From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ozlabs.org (ozlabs.org [IPv6:2401:3900:2:1::2]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by lists.ozlabs.org (Postfix) with ESMTPS id 3rsPsf3R63zDqGv for ; Sun, 17 Jul 2016 08:55:18 +1000 (AEST) Message-ID: <1468709715.20552.161.camel@kernel.crashing.org> Subject: Re: [PATCH 02/14] powerpc/irq: Add support for HV virtualization interrupts From: Benjamin Herrenschmidt To: Michael Ellerman , linuxppc-dev@ozlabs.org Date: Sat, 16 Jul 2016 17:55:15 -0500 In-Reply-To: <1468704837.20552.146.camel@kernel.crashing.org> References: <1467959838-6900-1-git-send-email-benh@kernel.crashing.org> <1467959838-6900-2-git-send-email-benh@kernel.crashing.org> <87ziphamta.fsf@@concordia.ellerman.id.au> <1468704837.20552.146.camel@kernel.crashing.org> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Sat, 2016-07-16 at 16:33 -0500, Benjamin Herrenschmidt wrote: > On Sat, 2016-07-16 at 20:40 +1000, Michael Ellerman wrote:s > > This is breaking allmodconfig with: > > > > arch/powerpc/kernel/exceptions-64s.S:778: Error: attempt to move > > .org backwards > > arch/powerpc/kernel/exceptions-64s.S:779: Error: attempt to move > > .org backwards > > arch/powerpc/kernel/exceptions-64s.S:803: Error: attempt to move > > .org backwards > > arch/powerpc/kernel/exceptions-64s.S:804: Error: attempt to move > > .org backwards > > arch/powerpc/kernel/exceptions-64s.S:823: Error: attempt to move > > .org backwards > > arch/powerpc/kernel/exceptions-64s.S:833: Error: attempt to move > > .org backwards > > > > http://kisskb.ozlabs.ibm.com/kisskb/buildresult/12746513/ > > Yuck. Is that even fixable without completely shuffling all the > IRQ stuff ? We really need to find a way to sort that shit once > and for all... I reproduced... We need to move more of the "common" ones out of the way, without making the conditional branches lose it. Ugh....  Another quick and easy option: Get rid of Cell :-) There is 3 cell ones right there  we could get rid of ;-)  ... except that's not enough. Now the OOL handlers are bumping into the FWNMI page. Argh. Moved a couple more, seems to work, I'm sending you a patch in the next few monites to test that needs to be added *before* the HV virtualization interrupts one. It works with the few variants of allmodconfigs I managed to try on the laptop (basically with and without RELOCATABLE). Let me know how it goes in kisskb. Cheers, Ben.