| Message ID | 20250124173926.2369699-1-richard.purdie@linuxfoundation.org |
|---|---|
| State | Accepted, archived |
| Commit | 9d91a5674c515a43ae76d8615f72e5e2dc16c961 |
| Headers | show |
| Series | [1/2] oeqa/sshcontrol: Handle empty reads | expand |
diff --git a/meta/lib/oeqa/utils/sshcontrol.py b/meta/lib/oeqa/utils/sshcontrol.py index 36c2ecb3dba..6c5648779a5 100644 --- a/meta/lib/oeqa/utils/sshcontrol.py +++ b/meta/lib/oeqa/utils/sshcontrol.py @@ -57,8 +57,10 @@ class SSHProcess(object): if select.select([self.process.stdout], [], [], 5)[0] != []: data = os.read(self.process.stdout.fileno(), 1024) if not data: - self.process.stdout.close() - eof = True + self.process.poll() + if self.process.returncode is None: + self.process.stdout.close() + eof = True else: data = data.decode("utf-8") output += data
Looking at some of the autobuilder failures, it seems that somehow empty reads might be possible despite not being EOF. Tweak the code to be a little more robust in handling this. In theory this shouldn't be possible but python does handle signals a bit differently (e.g. transparrently retrying syscalls for EINTR) so adding this check and a bit of code safety at least rules out this problem. Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> --- meta/lib/oeqa/utils/sshcontrol.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)