From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-7.0 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS,URIBL_BLOCKED autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id AA3CAC43441 for ; Fri, 16 Nov 2018 01:23:35 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 6DF6D2145D for ; Fri, 16 Nov 2018 01:23:35 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 6DF6D2145D Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S2389194AbeKPLdw (ORCPT ); Fri, 16 Nov 2018 06:33:52 -0500 Received: from mga17.intel.com ([192.55.52.151]:56665 "EHLO mga17.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S2389117AbeKPLdv (ORCPT ); Fri, 16 Nov 2018 06:33:51 -0500 X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga003.fm.intel.com ([10.253.24.29]) by fmsmga107.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 15 Nov 2018 17:23:33 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.56,238,1539673200"; d="scan'208";a="96702593" Received: from allen-box.sh.intel.com (HELO [10.239.161.122]) ([10.239.161.122]) by FMSMGA003.fm.intel.com with ESMTP; 15 Nov 2018 17:23:30 -0800 Cc: baolu.lu@linux.intel.com, Joerg Roedel , David Woodhouse , ashok.raj@intel.com, sanjay.k.kumar@intel.com, jacob.jun.pan@intel.com, kevin.tian@intel.com, Jean-Philippe Brucker , yi.l.liu@intel.com, yi.y.sun@intel.com, peterx@redhat.com, tiwei.bie@intel.com, Zeng Xin , iommu@lists.linux-foundation.org, kvm@vger.kernel.org, linux-kernel@vger.kernel.org, Jacob Pan Subject: Re: [PATCH v4 6/8] vfio/mdev: Add iommu place holders in mdev_device To: Kirti Wankhede , Alex Williamson References: <20181105073408.21815-1-baolu.lu@linux.intel.com> <20181105073408.21815-7-baolu.lu@linux.intel.com> <20181106165356.44b59ec3@w520.home> <7bfc2f2b-10e6-b53c-51f3-c6494b686aa3@nvidia.com> From: Lu Baolu Message-ID: <25b9a20f-06b5-57ac-81f1-fd933845c77f@linux.intel.com> Date: Fri, 16 Nov 2018 09:20:48 +0800 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.2.1 MIME-Version: 1.0 In-Reply-To: <7bfc2f2b-10e6-b53c-51f3-c6494b686aa3@nvidia.com> Content-Type: text/plain; charset=utf-8; format=flowed Content-Language: en-US Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi, On 11/16/18 5:31 AM, Kirti Wankhede wrote: > > > On 11/7/2018 7:18 AM, Lu Baolu wrote: >> Hi Alex, >> >> On 11/7/18 7:53 AM, Alex Williamson wrote: >>> On Mon,  5 Nov 2018 15:34:06 +0800 >>> Lu Baolu wrote: >>> >>>> A parent device might create different types of mediated >>>> devices. For example, a mediated device could be created >>>> by the parent device with full isolation and protection >>>> provided by the IOMMU. One usage case could be found on >>>> Intel platforms where a mediated device is an assignable >>>> subset of a PCI, the DMA requests on behalf of it are all >>>> tagged with a PASID. Since IOMMU supports PASID-granular >>>> translations (scalable mode in vt-d 3.0), this mediated >>>> device could be individually protected and isolated by an >>>> IOMMU. >>>> >>>> This patch adds two new members in struct mdev_device: >>>> * iommu_device >>>>    - This, if set, indicates that the mediated device could >>>>      be fully isolated and protected by IOMMU via attaching >>>>      an iommu domain to this device. If empty, it indicates >>>>      using vendor defined isolation. >>>> >>>> * iommu_domain >>>>    - This is a place holder for an iommu domain. A domain >>>>      could be store here for later use once it has been >>>>      attached to the iommu_device of this mdev. >>>> >>>> Below helpers are added to set and get above iommu device >>>> and iommu domain pointers. >>>> >>>> * mdev_set/get_iommu_device(dev, iommu_device) >>>>    - Set or get the iommu device which represents this mdev >>>>      in IOMMU's device scope. Drivers don't need to set the >>>>      iommu device if it uses vendor defined isolation. >>>> >>>> * mdev_set/get_iommu_domain(domain) >>>>    - A iommu domain which has been attached to the iommu >>>>      device in order to protect and isolate the mediated >>>>      device will be kept in the mdev data structure and >>>>      could be retrieved later. >>>> >>>> Cc: Ashok Raj >>>> Cc: Jacob Pan >>>> Cc: Kevin Tian >>>> Cc: Liu Yi L >>>> Suggested-by: Kevin Tian >>>> Suggested-by: Alex Williamson >>>> Signed-off-by: Lu Baolu >>>> --- >>>>   drivers/vfio/mdev/mdev_core.c    | 36 ++++++++++++++++++++++++++++++++ >>>>   drivers/vfio/mdev/mdev_private.h |  2 ++ >>>>   include/linux/mdev.h             | 23 ++++++++++++++++++++ >>>>   3 files changed, 61 insertions(+) >>>> >>>> diff --git a/drivers/vfio/mdev/mdev_core.c >>>> b/drivers/vfio/mdev/mdev_core.c >>>> index 0212f0ee8aea..5119809225c5 100644 >>>> --- a/drivers/vfio/mdev/mdev_core.c >>>> +++ b/drivers/vfio/mdev/mdev_core.c >>>> @@ -390,6 +390,42 @@ int mdev_device_remove(struct device *dev, bool >>>> force_remove) >>>>       return 0; >>>>   } >>>>   +int mdev_set_iommu_device(struct device *dev, struct device >>>> *iommu_device) >>>> +{ >>>> +    struct mdev_device *mdev = to_mdev_device(dev); >>>> + >>>> +    mdev->iommu_device = iommu_device; >>>> + >>>> +    return 0; >>>> +} >>>> +EXPORT_SYMBOL(mdev_set_iommu_device); >>>> + >>>> +struct device *mdev_get_iommu_device(struct device *dev) >>>> +{ >>>> +    struct mdev_device *mdev = to_mdev_device(dev); >>>> + >>>> +    return mdev->iommu_device; >>>> +} >>>> +EXPORT_SYMBOL(mdev_get_iommu_device); >>>> + >>>> +int mdev_set_iommu_domain(struct device *dev, void *domain) >>>> +{ >>>> +    struct mdev_device *mdev = to_mdev_device(dev); >>>> + >>>> +    mdev->iommu_domain = domain; >>>> + >>>> +    return 0; >>>> +} >>>> +EXPORT_SYMBOL(mdev_set_iommu_domain); >>>> + >>>> +void *mdev_get_iommu_domain(struct device *dev) >>>> +{ >>>> +    struct mdev_device *mdev = to_mdev_device(dev); >>>> + >>>> +    return mdev->iommu_domain; >>>> +} >>>> +EXPORT_SYMBOL(mdev_get_iommu_domain); >>>> + >>>>   static int __init mdev_init(void) >>>>   { >>>>       return mdev_bus_register(); >>>> diff --git a/drivers/vfio/mdev/mdev_private.h >>>> b/drivers/vfio/mdev/mdev_private.h >>>> index b5819b7d7ef7..c01518068e84 100644 >>>> --- a/drivers/vfio/mdev/mdev_private.h >>>> +++ b/drivers/vfio/mdev/mdev_private.h >>>> @@ -34,6 +34,8 @@ struct mdev_device { >>>>       struct list_head next; >>>>       struct kobject *type_kobj; >>>>       bool active; >>>> +    struct device *iommu_device; >>>> +    void *iommu_domain; >>>>   }; >>>>     #define to_mdev_device(dev)    container_of(dev, struct >>>> mdev_device, dev) >>>> diff --git a/include/linux/mdev.h b/include/linux/mdev.h >>>> index b6e048e1045f..c46777d3e568 100644 >>>> --- a/include/linux/mdev.h >>>> +++ b/include/linux/mdev.h >>>> @@ -14,6 +14,29 @@ >>>>   #define MDEV_H >>>>     struct mdev_device; >>>> +struct iommu_domain; >>>> + >>>> +/* >>>> + * Called by the parent device driver to set the PCI device which >>>> represents >>> >>> s/PCI // >>> >>> There is no requirement or expectation that the device is PCI. >>> >> >> Fair enough. >> >>>> + * this mdev in iommu protection scope. By default, the iommu device >>>> is NULL, >>>> + * that indicates using vendor defined isolation. >>>> + * >>>> + * @dev: the mediated device that iommu will isolate. >>>> + * @iommu_device: a pci device which represents the iommu for @dev. >>>> + * >>>> + * Return 0 for success, otherwise negative error value. >>>> + */ >>>> +int mdev_set_iommu_device(struct device *dev, struct device >>>> *iommu_device); >>>> + >>>> +struct device *mdev_get_iommu_device(struct device *dev); >>>> + >>>> +/* >>>> + * Called by vfio iommu modules to save the iommu domain after a >>>> domain being >>>> + * attached to the mediated device. >>>> + */ >>>> +int mdev_set_iommu_domain(struct device *dev, void *domain); >>>> + >>>> +void *mdev_get_iommu_domain(struct device *dev); >>> >>> I can't say I really understand the purpose of this, the cover letter >>> indicates this is a placeholder, should we add it separately when we >>> have a requirement for it? >> >> Oh, I am sorry that I used a wrong word. It's not a placeholder for >> something designed for future, but adding two members that will be used >> in the following patches. Since they will be used in anther modules >> (like vfio_iommu), we need function interfaces to get and set them. >> >> mdev->iommu_device: >>  -  This, if set, indicates that the mediated device could >>     be fully isolated and protected by IOMMU via attaching >>     an iommu domain to this device. If empty, it indicates >>     using vendor defined isolation. >> >> mdev->iommu_domain: >>   - This is used to save the pointer of an iommu domain. Once >>     a domain has been attached to the iommu_device, it should >>     be stored here. >> > > I don't see mdev->iommu_domain is used anywhere in this series of patch. > If this is not being used, then no need to save it. With that symbols > mdev_set/get_iommu_domain(domain) are not required. Yes. We won't use mdev->iommu_domain in this patch series. It should be used by mdev parent driver to retrieve the default pasid of the domain. Something like: domain = mdev_get_iommu_domain(dev) pasid = iommu_domain_get_attr(domain, DOMAIN_ATTR_AUXD_ID); reg_write(pasid_reg, pasid); I am okay if we remove mdev_set/get_iommu_domain from this patch series and add it later when the parent driver comes. > > Please keep symbols mdev_set/get_iommu_device(dev, iommu_device) non-GPL > same as other exported symbols from mdev_core module. Yes. It will be fixed in the next version. > > Thanks, > Kirti > Best regards, Lu Baolu