From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gnu.wildebeest.org (gnu.wildebeest.org [45.83.234.184]) by sourceware.org (Postfix) with ESMTPS id 4782B3858D28 for ; Fri, 8 Apr 2022 20:59:58 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org 4782B3858D28 Authentication-Results: sourceware.org; dmarc=none (p=none dis=none) header.from=klomp.org Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=klomp.org Received: from reform (deer0x0f.wildebeest.org [172.31.17.145]) (using TLSv1.2 with cipher ADH-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by gnu.wildebeest.org (Postfix) with ESMTPSA id 27F6430008BD; Fri, 8 Apr 2022 22:59:56 +0200 (CEST) Received: by reform (Postfix, from userid 1000) id A18662E80AC2; Fri, 8 Apr 2022 22:59:55 +0200 (CEST) Date: Fri, 8 Apr 2022 22:59:55 +0200 From: Mark Wielaard To: Milian Wolff Cc: "Frank Ch. Eigler" , elfutils-devel@sourceware.org Subject: Re: caching failed lookups of debuginfo? Message-ID: References: <4448277.fIUe8AKecr@milian-workstation> <20220408200527.GC23295@redhat.com> <2854938.4GpPBkblLQ@milian-workstation> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <2854938.4GpPBkblLQ@milian-workstation> X-Spam-Status: No, score=-3.8 required=5.0 tests=BAYES_00, KAM_DMARC_STATUS, SPF_HELO_NONE, SPF_PASS, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.4 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on server2.sourceware.org X-BeenThere: elfutils-devel@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Elfutils-devel mailing list List-Unsubscribe: , List-Archive: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 08 Apr 2022 20:59:59 -0000 Hi Milian, On Fri, Apr 08, 2022 at 10:45:10PM +0200, Milian Wolff wrote: > In the off-CPU flame graphs I clearly see this callchain when DEBUGINFOD_URLS > is set: > > ``` > dwfl_standard_find_debuginfo > debuginfod_query_server > curl_multi_wait > ``` > > This accounts for ~2.5s of off-CPU time in my case. If you are saying that > negative lookups are cached, then what is this? Why would a second run of the > same dwfl-using application with the same input data keep querying the server? That is certainly not what should happen once the cache is filled. Could you run with DEBUGINFOD_VERBOSE=1 to see what is being fetched and why? Thanks, Mark