From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Return-Path: Subject: Re: [PATCH resend v2] Extending kernel option pci=resource_alignment to be able to specify PCI device/vendor IDs To: "Koehrer Mathias (ETAS/ESW5)" References: <00e437fe05f54c6a9df12bf48cf28dc5@FE-MBX1012.de.bosch.com> Cc: Gavin Shan , Bjorn Helgaas , "gregkh@linuxfoundation.org" , "linux-pci@vger.kernel.org" , "bhelgaas@google.com" From: Yongji Xie Date: Tue, 26 Jul 2016 18:11:06 +0800 MIME-Version: 1.0 In-Reply-To: <00e437fe05f54c6a9df12bf48cf28dc5@FE-MBX1012.de.bosch.com> Content-Type: text/plain; charset=utf-8; format=flowed Message-Id: <75813d7d-cadd-828f-737e-4113da3427d9@linux.vnet.ibm.com> List-ID: Hi Mathias, On 2016/7/25 14:05, Koehrer Mathias (ETAS/ESW5) wrote: > Hi Gavin, > >> Mathias, The code change would conflict with Yongji's patch where the >> wildcard introduced to identify PCI devices. >> >> https://patchwork.ozlabs.org/patch/642473/ > That's true, however it is also true the other way round. The first sending of my patch was > 22-June 2016 and thus before Yongji's patch... > See also the comment of Bjorn Helgaas on my patch (v1): > http://marc.info/?l=linux-pci&m=146654644431550&w=2 > > However it looks as if there are a couple of (conflicting) approaches for the issue with the alignment of PCI addresses flying around... > My personal opinion is that the need to adapt the alignment of PCI addresses should be done on PCI device/vendor ids of relevant PCI devices and not on PCI addresses. > The PCI addresses are likely to change e.g. by inserting another PCI board. The device/vendor id of the devices will not change. Yes, your opinion is right. I think I will update my patches based on your change. With your patch, seems like we can specify alignment for all PCI devices by using something like: pci=resource_alignment=4096@pci:0:0. We'd better document it. Thanks, Yongji