From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-ej1-f53.google.com (mail-ej1-f53.google.com [209.85.218.53]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id AAF443FCD for ; Wed, 1 Sep 2021 13:56:24 +0000 (UTC) Received: by mail-ej1-f53.google.com with SMTP id i21so6910751ejd.2 for ; Wed, 01 Sep 2021 06:56:24 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=BRrNhgrDIwSehOjM+0QWWad7e+b+lvVLJyyUCl9/Exk=; b=FewNOuSfaXBPf5PjzcdFZgDeR+uXL3OAdSKgY1radiJb0OnEwmABD866so4RSH96Iz PBSqXj/lVJs0NSGYjeucv81XG/L0GBO3ytFxLGqDvL2rm8xczWiyIDoHWQlB2UFWPAD7 AfkW3KS1y+E6EvfBl3nVFaOlsKEp+n9NB0/JIE6wRcOgcWZjnKzr1rkS/YDoN5OZHQxZ bv2BkbLlZj+UiN+knYKxTXfOfYDqCyyheI7ka2b9QFQ2if58ZbIFqhPrx+mUSfzmNyFe 9vjOIgFzR+8MOLauC57u82SGh/+nvGje5Voo4X/ryP0p9pJWnWUtqmgSDRCIXMcvyT6a em3A== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=BRrNhgrDIwSehOjM+0QWWad7e+b+lvVLJyyUCl9/Exk=; b=Y1OEDBgS1NgGr/9MyoLPySoMWQrluZUSouhKn9pm6uWXo+LCrkr/+dTRGVl4bRS8Al bunC92runy2j7cZnq/ttnPQx8nm7s7HHfWJ9NZRVcWcHjYb1n3ObqE9HoJsKKkwSfaSL qexKtGp80RIVl1TNzSnOGOd7O7GADZLP6hXBDXYiQ6XCpw09/V0pPY0qeRrx1zu+GMZh HfXMDxn8Sir2ruFrVshyc65GnD0BBem5XvzFQWCyiDB7DPymX9X/mFGK9khVvbijNQZe W6LhYsQFUgxh/sVCwvOMFD+trJnMpkin7PyLxlf6YCzMaQp7C4pf6vNpgFvfiKEuxtc0 qKsQ== X-Gm-Message-State: AOAM5337QqpuHjN5u1btL8EroHaYmVUw9AERef7YnQUFYHexQPOlggu4 iE3AD1Ys7Hex17D2uxlx6wY= X-Google-Smtp-Source: ABdhPJx/c4ausu3jMt9m8i3ag5V/dPMIynN1X73PIVTKfku986poAoon7KEdUGEkoyOhKqDMx2BLKw== X-Received: by 2002:a17:906:b14d:: with SMTP id bt13mr37365394ejb.39.1630504582933; Wed, 01 Sep 2021 06:56:22 -0700 (PDT) Received: from localhost.localdomain (host-79-22-100-164.retail.telecomitalia.it. [79.22.100.164]) by smtp.gmail.com with ESMTPSA id bt24sm20384ejb.77.2021.09.01.06.56.21 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 01 Sep 2021 06:56:22 -0700 (PDT) From: "Fabio M. De Francesco" To: Johan Hovold , Alex Elder Cc: Matthew Wilcox , Alex Elder , linux-staging@lists.linux.dev, linux-kernel@vger.kernel.org, greybus-dev@lists.linaro.org Subject: Re: [greybus-dev] [PATCH v4] staging: greybus: Convert uart.c from IDR to XArray Date: Wed, 01 Sep 2021 15:56:20 +0200 Message-ID: <8914101.vIO1HAjRha@localhost.localdomain> In-Reply-To: <794b3ff8-0240-ff14-8721-cdf510f52be3@linaro.org> References: <20210829092250.25379-1-fmdefrancesco@gmail.com> <6155058.TBsaUTXu4T@localhost.localdomain> <794b3ff8-0240-ff14-8721-cdf510f52be3@linaro.org> Precedence: bulk X-Mailing-List: linux-staging@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: 7Bit Content-Type: text/plain; charset="utf-8" On Wednesday, September 1, 2021 2:09:16 PM CEST Alex Elder wrote: > On 8/31/21 6:50 AM, Fabio M. De Francesco wrote: > > I was wrong in assuming that trivial patches to Greybus are welcome as they > > are for other drivers. > > This is not a correct statement. Yes, I agree: it's not a correct statement. Please let me explain what I was trying to convey with that consideration... The Mutexes were there around idr_find() and I decided to leave the code as it was. Who am I to say that they are not necessary? I must stay on the safe side. First because I don't know how the drivers work (can that critical section really be entered by different threads that could possibly share the gb_tty that is retrieved by xa_load()? Even if xa_load() always give you back the right gb_tty, how do I know if in the while other threads change its fields or destroy the object? I guess I should stay on the safe side and leave the Mutexes there, exactly were they were. These are the reason why v1 was indeed a trivial patch. But v2 *was not* because you wrote that you were pretty sure they were unneeded and you asked me to leave them or remove them and in either case I had to provide a reason why. I guess that in v1 I should not provide a reason why they are still there, as well as I don't have to provide any reason on why the greybus code (line by line) is as it is: it is out of the scope of my patch. Am I wrong? Your note about the possibility that the mutexes could be removed pushed me beyond what I need to know to accomplish the intended task. Anyway I tried to reason about it. I perfectly know what is required to protect critical sections of code, but I don't know how drivers work; I mean I don't know whether or not different threads that run concurrently could really interfere in that specific section. This is because I simply reason in terms of general rules of protection of critical section but I really don't know how Greybus works or (more in general) how drivers work. I still think that if I stayed within the bounds of my original purpose I didn't have to reason about this topic and that the v1 patch was trivial. v2 was not! I'm sorry because I'm still not sure if I was able to conveyed what I thought and still think. > But as Johan pointed out, even for a trivial patch if you > must understand the consequences of what the change does. > If testing is not possible, you must work extra hard to > ensure your patch is correct. Again, I don't see any possible harm with the mutexes in place :) > In the first (or an early) version of your patch I pointed > out a bug. Later, I suggested > the lock might not be necessary > and asked you to either confirm > it was or explain why it was > not, but you didn't do that. This was beyond my knowledge and perhaps unnecessary (sorry if I insist on that :)). > I agree that the change appeared trivial, and even sensible, > but even trivial patches must result in correct code. And > all patches should have good and complete explanations. > > - Alex Is v2 correct with the mutexes restored where they were? I guess it is. Thanks for you kind review and the time you spent for me. I appreciated it, seriously. Fabio