		return nproc;
>>
>> I needed
>>
>> +#ifndef BLKROTATIONAL
>> +#define BLKROTATIONAL _IO(0x12,126)
>> +#endif
>>
>> to make this compile on my not /that/ ancient (?) rhel6 box ;)
> 
> Hmm... well, since I don't see backporting xfs kernel scrub to 2.6.32
> maybe xfsprogs' build system should just turn off xfs_scrub on old
> systems?
> 
> In any case, I #ifdef BLKROTATIONAL'd out the entire clause.

ok.  well, other distros are making noise about using bleeding edge progs
w/ older distro kernels (hence the mkfs config file wishes) so it's probably
good to consider building against older environments.

Thanks,
-Eric
--
To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Darrick J. Wong Jan. 12, 2018, 1:27 a.m. UTC | #4 
On Thu, Jan 11, 2018 at 06:04:38PM -0600, Eric Sandeen wrote:
> On 1/11/18 5:59 PM, Darrick J. Wong wrote:
> > On Thu, Jan 11, 2018 at 05:24:58PM -0600, Eric Sandeen wrote:
> ...
> 
> >>> +	/* Non-rotational device?  Throw all the CPUs. */
> >>> +	rot = 1;
> >>> +	error = ioctl(disk->d_fd, BLKROTATIONAL, &rot);
> >>> +	if (error == 0 && rot == 0)
> >>> +		return nproc;
> >>
> >> I needed
> >>
> >> +#ifndef BLKROTATIONAL
> >> +#define BLKROTATIONAL _IO(0x12,126)
> >> +#endif
> >>
> >> to make this compile on my not /that/ ancient (?) rhel6 box ;)
> > 
> > Hmm... well, since I don't see backporting xfs kernel scrub to 2.6.32
> > maybe xfsprogs' build system should just turn off xfs_scrub on old
> > systems?
> > 
> > In any case, I #ifdef BLKROTATIONAL'd out the entire clause.
> 
> ok.  well, other distros are making noise about using bleeding edge progs
> w/ older distro kernels (hence the mkfs config file wishes) so it's probably
> good to consider building against older environments.

<shrug> ok I can patch it in like that...

--D

> Thanks,
> -Eric
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/scrub/Makefile b/scrub/Makefile
index 097ec84..c3a9986 100644
--- a/scrub/Makefile
+++ b/scrub/Makefile
@@ -17,10 +17,12 @@  endif	# scrub_prereqs
 
 HFILES = \
 common.h \
+disk.h \
 xfs_scrub.h
 
 CFILES = \
 common.c \
+disk.c \
 xfs_scrub.c
 
 LLDLIBS += $(LIBHANDLE) $(LIBFROG) $(LIBPTHREAD)
diff --git a/scrub/disk.c b/scrub/disk.c
new file mode 100644
index 0000000..d4bf81f
--- /dev/null
+++ b/scrub/disk.c
@@ -0,0 +1,164 @@ 
+/*
+ * Copyright (C) 2018 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include <stdio.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/ioctl.h>
+#include <sys/statvfs.h>
+#include <sys/vfs.h>
+#include <linux/fs.h>
+#include "platform_defs.h"
+#include "libfrog.h"
+#include "xfs_scrub.h"
+#include "disk.h"
+
+/*
+ * Disk Abstraction
+ *
+ * These routines help us to discover the geometry of a block device,
+ * estimate the amount of concurrent IOs that we can send to it, and
+ * abstract the process of performing read verification of disk blocks.
+ */
+
+/* Figure out how many disk heads are available. */
+static unsigned int
+__disk_heads(
+	struct disk		*disk)
+{
+	int			iomin;
+	int			ioopt;
+	unsigned short		rot;
+	int			error;
+
+	/* If it's not a block device, throw all the CPUs at it. */
+	if (!S_ISBLK(disk->d_sb.st_mode))
+		return nproc;
+
+	/* Non-rotational device?  Throw all the CPUs. */
+	rot = 1;
+	error = ioctl(disk->d_fd, BLKROTATIONAL, &rot);
+	if (error == 0 && rot == 0)
+		return nproc;
+
+	/*
+	 * Sometimes we can infer the number of devices from the
+	 * min/optimal IO sizes.
+	 */
+	iomin = ioopt = 0;
+	if (ioctl(disk->d_fd, BLKIOMIN, &iomin) == 0 &&
+	    ioctl(disk->d_fd, BLKIOOPT, &ioopt) == 0 &&
+	    iomin > 0 && ioopt > 0) {
+		return min(nproc, max(1, ioopt / iomin));
+	}
+
+	/* Rotating device?  I guess? */
+	return 2;
+}
+
+/* Figure out how many disk heads are available. */
+unsigned int
+disk_heads(
+	struct disk		*disk)
+{
+	if (nr_threads)
+		return nr_threads;
+	return __disk_heads(disk);
+}
+
+/* Open a disk device and discover its geometry. */
+struct disk *
+disk_open(
+	const char		*pathname)
+{
+	struct disk		*disk;
+	int			lba_sz;
+	int			error;
+
+	disk = calloc(1, sizeof(struct disk));
+	if (!disk)
+		return NULL;
+
+	disk->d_fd = open(pathname, O_RDONLY | O_DIRECT | O_NOATIME);
+	if (disk->d_fd < 0)
+		goto out_free;
+
+	/* Try to get LBA size. */
+	error = ioctl(disk->d_fd, BLKSSZGET, &lba_sz);
+	if (error)
+		lba_sz = 512;
+	disk->d_lbalog = log2_roundup(lba_sz);
+
+	/* Obtain disk's stat info. */
+	error = fstat(disk->d_fd, &disk->d_sb);
+	if (error)
+		goto out_close;
+
+	/* Determine bdev size, block size, and offset. */
+	if (S_ISBLK(disk->d_sb.st_mode)) {
+		error = ioctl(disk->d_fd, BLKGETSIZE64, &disk->d_size);
+		if (error)
+			disk->d_size = 0;
+		error = ioctl(disk->d_fd, BLKBSZGET, &disk->d_blksize);
+		if (error)
+			disk->d_blksize = 0;
+		disk->d_start = 0;
+	} else {
+		disk->d_size = disk->d_sb.st_size;
+		disk->d_blksize = disk->d_sb.st_blksize;
+		disk->d_start = 0;
+	}
+
+	return disk;
+out_close:
+	close(disk->d_fd);
+out_free:
+	free(disk);
+	return NULL;
+}
+
+/* Close a disk device. */
+int
+disk_close(
+	struct disk		*disk)
+{
+	int			error = 0;
+
+	if (disk->d_fd >= 0)
+		error = close(disk->d_fd);
+	disk->d_fd = -1;
+	free(disk);
+	return error;
+}
+
+/* Read-verify an extent of a disk device. */
+ssize_t
+disk_read_verify(
+	struct disk		*disk,
+	void			*buf,
+	uint64_t		start,
+	uint64_t		length)
+{
+	return pread(disk->d_fd, buf, length, start);
+}
diff --git a/scrub/disk.h b/scrub/disk.h
new file mode 100644
index 0000000..834678e
--- /dev/null
+++ b/scrub/disk.h
@@ -0,0 +1,39 @@ 
+/*
+ * Copyright (C) 2018 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#ifndef XFS_SCRUB_DISK_H_
+#define XFS_SCRUB_DISK_H_
+
+struct disk {
+	struct stat	d_sb;
+	int		d_fd;
+	int		d_lbalog;
+	unsigned int	d_flags;
+	unsigned int	d_blksize;	/* bytes */
+	uint64_t	d_size;		/* bytes */
+	uint64_t	d_start;	/* bytes */
+};
+
+unsigned int disk_heads(struct disk *disk);
+struct disk *disk_open(const char *pathname);
+int disk_close(struct disk *disk);
+ssize_t disk_read_verify(struct disk *disk, void *buf, uint64_t startblock,
+		uint64_t blockcount);
+
+#endif /* XFS_SCRUB_DISK_H_ */