From: Robin Murphy <robin.murphy@arm.com>
To: Joerg Roedel <joro@8bytes.org>, Eric Auger <eric.auger@redhat.com>
Cc: eric.auger.pro@gmail.com, christoffer.dall@linaro.org,
marc.zyngier@arm.com, alex.williamson@redhat.com,
will.deacon@arm.com, tglx@linutronix.de, jason@lakedaemon.net,
linux-arm-kernel@lists.infradead.org, kvm@vger.kernel.org,
drjones@redhat.com, linux-kernel@vger.kernel.org,
pranav.sawargaonkar@gmail.com, iommu@lists.linux-foundation.org,
punit.agrawal@arm.com, diana.craciun@nxp.com
Subject: Re: [RFC v2 8/8] iommu/arm-smmu: implement add_reserved_regions callback
Date: Thu, 10 Nov 2016 16:07:08 +0000 [thread overview]
Message-ID: <b347ddbc-5066-641c-8101-289fb28624a3@arm.com> (raw)
In-Reply-To: <20161110154606.GH2078@8bytes.org>
On 10/11/16 15:46, Joerg Roedel wrote:
> On Fri, Nov 04, 2016 at 11:24:06AM +0000, Eric Auger wrote:
>> The function populates the list of reserved regions with the
>> PCI host bridge windows and the MSI IOVA range.
>>
>> At the moment an arbitray MSI IOVA window is set at 0x8000000
>> of size 1MB.
>>
>> Signed-off-by: Eric Auger <eric.auger@redhat.com>
>>
>> ---
>>
>> RFC v1 -> v2: use defines for MSI IOVA base and length
>> ---
>> drivers/iommu/arm-smmu.c | 66 ++++++++++++++++++++++++++++++++++++++++++++++++
>> 1 file changed, 66 insertions(+)
>>
>> diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c
>> index c841eb7..c07ea41 100644
>> --- a/drivers/iommu/arm-smmu.c
>> +++ b/drivers/iommu/arm-smmu.c
>> @@ -278,6 +278,9 @@ enum arm_smmu_s2cr_privcfg {
>>
>> #define FSYNR0_WNR (1 << 4)
>>
>> +#define MSI_IOVA_BASE 0x8000000
>> +#define MSI_IOVA_LENGTH 0x100000
>> +
>> static int force_stage;
>> module_param(force_stage, int, S_IRUGO);
>> MODULE_PARM_DESC(force_stage,
>> @@ -1533,6 +1536,68 @@ static int arm_smmu_of_xlate(struct device *dev, struct of_phandle_args *args)
>> return iommu_fwspec_add_ids(dev, &fwid, 1);
>> }
>>
>> +static int add_pci_window_reserved_regions(struct iommu_domain *domain,
>> + struct pci_dev *dev)
>> +{
>> + struct pci_host_bridge *bridge = pci_find_host_bridge(dev->bus);
>> + struct iommu_reserved_region *region;
>> + struct resource_entry *window;
>> + phys_addr_t start;
>> + size_t length;
>> +
>> + resource_list_for_each_entry(window, &bridge->windows) {
>> + if (resource_type(window->res) != IORESOURCE_MEM &&
>> + resource_type(window->res) != IORESOURCE_IO)
>> + continue;
>
> Why do you care about IO resources?
[since this is essentially code I wrote]
Because they occupy some area of the PCI address space, therefore I
assumed that, like memory windows, they would be treated as P2P. Is that
not the case?
>> +
>> + start = window->res->start - window->offset;
>> + length = window->res->end - window->res->start + 1;
>> +
>> + iommu_reserved_region_for_each(region, domain) {
>> + if (region->start == start && region->length == length)
>> + continue;
>> + }
>> + region = kzalloc(sizeof(*region), GFP_KERNEL);
>> + if (!region)
>> + return -ENOMEM;
>> +
>> + region->start = start;
>> + region->length = length;
>> +
>> + list_add_tail(®ion->list, &domain->reserved_regions);
>> + }
>> + return 0;
>> +}
>> +
>> +static int arm_smmu_add_reserved_regions(struct iommu_domain *domain,
>> + struct device *device)
>> +{
>> + struct iommu_reserved_region *region;
>> + int ret = 0;
>> +
>> + /* An arbitrary 1MB region starting at 0x8000000 is reserved for MSIs */
>> + if (!domain->iova_cookie) {
>> +
>> + region = kzalloc(sizeof(*region), GFP_KERNEL);
>> + if (!region)
>> + return -ENOMEM;
>> +
>> + region->start = MSI_IOVA_BASE;
>> + region->length = MSI_IOVA_LENGTH;
>> + list_add_tail(®ion->list, &domain->reserved_regions);
>> +
>> + ret = iommu_get_dma_msi_region_cookie(domain,
>> + region->start, region->length);
>> + if (ret)
>> + return ret;
>
> Gah, I hate this. Is there no other and simpler way to get the MSI
> region than allocating an iova-domain? In that regard, I also _hate_ the
> patch before introducing this weird iommu_get_dma_msi_region_cookie()
> function.
Mea culpa ;)
> Allocation an iova-domain is pretty expensive, as it also includes
> per-cpu data-structures and all, so please don't do this just for the
> purpose of compiling a list of reserved regions.
Yeah, this was really more of a "get it working" thing - the automatic
MSI mapping stuff is already plumbed in for DMA domains, so faking up a
DMA domain is the easy way in. I'll have a look at factoring out the
relevant parts a bit so that MSI-only regions can have an alternate
cookie with a lightweight allocator.
Robin.
next prev parent reply other threads:[~2016-11-10 16:07 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-11-04 11:23 [RFC v2 0/8] KVM PCIe/MSI passthrough on ARM/ARM64 - Alt II Eric Auger
2016-11-04 11:23 ` [RFC v2 1/8] vfio: fix vfio_info_cap_add/shift Eric Auger
2016-11-04 11:24 ` [RFC v2 2/8] iommu/iova: fix __alloc_and_insert_iova_range Eric Auger
2016-11-10 15:22 ` Joerg Roedel
2016-11-10 15:41 ` Auger Eric
2016-11-04 11:24 ` [RFC v2 3/8] iommu/dma: Allow MSI-only cookies Eric Auger
2016-11-14 12:36 ` Robin Murphy
2016-11-14 23:23 ` Auger Eric
2016-11-15 14:52 ` Robin Murphy
2016-11-04 11:24 ` [RFC v2 4/8] iommu: Add a list of iommu_reserved_region in iommu_domain Eric Auger
2016-11-04 14:00 ` Robin Murphy
2016-11-10 11:22 ` Auger Eric
2016-11-10 11:54 ` Robin Murphy
2016-11-10 12:14 ` Auger Eric
2016-11-10 12:48 ` Robin Murphy
2016-11-10 15:37 ` Joerg Roedel
2016-11-10 15:42 ` Auger Eric
2016-11-04 11:24 ` [RFC v2 5/8] vfio/type1: Introduce RESV_IOVA_RANGE capability Eric Auger
2016-11-04 11:24 ` [RFC v2 6/8] iommu: Handle the list of reserved regions Eric Auger
2016-11-04 11:24 ` [RFC v2 7/8] iommu/vt-d: Implement add_reserved_regions callback Eric Auger
2016-11-04 11:24 ` [RFC v2 8/8] iommu/arm-smmu: implement " Eric Auger
2016-11-04 14:16 ` Robin Murphy
2016-11-10 15:46 ` Joerg Roedel
2016-11-10 15:57 ` Auger Eric
2016-11-10 16:13 ` Joerg Roedel
2016-11-10 18:00 ` Auger Eric
2016-11-11 11:42 ` Joerg Roedel
2016-11-11 15:47 ` Auger Eric
2016-11-11 16:22 ` Joerg Roedel
2016-11-11 16:45 ` Auger Eric
2016-11-14 15:31 ` Joerg Roedel
2016-11-14 16:08 ` Auger Eric
2016-11-14 16:20 ` Joerg Roedel
2016-11-14 16:57 ` Auger Eric
2016-11-10 16:07 ` Robin Murphy [this message]
2016-11-10 16:16 ` Joerg Roedel
2016-11-11 14:34 ` Robin Murphy
2016-11-11 15:03 ` Joerg Roedel
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=b347ddbc-5066-641c-8101-289fb28624a3@arm.com \
--to=robin.murphy@arm.com \
--cc=alex.williamson@redhat.com \
--cc=christoffer.dall@linaro.org \
--cc=diana.craciun@nxp.com \
--cc=drjones@redhat.com \
--cc=eric.auger.pro@gmail.com \
--cc=eric.auger@redhat.com \
--cc=iommu@lists.linux-foundation.org \
--cc=jason@lakedaemon.net \
--cc=joro@8bytes.org \
--cc=kvm@vger.kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=marc.zyngier@arm.com \
--cc=pranav.sawargaonkar@gmail.com \
--cc=punit.agrawal@arm.com \
--cc=tglx@linutronix.de \
--cc=will.deacon@arm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®