From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from lithops.sigma-star.at ([195.201.40.130]) by bombadil.infradead.org with esmtps (Exim 4.90_1 #2 (Red Hat Linux)) id 1gfAlN-0007pa-6M for linux-um@lists.infradead.org; Thu, 03 Jan 2019 21:42:50 +0000 From: Richard Weinberger Subject: Re: [PATCH] um: Try to avoid kmalloc in signal handling Date: Thu, 03 Jan 2019 22:42:45 +0100 Message-ID: <8232816.4fJl4d8jsT@blindfold> In-Reply-To: References: <20190103160917.10217-1-anton.ivanov@cambridgegreys.com> MIME-Version: 1.0 List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: "linux-um" Errors-To: linux-um-bounces+geert=linux-m68k.org@lists.infradead.org To: Anton Ivanov Cc: linux-um@lists.infradead.org Am Donnerstag, 3. Januar 2019, 22:29:17 CET schrieb Anton Ivanov: > Second thought - this should not use kmalloc at all if it can. It should > either go on the stack the way it used to do before > b6024b21fec8367ef961a771cc9dde31f1831965 or it should have a sufficient > "register stack" to push registers for all use cases. > > Kmalloc/ATOMIC can fail just because it feels like it which means that > this can throw a panic() under very heavy IO. Not good. :-( We cannot go back to stack based allocation. The stack is too small for AVX register sets. Maybe we can preallocate it. (And check what arch/x86/ does) Thanks, //richard _______________________________________________ linux-um mailing list linux-um@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-um