On Tue, Aug 18, 2026 at 05:16:33PM +0200, Peter Krempa via Devel wrote:

From: Peter Krempa <pkrempa@redhat.com>

Preserve the error on cleanup path as cleanup of the 'libivrt_iohelper' process may overwrite it with an unactionable error, such as when attempting to restore a save image while the VM is already running:

$ virsh list Id Name State -------------------------- 9 cd running

$ virsh restore /tmp/cd.save error: Failed to restore domain from /tmp/cd.save error: internal error: Child process (LIBVIRT_LOG_OUTPUTS=1:stderr /home/pipo/build/libvirt/gcc/src/libvirt_iohelper /tmp/cd.save 0) unexpected fatal signal 13

After this patch the original error is reported:

$ virsh restore /tmp/cd.save error: Failed to restore domain from /tmp/cd.save error: Requested operation is not valid: domain 'cd' is already active

Signed-off-by: Peter Krempa <pkrempa@redhat.com>

Reviewed-by: Martin Kletzander <mkletzan@redhat.com>

--- src/qemu/qemu_driver.c | 3 +++ 1 file changed, 3 insertions(+)

diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 61d95bfbbf..25ba6695bb 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -5985,6 +5985,7 @@ qemuDomainRestoreInternal(virQEMUDriver *driver, bool sparse = false; bool bypass_cache = (flags & VIR_DOMAIN_SAVE_BYPASS_CACHE) != 0; g_autoptr(qemuMigrationParams) restoreParams = NULL; + virErrorPtr save_err = NULL;

virCheckFlags(QEMU_DOMAIN_RESTORE_FLAGS, -1);

@@ -6109,6 +6110,7 @@ qemuDomainRestoreInternal(virQEMUDriver *driver, qemuProcessEndJob(vmNew);

cleanup: + virErrorPreserveLast(&save_err); VIR_FORCE_CLOSE(fd); if (virFileWrapperFdClose(wrapperFd) < 0) ret = -1; @@ -6117,6 +6119,7 @@ qemuDomainRestoreInternal(virQEMUDriver *driver, if (vmNew && ret < 0) qemuDomainRemoveInactive(vmNew, 0, false); virDomainObjEndAPI(&vmNew); + virErrorRestore(&save_err); return ret; }

-- 2.55.0