From 5f032256db59144373cb714473b251de3d63b24f Mon Sep 17 00:00:00 2001
From: Paul Holzinger <pholzing@redhat.com>
Date: Wed, 4 Jan 2023 16:34:50 +0100
Subject: [PATCH] podman logs: journald fix --since and --follow

The `containerCouldBeLogging` bool should not be false by default, when
--since is used we seek in the journal and can miss the start event so
that bool would stay false forever. This means that a running container
is not followed even when it should.

To fix this we can just set the `containerCouldBeLogging` bool based on
the current contianer state.

Fixes #16950

Signed-off-by: Paul Holzinger <pholzing@redhat.com>
---
 libpod/container_log_linux.go | 13 +++++++++-
 test/system/035-logs.bats     | 46 +++++++++++++++++++++++++++++++++++
 2 files changed, 58 insertions(+), 1 deletion(-)

diff --git a/libpod/container_log_linux.go b/libpod/container_log_linux.go
index 7e95f2449f..e8f957322b 100644
--- a/libpod/container_log_linux.go
+++ b/libpod/container_log_linux.go
@@ -97,7 +97,18 @@ func (c *Container) readFromJournal(ctx context.Context, options *logs.LogOption
 	// exponential backoff.
 	var cursor string
 	var cursorError error
-	var containerCouldBeLogging bool
+
+	c.lock.Lock()
+	if err := c.syncContainer(); err != nil {
+		c.lock.Unlock()
+		return err
+	}
+	// The initial "containerCouldBeLogging" state must be correct, we cannot rely on the start event being still in the journal.
+	// This can happen if the journal was rotated after the container was started or when --since is used.
+	// https://github.com/containers/podman/issues/16950
+	containerCouldBeLogging := c.ensureState(define.ContainerStateRunning, define.ContainerStateStopping)
+	c.lock.Unlock()
+
 	for i := 1; i <= 3; i++ {
 		cursor, cursorError = journal.GetCursor()
 		hundreds := 1
diff --git a/test/system/035-logs.bats b/test/system/035-logs.bats
index f9ac33d2c7..7e0efc6e98 100644
--- a/test/system/035-logs.bats
+++ b/test/system/035-logs.bats
@@ -298,4 +298,50 @@ $contentC" "logs -f on exitted container works"
 
     _log_test_follow journald
 }
+
+function _log_test_follow_since() {
+    local driver=$1
+    cname=$(random_string)
+    content=$(random_string)
+    local events_backend=$(_additional_events_backend $driver)
+
+    if [[ -n "${events_backend}" ]]; then
+        skip_if_remote "remote does not support --events-backend"
+    fi
+
+    run_podman ${events_backend} run --log-driver=$driver --name $cname $IMAGE echo "$content"
+    # Using --since 0s can flake because the log might written in the second as the logs call is made.
+    # The -1s makes sure we only read logs that would be created 1s in the future which cannot happen.
+    run_podman ${events_backend} logs --since -1s -f $cname
+    assert "$output" == "" "logs --since -f on exited container works"
+
+    run_podman ${events_backend} rm -t 0 -f $cname
+
+    # Now do the same with a running container to check #16950.
+    run_podman ${events_backend} run --log-driver=$driver --name $cname -d $IMAGE \
+        sh -c "sleep 0.5; while :; do echo $content && sleep 3; done"
+
+    # sleep is required to make sure the podman event backend no longer sees the start event in the log
+    # This value must be greater or equal than the the value given in --since below
+    sleep 0.2
+
+    # Make sure podman logs actually follows by giving a low timeout and check that the command times out
+    PODMAN_TIMEOUT=2 run_podman 124 ${events_backend} logs --since 0.1s -f $cname
+    assert "$output" =~ "^$content
+timeout: sending signal TERM to command.*" "logs --since -f on running container works"
+
+    run_podman ${events_backend} rm -t 0 -f $cname
+}
+
+@test "podman logs - --since --follow k8s-file" {
+    _log_test_follow_since k8s-file
+}
+
+@test "podman logs - --since --follow journald" {
+    # We can't use journald on RHEL as rootless: rhbz#1895105
+    skip_if_journald_unavailable
+
+    _log_test_follow_since journald
+}
+
 # vim: filetype=sh