When using tar-in or tools built around it such as virt-make-fs, if
the target filesystem is vfat then unpacking the tarball would fail
because tar tries to chown(2) files and fails.
You would see errors such as:
tar: <file>: Cannot change ownership to uid 500, gid 500: Operation not permitted
This change detects whether the target filesystem supports chown(2).
If not, it adds the --no-same-owner flag to tar to stop it from trying
to change the ownership of newly created files.
Add xfs_info to show the geometry of the xfs filesystem.
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
RWMJ:
- Updated po/POTFILES.
- Use xfs_ prefix for all struct fields.
- Return uninitialized fields as -1 / empty string.
- Copyedit the description.
case_sensitive_path is undefined when the final path element doesn't
exist. Currently it returns an error, but this means that creating a
new file doesn't work as expected:
$ guestfish --rw -i -d windows touch 'win:c:\blah'
libguestfs: error: case_sensitive_path: blah no file or directory found with this name
We should allow this case (provided there is no trailing slash) so
that new files or directories can be created.
On Linux, sync(2) does not actually issue a write barrier, thus it
doesn't force a flush of the underlying hardware write cache (or
qemu's disk cache in the virtual case).
This can be a problem, because libguestfs relies on running sync in
the appliance, followed by killing qemu (using SIGTERM).
In most cases, this is fine, because killing qemu with SIGTERM should
cause it to flush out the disk cache before it exits. However we have
found various bugs in qemu which cause qemu to crash while doing the
flush, leaving the data unwritten (see RHBZ#836913).
The solution is to issue fsync(2) to the block devices. This has a
write barrier, so it ensures that qemu writes out its cache long
before we get around to killing qemu.
This returns the number of whole block devices added. It is usually
simpler to call this than to list the devices and count them, which
is what we do in some places in the current codebase.
The original fix for this in
commit 511c82df46 was not complete, in
that it did not fix the case of the old (pre '-m' option) parted.
This doesn't matter for Fedora, but it matters for RHEL 5 which has
this ancient parted.
This returns the index of the device, eg. /dev/sdb => 1.
Or you can think of it as the order that the device was
added, or the index of the device in guestfs_list_devices.
Apparently e2fsprogs only knows that "/dev/sda" is a whole device, but
doesn't think that "/dev/vda" is. On switching the default device
over to virtio-scsi, that causes mke2fs without -F option to complain
and ask for an interactive prompt. Adding -F forces it to go ahead
anyway.
This caused several less-used APIs to break with virtio-scsi.
This is closer to the real meaning of "availability of btrfs", since
just having the btrfs tool doesn't help much if it's not supported by
the kernel too.
Add the new API btrfs-fsck to check the btrfs filesystem.
Btrfs is currently under heavy development, and not suitable for
any uses other than benchmarking and review. But it'll be useful
in the near future.
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
I used scsi_debug to create a 4k sector virtual disk:
modprobe scsi_debug dev_size_mb=128 sector_size=4096
I then used 'gdisk' to create lots of partitions, and used 'hexdump'
to examine what was written to disk.
Bind the easy parts of the 'btrfs' program.
The new APIs are:
btrfs-device-add: add devices to a btrfs filesystem
btrfs-device-delete: remove devices from a btrfs filesystem
btrfs-filesystem-sync: sync a btrfs filesystem
btrfs-filesystem-balance: balance a btrfs filesystem
btrfs-subvolume-create: create a btrfs snapshot
btrfs-subvolume-delete: delete a btrfs snapshot
btrfs-subvolume-list: list btrfs snapshots and subvolumes
btrfs-subvolume-set-default: set default btrfs subvolume
btrfs-subvolume-snapshot: create a writable btrfs snapshot
The new APIs are:
get-e2attrs: List ext2 file attributes of a file.
set-e2attrs: Set or clear ext2 file attributes of a file.
get-e2generation: Get ext2 file generation of a file.
set-e2generation: Set ext2 file generation of a file.
These are implemented using the lsattr and chattr programs from
e2fsprogs.
This function allows you to pass an explicit errno back to the
library. reply_with_error is redefined as a macro that calls
reply_with_error_errno with errno == 0.
Add an API for doing what virt-sparsify was doing: freeing up free
space in a filesystem.
The current implementation is simple-minded: we create a file, fill it
with zeroes until we run out of space, then delete the file. However
the description leaves it open to do a better implementation, eg.
using sparsification support that is currently being worked on in ext4
and qemu.
The implementation also sends progress notifications, which is an
advantage over the old 'dd' method.
The presumption is that all file descriptors should be created with
the close-on-exec flag set. The only exception are file descriptors
that we want passed through to exec'd subprocesses (mainly pipes and
stdin/stdout/stderr).
For open calls, we pass O_CLOEXEC as an extra flag, eg:
fd = open ("foo", O_RDONLY|O_CLOEXEC);
This is a Linux-ism, but using a macro we can easily make it portable.
For sockets, similarly:
sock = socket (..., SOCK_STREAM|SOCK_CLOEXEC, ...);
For accepted sockets, we use the Linux accept4 system call which
allows flags to be supplied, but we use the Gnulib 'accept4' module to
make this portable.
For dup, dup2, we use the Linux dup3 system call, and the Gnulib
modules 'dup3' and 'cloexec'.
Previously a lot of daemon code used three variables (a string list,
'int size' and 'int alloc') to track growable strings buffers. This
commit implements a simple struct containing the same variables, but
using size_t instead of int:
struct stringsbuf {
char **argv;
size_t size;
size_t alloc;
};
Use it like this:
DECLARE_STRINGSBUF (ret);
//...
if (add_string (&ret, str) == -1)
return NULL;
//...
if (end_stringsbuf (&ret) == -1)
return NULL;
return ret.argv;
