From: Niklas Schnelle <schnelle@linux.ibm.com>
To: Matthew Rosato <mjrosato@linux.ibm.com>,
iommu@lists.linux.dev, Joerg Roedel <joro@8bytes.org>,
Will Deacon <will@kernel.org>,
Robin Murphy <robin.murphy@arm.com>,
Jason Gunthorpe <jgg@nvidia.com>
Cc: Gerd Bayer <gbayer@linux.ibm.com>,
Pierre Morel <pmorel@linux.ibm.com>,
linux-s390@vger.kernel.org, borntraeger@linux.ibm.com,
hca@linux.ibm.com, gor@linux.ibm.com,
gerald.schaefer@linux.ibm.com, agordeev@linux.ibm.com,
svens@linux.ibm.com, linux-kernel@vger.kernel.org
Subject: [PATCH v2 2/5] iommu/s390: Add I/O TLB ops
Date: Wed, 9 Nov 2022 15:29:00 +0100 [thread overview]
Message-ID: <20221109142903.4080275-3-schnelle@linux.ibm.com> (raw)
In-Reply-To: <20221109142903.4080275-1-schnelle@linux.ibm.com>
Currently s390-iommu does an I/O TLB flush (RPCIT) for every update of
the I/O translation table explicitly. For one this is wasteful since
RPCIT can be skipped after a mapping operation if zdev->tlb_refresh is
unset. Moreover we can do a single RPCIT for a range of pages including
whne doing lazy unmapping.
Thankfully both of these optimizations can be achieved by implementing
the IOMMU operations common code provides for the different types of I/O
tlb flushes:
* flush_iotlb_all: Flushes the I/O TLB for the entire IOVA space
* iotlb_sync: Flushes the I/O TLB for a range of pages that can be
gathered up, for example to implement lazy unmapping.
* iotlb_sync_map: Flushes the I/O TLB after a mapping operation
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
---
v1->v2:
- Don't skip IOTLB flushes for other devices on IOTLB flush failure (Jason)
drivers/iommu/s390-iommu.c | 67 +++++++++++++++++++++++++++++++-------
1 file changed, 56 insertions(+), 11 deletions(-)
diff --git a/drivers/iommu/s390-iommu.c b/drivers/iommu/s390-iommu.c
index e2c886bc4376..9771bce86e94 100644
--- a/drivers/iommu/s390-iommu.c
+++ b/drivers/iommu/s390-iommu.c
@@ -199,14 +199,63 @@ static void s390_iommu_release_device(struct device *dev)
__s390_iommu_detach_device(zdev);
}
+static void s390_iommu_flush_iotlb_all(struct iommu_domain *domain)
+{
+ struct s390_domain *s390_domain = to_s390_domain(domain);
+ struct zpci_dev *zdev;
+ unsigned long flags;
+
+ spin_lock_irqsave(&s390_domain->list_lock, flags);
+ list_for_each_entry(zdev, &s390_domain->devices, iommu_list) {
+ zpci_refresh_trans((u64)zdev->fh << 32, zdev->start_dma,
+ zdev->end_dma - zdev->start_dma + 1);
+ }
+ spin_unlock_irqrestore(&s390_domain->list_lock, flags);
+}
+
+static void s390_iommu_iotlb_sync(struct iommu_domain *domain,
+ struct iommu_iotlb_gather *gather)
+{
+ struct s390_domain *s390_domain = to_s390_domain(domain);
+ size_t size = gather->end - gather->start + 1;
+ struct zpci_dev *zdev;
+ unsigned long flags;
+
+ /* If gather was never added to there is nothing to flush */
+ if (!gather->end)
+ return;
+
+ spin_lock_irqsave(&s390_domain->list_lock, flags);
+ list_for_each_entry(zdev, &s390_domain->devices, iommu_list) {
+ zpci_refresh_trans((u64)zdev->fh << 32, gather->start,
+ size);
+ }
+ spin_unlock_irqrestore(&s390_domain->list_lock, flags);
+}
+
+static void s390_iommu_iotlb_sync_map(struct iommu_domain *domain,
+ unsigned long iova, size_t size)
+{
+ struct s390_domain *s390_domain = to_s390_domain(domain);
+ struct zpci_dev *zdev;
+ unsigned long flags;
+
+ spin_lock_irqsave(&s390_domain->list_lock, flags);
+ list_for_each_entry(zdev, &s390_domain->devices, iommu_list) {
+ if (!zdev->tlb_refresh)
+ continue;
+ zpci_refresh_trans((u64)zdev->fh << 32,
+ iova, size);
+ }
+ spin_unlock_irqrestore(&s390_domain->list_lock, flags);
+}
+
static int s390_iommu_update_trans(struct s390_domain *s390_domain,
phys_addr_t pa, dma_addr_t dma_addr,
unsigned long nr_pages, int flags)
{
phys_addr_t page_addr = pa & PAGE_MASK;
- dma_addr_t start_dma_addr = dma_addr;
unsigned long irq_flags, i;
- struct zpci_dev *zdev;
unsigned long *entry;
int rc = 0;
@@ -225,15 +274,6 @@ static int s390_iommu_update_trans(struct s390_domain *s390_domain,
dma_addr += PAGE_SIZE;
}
- spin_lock(&s390_domain->list_lock);
- list_for_each_entry(zdev, &s390_domain->devices, iommu_list) {
- rc = zpci_refresh_trans((u64)zdev->fh << 32,
- start_dma_addr, nr_pages * PAGE_SIZE);
- if (rc)
- break;
- }
- spin_unlock(&s390_domain->list_lock);
-
undo_cpu_trans:
if (rc && ((flags & ZPCI_PTE_VALID_MASK) == ZPCI_PTE_VALID)) {
flags = ZPCI_PTE_INVALID;
@@ -340,6 +380,8 @@ static size_t s390_iommu_unmap_pages(struct iommu_domain *domain,
if (rc)
return 0;
+ iommu_iotlb_gather_add_range(gather, iova, size);
+
return size;
}
@@ -384,6 +426,9 @@ static const struct iommu_ops s390_iommu_ops = {
.detach_dev = s390_iommu_detach_device,
.map_pages = s390_iommu_map_pages,
.unmap_pages = s390_iommu_unmap_pages,
+ .flush_iotlb_all = s390_iommu_flush_iotlb_all,
+ .iotlb_sync = s390_iommu_iotlb_sync,
+ .iotlb_sync_map = s390_iommu_iotlb_sync_map,
.iova_to_phys = s390_iommu_iova_to_phys,
.free = s390_domain_free,
}
--
2.34.1
next prev parent reply other threads:[~2022-11-09 14:29 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-11-09 14:28 [PATCH v2 0/5] iommu/s390: Further improvements Niklas Schnelle
2022-11-09 14:28 ` [PATCH v2 1/5] iommu/s390: Make attach succeed even if the device is in error state Niklas Schnelle
2022-11-14 15:05 ` Gerd Bayer
2022-11-14 15:30 ` Matthew Rosato
2022-11-09 14:29 ` Niklas Schnelle [this message]
2022-11-09 14:29 ` [PATCH v2 3/5] iommu/s390: Use RCU to allow concurrent domain_list iteration Niklas Schnelle
2022-11-09 14:29 ` [PATCH v2 4/5] iommu/s390: Optimize IOMMU table walking Niklas Schnelle
2022-11-09 14:29 ` [PATCH v2 5/5] s390/pci: use lock-free I/O translation updates Niklas Schnelle
2022-11-19 9:28 ` [PATCH v2 0/5] iommu/s390: Further improvements 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=20221109142903.4080275-3-schnelle@linux.ibm.com \
--to=schnelle@linux.ibm.com \
--cc=agordeev@linux.ibm.com \
--cc=borntraeger@linux.ibm.com \
--cc=gbayer@linux.ibm.com \
--cc=gerald.schaefer@linux.ibm.com \
--cc=gor@linux.ibm.com \
--cc=hca@linux.ibm.com \
--cc=iommu@lists.linux.dev \
--cc=jgg@nvidia.com \
--cc=joro@8bytes.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-s390@vger.kernel.org \
--cc=mjrosato@linux.ibm.com \
--cc=pmorel@linux.ibm.com \
--cc=robin.murphy@arm.com \
--cc=svens@linux.ibm.com \
--cc=will@kernel.org \
/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