回合上游社区补丁,补丁数量：5

需要同步的补丁如下：<table ><tr><td colSpan="1" rowSpan="1">CommitId</td><td colSpan="1" rowSpan="1">描述</td></tr><tr><td colSpan="1" rowSpan="1">230e7757a7805c7c530d0914936f353882bd504e</td><td colSpan="1" rowSpan="1">LOGROTATE: logrotate should also signal sssd_kcm sssd_kcm is not registered with SSSD&#x27;s monitor, so it is not signaled when it must restart the log. Adding this command will directly signal sssd_kcm (in addition to the monitor). If sssd_kcm is also running in one or more containers, they will also receive the signal. Because only the log files in the host where rotated, the instances in the containers will go on using the same log files. Nothing will happen except for the Received SIGHUP. Rotating logfiles. message in the log files. If we want to avoid this, we should implement a PID file. Reviewed-by: Sumit Bose Reviewed-by: Tomáš Halman </td></tr><tr><td colSpan="1" rowSpan="1">3cba6d1153c102f9596335db28cc017e8338e868</td><td colSpan="1" rowSpan="1">KCM: Fixed a wrong check The pointer to the newly allocated iobuffer is stored into state-&gt;op_ctx-&gt;reply but the check for NULL is done on state-&gt;reply, which we already know is not NULL because it was checked before and not modified after that. Reviewed-by: Alexey Tikhonov Reviewed-by: Tomáš Halman </td></tr><tr><td colSpan="1" rowSpan="1">ae6b9163be0a5a8846e8dbf2e0da2c29221781b9</td><td colSpan="1" rowSpan="1">pam_sss: fix passthrow of old authtok from another pam modules at PAM_PRELIM_CHECK pam_sss ignored old authtoks passed from another pam modules Resolves: https://github.com/SSSD/sssd/issues/7007 Resolves: https://github.com/SSSD/sssd/issues/5418 Reviewed-by: Iker Pedrosa Reviewed-by: Sumit Bose </td></tr><tr><td colSpan="1" rowSpan="1">cffe6e09c6b4cd8afa049365bbd432ace5d2a9d9</td><td colSpan="1" rowSpan="1">nssidmap: fix sss_nss_getgrouplist_timeout() with empty secondary group list sss_nss_getgrouplist_timeout() is intended as a replacement for getgrouplist() which only gets secondary groups from SSSD. Currently it returns an ENOENT error if there are no secondary groups returned by SSSD. However, as with getgrouplist(), there is the second parameter which expects a single GID which will be added to the result. This means that sss_nss_getgrouplist_timeout() will always return at least this GID as a result and an ENOENT error does not make sense. With this patch sss_nss_getgrouplist_timeout() will not return an error anymore if there are no secondary groups but just a result with the single GID from the second parameter. Reviewed-by: Alejandro López Reviewed-by: Tomáš Halman </td></tr><tr><td colSpan="1" rowSpan="1">cbae6855320b53f3f2bdc0e11c5a9c8eb84daf87</td><td colSpan="1" rowSpan="1">KCM: Fix a memory leak When an operation is processed, a buffer is allocated for the reply and its parent is the client context (struct cli_ctx). This buffer is not explicitly freed but it is released when the client context is freed. With each operation a new buffer is allocated and the previous one gets lost. This is not an actual leak because the lost buffers are released by talloc once the client context is freed, when the connection is closed. But on long-lived connections this can consume a large amount of memory before the connection is closed. To solve this, the request context (struct kcm_req_ctx) is the new parent of the buffer. The request is freed as soon as the operation is completed and no buffer gets lost. Resolves: https://github.com/SSSD/sssd/issues/7072 Reviewed-by: Iker Pedrosa Reviewed-by: Sumit Bose </td></tr></table>

src-openEuler/sssd
Closed

Content Risk Flag

Comments (2)

src-openEuler/sssdClosed .gitee-modal { width: 500px !important; }

Content Risk Flag