similar to: [PATCH v3] lib: libvirt: Convert all drive socket parameters to an absolute path (RHBZ#1588451).

--- lib/launch-libvirt.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/lib/launch-libvirt.c b/lib/launch-libvirt.c index b279aa6b9..48404ef93 100644 --- a/lib/launch-libvirt.c +++ b/lib/launch-libvirt.c @@ -23,6 +23,7 @@ #include <stdarg.h> #include <stdbool.h> #include <unistd.h> +#include <limits.h> #include <fcntl.h>

One interesting omission is that we don't allow sockets in the abstract namespace. The API won't let you pass these sockets because they contain a '\0' character in the middle of the string. Therefore this patch doesn't need to deal with those. However we should in future allow that, probably using the '@' character to stand in for the NUL byte, as is used in a few

Also fix two tests which assumed that you could add a non-existent socket. --- fish/test-add-uri.sh | 16 ++++++++++++---- lib/drives.c | 14 +++++++++++++- tests/disks/test-qemu-drive.sh | 19 ++++++++++++++----- 3 files changed, 39 insertions(+), 10 deletions(-) diff --git a/fish/test-add-uri.sh b/fish/test-add-uri.sh index 756df997b..cb4d40199 100755 ---

QEMU upstream has broken snapshot=on ... again. These two patches stop using it entirely. Instead we run 'qemu-img create' to create overlay disks as required. Note that the libvirt and UML backends were already doing this: The libvirt backend because <transient/> has never worked, and the UML backend was running uml_mkcow because the UML-equivalent syntax of snapshot=on was

Consolidate and extend the use of funky start_element() etc macros. Rich.

v1 was here: https://www.redhat.com/archives/libguestfs/2018-October/msg00047.html However it was broken in a few ways. First of all the documentation was broken because "/**" enhanced comments were not permitted on macros. This is fixed in the new 1/4 patch. Secondly we didn't use single_element() everywhere possible, which is fixed in the new 4/4 patch. Lastly I've

v1 -> v2: - Rebase on top of Pino's version work. Two patches went upstream, these are the two remaining patches. Note the generation number is still inside the qemu.stat file. We could put it in the filename, I have no particular preference. Rich.

Don't include the one from /usr/include. Fixes commit 05a8c47b9be8a6b8ffc091eab461fece9d40dcd0. Thanks: Nir Soffer. --- tests/Makefile.am | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/tests/Makefile.am b/tests/Makefile.am index 80b5c6d..86bef71 100644 --- a/tests/Makefile.am +++ b/tests/Makefile.am @@ -553,6 +553,7 @@ noinst_LTLIBRARIES += \ test_layers_plugin_la_SOURCES = \

Doing qemu feature detection in the direct backend takes ~100ms because we need to run `qemu -help' and `qemu -devices ?', and each of those interacts with glibc's very slow link loader. Fixing the link loader is really hard. Instead memoize the output of those two commands. This patch series first separates all the code dealing with qemu into a separate module (src/qemu.c) and

This patch causes aarch64 to use virtio-pci instead of virtio-mmio. Virtio-pci is considerably faster than virtio-mmio, it's more like how other architectures work, and it supports hotplugging (although it's not likely we'd use the latter feature). I'm not necessarily suggesting that we apply this. Laine (CC'd) has some further patches to libvirt lined up which AIUI would

In some places when generating XML output in C code we use some clever macros: start_element ("memory") { attribute ("unit", "MiB"); string ("%d", g->memsize); } end_element (); This commit which is mostly refactoring moves the repeated definitions of these macros into a common header file. I also took this opportunity to change / clean up

Parse the query string from URLs, and pass it as new add_drive optarg; use it when building the file= URL for qemu. Accept query string only for http and https protocols, for now. --- Possibly it looks like an ad-hoc solution for http(s), although I'm not sure how it could possibly be generalized somehow (maybe "extra params" which would be the query string for http(s)?).

This request stems off a bug report I posted https://bugs.r-project.org/bugzilla/show_bug.cgi?id=17757 where it was determined the current behaviour is as expected. To recap: when given a real file, normalizePath() always* returns the full absolute path. When given a non-existent file, normalizePath() returns a full path on Windows but it returns the input on other systems*. I'd argue that

The dummy library can be compiled this way on all platforms since it doesn't rely on external undefined symbols. Reverts commit ba000b469c24ad69604137de975b255bbe70d056. --- configure.ac | 9 --------- tests/Makefile.am | 6 +----- 2 files changed, 1 insertion(+), 14 deletions(-) diff --git a/configure.ac b/configure.ac index 0a0ac60a..92e0d4e3 100644 --- a/configure.ac +++

On Fri, Aug 21, 2020 at 03:15:34PM +0200, Jean-Philippe Brucker wrote: > Add a topology description to the virtio-iommu driver and enable x86 > platforms. > > Since [v2] we have made some progress on adding ACPI support for > virtio-iommu, which is the preferred boot method on x86. It will be a > new vendor-agnostic table describing para-virtual topologies in a > minimal

Hi, On 8/21/20 3:15 PM, Jean-Philippe Brucker wrote: > Add a topology description to the virtio-iommu driver and enable x86 > platforms. > > Since [v2] we have made some progress on adding ACPI support for > virtio-iommu, which is the preferred boot method on x86. It will be a > new vendor-agnostic table describing para-virtual topologies in a > minimal format. However some

As this Caboose article [] suggests, I am not the only one who had this problem. RAILS_ROOT is a relative path on Windows, and it can bite you in unexpected ways. I had that problem while working on CruiseControl.rb. There is a trivial patch fixing it (by simply applying File.expand_path to the path before assigning it into the RAILS_ROOT constant): http://dev.rubyonrails.org/ticket/7259 Best

This is an incomplete port to Windows. Currently the server compiles and starts up successfully, but goes into an infinite loop when you connect to it. Nevertheless I think the approach is ready for feedback. This being Windows the changes go quite deep. Rich.

On 3/26/20 1:25 PM, Richard W.M. Jones wrote: > Note that this probably isn't testing anything because at least on > Linux the -undefined/-no-undefined flags appear to have no effect on > linking. > [skipping review of 5/9 for now, due to its size. I don't know how easy or hard it would be to split it into bisectable pieces: first introducing the library with just the

On 1/14/19 6:15 AM, Richard W.M. Jones wrote: > In commit 3775916120749e935d287f35cb29ff6c586656df (and before that in > discussions with Eric) we decided that we should try to allow plugins > to be compiled with C compilers other than GCC or Clang (or with very > old / peculiar / incompatible versions of those compilers). > > However until we test this we cannot guarantee that