From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from conssluserg-02.nifty.com (conssluserg-02.nifty.com [210.131.2.81]) by sourceware.org (Postfix) with ESMTPS id 342AE396D83F for ; Wed, 23 Jun 2021 03:10:12 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org 342AE396D83F Received: from Express5800-S70 (y084232.dynamic.ppp.asahi-net.or.jp [118.243.84.232]) (authenticated) by conssluserg-02.nifty.com with ESMTP id 15N38mPW004599; Wed, 23 Jun 2021 12:08:48 +0900 DKIM-Filter: OpenDKIM Filter v2.10.3 conssluserg-02.nifty.com 15N38mPW004599 X-Nifty-SrcIP: [118.243.84.232] Date: Wed, 23 Jun 2021 12:08:58 +0900 From: Takashi Yano To: cygwin@cygwin.com Cc: Viet-Duc Le , Jon Turney Subject: Re: xwin-xdg-menu high cpu usage with rxvt-unicode Message-Id: <20210623120858.83e471aa7ddbc20cd41efb2c@nifty.ne.jp> In-Reply-To: References: X-Mailer: Sylpheed 3.7.0 (GTK+ 2.24.30; i686-pc-mingw32) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Spam-Status: No, score=-1.6 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, NICE_REPLY_A, RCVD_IN_BARRACUDACENTRAL, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_PASS, TXREP autolearn=no autolearn_force=no version=3.4.2 X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on server2.sourceware.org X-BeenThere: cygwin@cygwin.com X-Mailman-Version: 2.1.29 Precedence: list List-Id: General Cygwin discussions and problem reports List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 23 Jun 2021 03:10:14 -0000 On Mon, 21 Jun 2021 05:49:28 +0000 Viet-Duc Le wrote: > Hi, > > I've encountered a peculiar issue with rxvt-unicode launched from xwin-xdg-menu (20170321-1) > The image below corresponds to two rxvt-unicode instances, each occupies 30% of system resouce. > https://imgur.com/a/FTFwIrZ > I'm using the the latest version of cygwin. 'cygcheck.out' is attached per guideline. > > Steps to reproduce: > - Start XWin Server from Start menu > - X applications menu -> System Tools -> rxvt-unicode > - Kill rxvt-unicode to free the resource again > > Other observations: > - XTerm does not incur high resources. > - Manually launching rxvt-unicode from XTerm does not incur high resources. > > I think this is not an issue with rxvt but with xwin-xdg-menu. > On laptop, it is very noticeable when the fan kicks in. > Insights and suggestions are much appreciated. > > Thanks. > Viet-Duc I looked into this problem and found the cause. urxvt seems to close stdout while xwin-xdg-menu use select() to read stdout of the launched apps. Since stdout is closed, select() returns repeatedly without any output. This causes high cpu load. I also confirmed that the following patch for xwin-xdg-menu resolves the issue. --- execute.c.orig 2017-03-22 05:31:53.000000000 +0900 +++ execute.c 2021-06-22 10:47:28.917675400 +0900 @@ -100,30 +100,38 @@ default: /* parent */ { + int stdout_ok = TRUE, stderr_ok = TRUE; close(stdout_filedes[1]); close(stderr_filedes[1]); printf("executing '%s', pid %d\n", (char *) cmd, pid); /* read from pipes, write to log, until both are closed */ - while (TRUE) { + while (stdout_ok || stderr_ok) { fd_set readfds, errorfds; - int nfds = max(stdout_filedes[0], stderr_filedes[0]) + 1; + int nfds = 0; FD_ZERO(&readfds); - FD_SET(stdout_filedes[0], &readfds); - FD_SET(stderr_filedes[0], &readfds); + if (stdout_ok) { + FD_SET(stdout_filedes[0], &readfds); + nfds = max(nfds, stdout_filedes[0] + 1); + } + if (stderr_ok) { + FD_SET(stderr_filedes[0], &readfds); + nfds = max(nfds, stderr_filedes[0] + 1); + } errorfds = readfds; if (select(nfds, &readfds, NULL, &errorfds, NULL) > 0) { - if (FD_ISSET(stdout_filedes[0], &readfds)) + if (FD_ISSET(stdout_filedes[0], &errorfds)) + stdout_ok = FALSE; + else if (FD_ISSET(stdout_filedes[0], &readfds)) LogLineFromFd(stdout_filedes[0], "stdout", pid); - if (FD_ISSET(stderr_filedes[0], &readfds)) - LogLineFromFd(stderr_filedes[0], "stderr", pid); - if (FD_ISSET(stdout_filedes[0], &errorfds) && - FD_ISSET(stderr_filedes[0], &errorfds)) - break; + if (FD_ISSET(stderr_filedes[0], &errorfds)) + stderr_ok = FALSE; + else if (FD_ISSET(stderr_filedes[0], &readfds)) + LogLineFromFd(stderr_filedes[0], "stderr", pid); } else { break; Jon, could you please have a look? -- Takashi Yano