similar to: [nbdkit PATCH] file: Diagnose a missing file earlier

Add a new mode to the file plugin, using directory=DIR instead of [file=]FILE, to allow it to serve all regular/block files in a given directory, as well as advertising the names of those files it will be serving. Signed-off-by: Eric Blake <eblake@redhat.com> --- plugins/file/nbdkit-file-plugin.pod | 28 +++++- tests/Makefile.am | 4 +- plugins/file/file.c

Introduce a new helper function to resolve a path name, calling nbdkit_error on failure: other than doing what nbdkit_absolute_path does, it also checks that the file exists (and thus avoids errors later on). To help distinguish it from nbdkit_absolute_path, improve the documentation of the latter. Apply it where an existing path is required, both in nbdkit itself and in plugins. Related to:

Hi, this is the v3 of my series for nbdkit_realpath; this series adds an extra documentation update to nbdkit_absolute_path, documenting when it can only be used, and then adds nbdkit_realpath. Thanks, Pino Toscano (2): docs: improve documentation of nbdkit_absolute_path plugin: add and use nbdkit_realpath docs/nbdkit-plugin.pod | 24 +++++++++++++++++++++++- include/nbdkit-common.h

Introduce a new helper function to resolve a path name, calling nbdkit_error on failure: other than doing what nbdkit_absolute_path does, it also checks that the file exist (and thus avoid errors later on). Apply it where an existing path is required, both in nbdkit itself and in plugins. Related to: https://bugzilla.redhat.com/show_bug.cgi?id=1527334 --- docs/nbdkit-plugin.pod | 13

Instead of needing to provide .unload, we can let nbdkit manage the lifetime for us. However, if we like this approach, it may be wiser to introduce new variants of nbdkit_realpath/absolute_path which return const char * already intern'd. Signed-off-by: Eric Blake <eblake@redhat.com> --- See the cover letter: if we like this, there are several other plugins that could also reduce

On 08/02/2018 02:05 PM, Nir Soffer wrote: > When using file systems not supporting ZERO_RANGE (e.g. NFS 4.2) or > block device on kernel < 4.9, we used to call fallocate() for every > zero, fail with EOPNOTSUPP, and fallback to manual zeroing. When > trimming, we used to try unsupported fallocate() on every call. > > Change file handle to remember if punching holes or

This is the second version to support efficient zero for block devices on older kernels (e.g. RHEL 7.5), and file systems that do not support yet FALLOC_FS_ZERO_RANGE (e.g. NFS 4.2). Changes since v1: - Split to smaller patches - Skip linux only includes on other systems - Skip code using BLKZEROOUT if the macro is not defined - Try BLKZEROOUT only if the offset and count are aligned to device

Previously the file plugin supported ‘rdelay’ and ‘wdelay’ parameters for injecting delays (for testing) into read and write requests. This moves the functionality to a new delay filter so that it can be used with any plugin. --- TODO | 3 - configure.ac | 1 + filters/Makefile.am | 1 + filters/delay/Makefile.am

This version addresses comments on v3. Changes since v3: - Finally got spacing right (Eric) - Reorder includes (Richard) - Return 0 or -1 instead of r (Richard) - Add common/include/isaligned.h to Makefile.am (Richard) v3 was here: https://www.redhat.com/archives/libguestfs/2018-August/msg00177.html Nir Soffer (4): file: Avoid unsupported fallocate() calls file: Support zero without

Here's changes to the file plugin (which I'm happy with) and a new exportname filter (which is still at RFC stage; I need to finish implementing strict mode in .open, and add tests). I also discovered that we really want .list_exports and .open to know when they are used on plaintext vs. tls clients for --tls=on, and we may want to split out a new .default_export callback rather than

fallocate(FALLOC_FL_ZERO_RANGE) is supported for block devices with modern kernel, but when it is not, we fall back to manual zeroing. For block device, try also to use ioctl(BLKZEROOUT) if offset and count are aligned to block device sector size. Here is an example run without this change on RHEL 7.5: $ export SOCK=/tmp/nbd.sock $ export

fallocate(FALLOC_FL_ZERO_RANGE) is supportd for block devices with modern kernel, but when it is not, we fall back to manual zeroing. For block device, try also to use ioctl(BLKZEROOUT) if offset and count are aligned to block device sector size. Here is an example run without this change on RHEL 7.5: $ export SOCK=/tmp/nbd.sock $ export

fallocate(FALLOC_FL_ZERO_RANGE) is supportd for block devices with modern kernel, but when it is not, we fall back to manual zeroing. Check if the underlying file is a block device when opening the file, and fall back to ioctl(BLKZEROOUT) for aligned zero requests for a block device. Here is an example run without this change on RHEL 7.5: $ export SOCK=/tmp/nbd.sock $ export

When using file systems not supporting ZERO_RANGE (e.g. NFS 4.2) or block device on kernel < 4.9, we used to call fallocate() for every zero, fail with EOPNOTSUPP, and fallback to manual zeroing. When trimming, we used to try unsupported fallocate() on every call. Change file handle to remember if punching holes or zeroing range are supported, and avoid unsupported calls. - zero changed to:

When using file systems not supporting ZERO_RANGE (e.g. NFS 4.2) or block device on kernel < 4.9, we used to call fallocate() for every zero, fail with EOPNOTSUPP, and fallback to manual zeroing. When trimming, we used to try unsupported fallocate() on every call. Change file handle to remember if punching holes or zeroing range are supported, and avoid unsupported calls. - zero changed to:

When using file systems not supporting ZERO_RANGE (e.g. NFS 4.2) or block device on kernel < 4.9, we used to call fallocate() for every zero, fail with EOPNOTSUPP, and fallback to manual zeroing. When trimming, we used to try unsupported fallocate() on every call. Change file handle to remember if punching holes or zeroing range are supported, and avoid unsupported calls. - zero changed to:

This version addresses some of the comments on v2. Changes since v2: - file_zero: Add missing space in function call - is_aligned: Assert that align is indeed a power of 2 - Spelling in commit message Not changed: - Eric commented that spacing was off: https://www.redhat.com/archives/libguestfs/2018-August/msg00113.html but I could not find anything wrong. - Eric asked if ioctl.h will cause

If we may not trim, we tried ZERO_RANGE, but this is not well supported yet, for example it is not available on NFS 4.2. ZERO_RANGE and PUNCH_HOLE are supported now on block devices, but not on RHRL 7, so we fallback to slow manual zeroing there. Change the logic to support block devices on RHEL 7, and file systems that do not support ZERO_RANGE. The new logic: - If we may trim, try PUNCH_HOLE -

When using file systems not supporting ZERO_RANGE (e.g. NFS 4.2) or block device on kernel < 4.9, we used to call fallocate() for every zero, fail with EOPNOTSUPP, and fallback to manual zeroing. When trimming, we used to try unsupported fallocate() on every call. Change file handle to remember if punching holes or zeroing range are supported, and avoid unsupported calls. - can_trim changed

On Thu, Aug 2, 2018 at 10:30 PM Eric Blake <eblake@redhat.com> wrote: > On 08/02/2018 02:05 PM, Nir Soffer wrote: > > When using file systems not supporting ZERO_RANGE (e.g. NFS 4.2) or > > block device on kernel < 4.9, we used to call fallocate() for every > > zero, fail with EOPNOTSUPP, and fallback to manual zeroing. When > > trimming, we used to try