From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from firstgate.proxmox.com (firstgate.proxmox.com [212.224.123.68]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits)) (No client certificate requested) by lists.proxmox.com (Postfix) with ESMTPS id CAAF38F41 for ; Tue, 7 Mar 2023 19:50:16 +0100 (CET) Received: from firstgate.proxmox.com (localhost [127.0.0.1]) by firstgate.proxmox.com (Proxmox) with ESMTP id A461C103C0 for ; Tue, 7 Mar 2023 19:49:46 +0100 (CET) Received: from proxmox-new.maurer-it.com (proxmox-new.maurer-it.com [94.136.29.106]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by firstgate.proxmox.com (Proxmox) with ESMTPS for ; Tue, 7 Mar 2023 19:49:44 +0100 (CET) Received: from proxmox-new.maurer-it.com (localhost.localdomain [127.0.0.1]) by proxmox-new.maurer-it.com (Proxmox) with ESMTP id 682EA45865 for ; Tue, 7 Mar 2023 19:49:44 +0100 (CET) Message-ID: <9e245a3d-98ea-8ea0-e953-d1dbc0496393@proxmox.com> Date: Tue, 7 Mar 2023 19:49:43 +0100 MIME-Version: 1.0 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:111.0) Gecko/20100101 Thunderbird/111.0 Content-Language: en-GB, de-AT To: Proxmox VE development discussion , Friedrich Weber References: <20230306140314.1150179-1-f.weber@proxmox.com> From: Thomas Lamprecht In-Reply-To: <20230306140314.1150179-1-f.weber@proxmox.com> Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit X-SPAM-LEVEL: Spam detection results: 0 AWL -0.050 Adjusted score from AWL reputation of From: address BAYES_00 -1.9 Bayes spam probability is 0 to 1% KAM_DMARC_STATUS 0.01 Test Rule for DKIM or SPF Failure with Strict Alignment NICE_REPLY_A -0.001 Looks like a legit reply (A) SPF_HELO_NONE 0.001 SPF: HELO does not publish an SPF Record SPF_PASS -0.001 SPF: sender matches SPF record URIBL_BLOCKED 0.001 ADMINISTRATOR NOTICE: The query to URIBL was blocked. See http://wiki.apache.org/spamassassin/DnsBlocklists#dnsbl-block for more information. [data.store, proxmox.com] Subject: Re: [pve-devel] [PATCH widget-toolkit] fix #4421: ui: guard setProxy against races of slow vs fast requests X-BeenThere: pve-devel@lists.proxmox.com X-Mailman-Version: 2.1.29 Precedence: list List-Id: Proxmox VE development discussion List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 07 Mar 2023 18:50:16 -0000 Am 06/03/2023 um 15:03 schrieb Friedrich Weber: > Some UI components use `Ext.data.Store.setProxy` to change their > associated API endpoint URL in reaction to user input. One example is > `BackupView`, which calls `setProxy` when the user switches from > listing backups on storage A to listing backups on storage B. However, > if A is slow, the UI may receive the response for A *after* the > response for B. It will then display the contents of A as if they were > the contents of B, resulting in a UI inconsistency. > > The reason is that `Ext.data.Store` still processes the slow response > for A, even though it is obsolete. This patch overrides the > responsible callback of `Ext.data.Store` to only process responses > belonging to the currently active proxy object. This should rule out > similar race conditions in all components that use the `setProxy` API. > In the above example, the patch results in the response for A being > ignored. > > Ignored responses are logged to the browser console. > > Note that this patch only concerns components that use `setProxy` for > changing API endpoints. Other components (e.g. those using > `proxy.setURL` for the same purpose) may be open to similar race > conditions. > > Signed-off-by: Friedrich Weber > --- > The original report only concerns the backup view [1], where the race > condition is easy to trigger. While ruling out this particular race is > simple, I thought it would be worthwhile to rule out race condition of > this category for all components. Hence this patch. However, most of the > other races are much harder to trigger, so it may be questionable > whether a general fix is needed. So if wanted, I can alternatively > submit a patch that only fixes the backup view. IMO a general fix/future proofing can be OK, so besides a small nit inline: LGTM, but did not checked/tested this too closely - @Dominik, what do you think on this? > > Also, there are several occurrences of the `proxy.setURL` or `proxy.url = ...` > patterns (see [1]) which are also susceptible to race conditions, and which > are not fixed by this patch. However, for those, I have not found a nice > solution that does not involve changing a lot of call sites. If wanted, I can > give it another try, or alternatively only submit patches for components for > which triggering the race conditions seems realistic. > > [1] https://bugzilla.proxmox.com/show_bug.cgi?id=4421 > > src/Utils.js | 15 +++++++++++++++ > 1 file changed, 15 insertions(+) > > diff --git a/src/Utils.js b/src/Utils.js > index f55b9a5..8a97487 100644 > --- a/src/Utils.js > +++ b/src/Utils.js > @@ -1451,3 +1451,18 @@ Ext.define('Proxmox.Async', { > return new Promise((resolve, _reject) => setTimeout(resolve, millis)); > }, > }); > + > +Ext.override(Ext.data.Store, { > + // If the store's proxy is changed while it is waiting for an AJAX > + // response, `onProxyLoad` will still be called for the outdated response. > + // To avoid displaying inconsistent information, only process responses > + // belonging to the current proxy. > + onProxyLoad: function(operation) { > + let me = this; > + if (operation.getProxy() === me.getProxy()) { > + me.callParent(arguments); > + } else { > + console.log(`ignored outdated response: ${operation.getRequest().getUrl()}`); maybe warn? also, no need to interpolate text on the console helpers, just add as it's own param, which often gives a much better dev experience (as object can then be inspected in full) - here it's just an URL (maybe dump the full thing?), so YNNV. console.warn('ignored outdated response:', operation.getRequest().getUrl()); > + } > + }, > +});