On Fri, Jul 29, 2022 at 20:36:57 +0000, Praveen K Paladugu wrote:
explicitly set the chr device type to pty to pass the
checks of ch driver.
https://gitlab.com/libvirt/libvirt/-/issues/344
This is not a persuading enough enough commit message ...
Signed-off-by: Praveen K Paladugu <prapal(a)linux.microsoft.com>
---
src/conf/domain_conf.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index e85cc1f809..abe9d8ae08 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -10078,6 +10078,7 @@ virDomainChrSourceDefNew(virDomainXMLOption *xmlopt)
if (!(def = virObjectNew(virDomainChrSourceDefClass)))
return NULL;
+ def->type = VIR_DOMAIN_CHR_TYPE_PTY;
... to justify a global change like this.
Could you please at least clarify which code path is problematic? I see
e.g. that virDomainChrDefParseXML sets VIR_DOMAIN_CHR_TYPE_PTY as the
source type if the user doesn't set one.
If this is a workaround for the stub console added by default in
virDomainDefAddConsoleCompat and the cloud hypervisor e.g. doesn't need
that or it's wrong for it's usage, the modification shoud IMO be limited
to the cloud hypervisor similarly to what I suggested in the issue.
Note that we require commit messages to be self-explanatory without
refering to e.g. outside discussions.