All of lore.kernel.org
 help / color / mirror / Atom feed
From: Robin Murphy <robin.murphy@arm.com>
To: Will Deacon <will@kernel.org>, Christoph Hellwig <hch@lst.de>
Cc: heikki.krogerus@linux.intel.com,
	thomas.hellstrom@linux.intel.com, peterz@infradead.org,
	benh@kernel.crashing.org, joonas.lahtinen@linux.intel.com,
	dri-devel@lists.freedesktop.org, chris@chris-wilson.co.uk,
	grant.likely@arm.com, paulus@samba.org,
	Frank Rowand <frowand.list@gmail.com>,
	mingo@kernel.org, Stefano Stabellini <sstabellini@kernel.org>,
	Saravana Kannan <saravanak@google.com>,
	mpe@ellerman.id.au,
	"Rafael J . Wysocki" <rafael.j.wysocki@intel.com>,
	Bartosz Golaszewski <bgolaszewski@baylibre.com>,
	bskeggs@redhat.com, linux-pci@vger.kernel.org,
	xen-devel@lists.xenproject.org,
	Thierry Reding <treding@nvidia.com>,
	intel-gfx@lists.freedesktop.org, matthew.auld@intel.com,
	linux-devicetree <devicetree@vger.kernel.org>,
	Jianxiong Gao <jxgao@google.com>, Daniel Vetter <daniel@ffwll.ch>,
	Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>,
	maarten.lankhorst@linux.intel.com, airlied@linux.ie,
	Dan Williams <dan.j.williams@intel.com>,
	linuxppc-dev@lists.ozlabs.org, jani.nikula@linux.intel.com,
	Nathan Chancellor <nathan@kernel.org>,
	Rob Herring <robh+dt@kernel.org>,
	rodrigo.vivi@intel.com, Bjorn Helgaas <bhelgaas@google.com>,
	Claire Chang <tientzu@chromium.org>,
	boris.ostrovsky@oracle.com,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	jgross@suse.com, Nicolas Boichat <drinkcat@chromium.org>,
	Greg KH <gregkh@linuxfoundation.org>,
	Randy Dunlap <rdunlap@infradead.org>,
	Qian Cai <quic_qiancai@quicinc.com>,
	lkml <linux-kernel@vger.kernel.org>,
	"list@263.net:IOMMU DRIVERS" <iommu@lists.linux-foundation.org>,
	Jim Quinlan <james.quinlan@broadcom.com>,
	xypron.glpk@gmx.de, Tom Lendacky <thomas.lendacky@amd.com>,
	bauerman@linux.ibm.com
Subject: Re: [PATCH v15 06/12] swiotlb: Use is_swiotlb_force_bounce for swiotlb data bouncing
Date: Tue, 6 Jul 2021 15:01:04 +0100	[thread overview]
Message-ID: <a59f771f-3289-62f0-ca50-8f3675d9b166@arm.com> (raw)
In-Reply-To: <20210706132422.GA20327@willie-the-truck>

On 2021-07-06 14:24, Will Deacon wrote:
> On Tue, Jul 06, 2021 at 06:48:48AM +0200, Christoph Hellwig wrote:
>> On Mon, Jul 05, 2021 at 08:03:52PM +0100, Will Deacon wrote:
>>> So at this point, the AMD IOMMU driver does:
>>>
>>> 	swiotlb        = (iommu_default_passthrough() || sme_me_mask) ? 1 : 0;
>>>
>>> where 'swiotlb' is a global variable indicating whether or not swiotlb
>>> is in use. It's picked up a bit later on by pci_swiotlb_late_init(), which
>>> will call swiotlb_exit() if 'swiotlb' is false.
>>>
>>> Now, that used to work fine, because swiotlb_exit() clears
>>> 'io_tlb_default_mem' to NULL, but now with the restricted DMA changes, I
>>> think that all the devices which have successfully probed beforehand will
>>> have stale pointers to the freed structure in their 'dev->dma_io_tlb_mem'
>>> field.
>>
>> Yeah.  I don't think we can do that anymore, and I also think it is
>> a bad idea to start with.
> 
> I've had a crack at reworking things along the following lines:
> 
>    - io_tlb_default_mem now lives in the BSS, the flexible array member
>      is now a pointer and that part is allocated dynamically (downside of
>      this is an extra indirection to get at the slots).
> 
>    - io_tlb_default_mem.nslabs tells you whether the thing is valid
> 
>    - swiotlb_exit() frees the slots array and clears the rest of the
>      structure to 0. I also extended it to free the actual slabs, but I'm
>      not sure why it wasn't doing that before.
> 
> So a non-NULL dev->dma_io_tlb_mem should always be valid to follow.

FWIW I was pondering the question of whether to do something along those 
lines or just scrap the default assignment entirely, so since I hadn't 
got round to saying that I've gone ahead and hacked up the alternative 
(similarly untested) for comparison :)

TBH I'm still not sure which one I prefer...

Robin.

----->8-----
diff --git a/drivers/base/core.c b/drivers/base/core.c
index ea5b85354526..394abf184c1a 100644
--- a/drivers/base/core.c
+++ b/drivers/base/core.c
@@ -2847,9 +2847,6 @@ void device_initialize(struct device *dev)
      defined(CONFIG_ARCH_HAS_SYNC_DMA_FOR_CPU_ALL)
  	dev->dma_coherent = dma_default_coherent;
  #endif
-#ifdef CONFIG_SWIOTLB
-	dev->dma_io_tlb_mem = io_tlb_default_mem;
-#endif
  }
  EXPORT_SYMBOL_GPL(device_initialize);

diff --git a/include/linux/swiotlb.h b/include/linux/swiotlb.h
index 39284ff2a6cd..620f16d89a98 100644
--- a/include/linux/swiotlb.h
+++ b/include/linux/swiotlb.h
@@ -107,16 +107,21 @@ struct io_tlb_mem {
  };
  extern struct io_tlb_mem *io_tlb_default_mem;

+static inline struct io_tlb_mem *dev_iotlb_mem(struct device *dev)
+{
+	return dev->dma_io_tlb_mem ?: io_tlb_default_mem;
+}
+
  static inline bool is_swiotlb_buffer(struct device *dev, phys_addr_t 
paddr)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);

  	return mem && paddr >= mem->start && paddr < mem->end;
  }

  static inline bool is_swiotlb_force_bounce(struct device *dev)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);

  	return mem && mem->force_bounce;
  }
@@ -167,7 +172,7 @@ bool swiotlb_free(struct device *dev, struct page 
*page, size_t size);

  static inline bool is_swiotlb_for_alloc(struct device *dev)
  {
-	return dev->dma_io_tlb_mem->for_alloc;
+	return dev_iotlb_mem(dev)->for_alloc;
  }
  #else
  static inline struct page *swiotlb_alloc(struct device *dev, size_t size)
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index b7f76bca89bf..f4942149f87d 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -359,7 +359,7 @@ static unsigned int swiotlb_align_offset(struct 
device *dev, u64 addr)
  static void swiotlb_bounce(struct device *dev, phys_addr_t tlb_addr, 
size_t size,
  			   enum dma_data_direction dir)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	int index = (tlb_addr - mem->start) >> IO_TLB_SHIFT;
  	phys_addr_t orig_addr = mem->slots[index].orig_addr;
  	size_t alloc_size = mem->slots[index].alloc_size;
@@ -440,7 +440,7 @@ static unsigned int wrap_index(struct io_tlb_mem 
*mem, unsigned int index)
  static int swiotlb_find_slots(struct device *dev, phys_addr_t orig_addr,
  			      size_t alloc_size)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned long boundary_mask = dma_get_seg_boundary(dev);
  	dma_addr_t tbl_dma_addr =
  		phys_to_dma_unencrypted(dev, mem->start) & boundary_mask;
@@ -522,7 +522,7 @@ phys_addr_t swiotlb_tbl_map_single(struct device 
*dev, phys_addr_t orig_addr,
  		size_t mapping_size, size_t alloc_size,
  		enum dma_data_direction dir, unsigned long attrs)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned int offset = swiotlb_align_offset(dev, orig_addr);
  	unsigned int i;
  	int index;
@@ -565,7 +565,7 @@ phys_addr_t swiotlb_tbl_map_single(struct device 
*dev, phys_addr_t orig_addr,

  static void swiotlb_release_slots(struct device *dev, phys_addr_t 
tlb_addr)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned long flags;
  	unsigned int offset = swiotlb_align_offset(dev, tlb_addr);
  	int index = (tlb_addr - offset - mem->start) >> IO_TLB_SHIFT;
@@ -682,7 +682,7 @@ size_t swiotlb_max_mapping_size(struct device *dev)

  bool is_swiotlb_active(struct device *dev)
  {
-	return dev->dma_io_tlb_mem != NULL;
+	return dev_iotlb_mem(dev) != NULL;
  }
  EXPORT_SYMBOL_GPL(is_swiotlb_active);

@@ -729,7 +729,7 @@ static void rmem_swiotlb_debugfs_init(struct 
reserved_mem *rmem)

  struct page *swiotlb_alloc(struct device *dev, size_t size)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	phys_addr_t tlb_addr;
  	int index;

@@ -792,7 +792,7 @@ static int rmem_swiotlb_device_init(struct 
reserved_mem *rmem,
  static void rmem_swiotlb_device_release(struct reserved_mem *rmem,
  					struct device *dev)
  {
-	dev->dma_io_tlb_mem = io_tlb_default_mem;
+	dev->dma_io_tlb_mem = NULL;
  }

  static const struct reserved_mem_ops rmem_swiotlb_ops = {

WARNING: multiple messages have this Message-ID (diff)
From: Robin Murphy <robin.murphy@arm.com>
To: Will Deacon <will@kernel.org>, Christoph Hellwig <hch@lst.de>
Cc: Jim Quinlan <james.quinlan@broadcom.com>,
	heikki.krogerus@linux.intel.com,
	linux-devicetree <devicetree@vger.kernel.org>,
	peterz@infradead.org, joonas.lahtinen@linux.intel.com,
	dri-devel@lists.freedesktop.org, chris@chris-wilson.co.uk,
	grant.likely@arm.com, paulus@samba.org,
	Frank Rowand <frowand.list@gmail.com>,
	mingo@kernel.org, Jianxiong Gao <jxgao@google.com>,
	Stefano Stabellini <sstabellini@kernel.org>,
	Saravana Kannan <saravanak@google.com>,
	"Rafael J . Wysocki" <rafael.j.wysocki@intel.com>,
	Bartosz Golaszewski <bgolaszewski@baylibre.com>,
	matthew.auld@intel.com, linux-pci@vger.kernel.org,
	xen-devel@lists.xenproject.org,
	Thierry Reding <treding@nvidia.com>,
	bskeggs@redhat.com, Nicolas Boichat <drinkcat@chromium.org>,
	thomas.hellstrom@linux.intel.com, jgross@suse.com,
	Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>,
	intel-gfx@lists.freedesktop.org,
	maarten.lankhorst@linux.intel.com, jani.nikula@linux.intel.com,
	Nathan Chancellor <nathan@kernel.org>,
	Rob Herring <robh+dt@kernel.org>,
	rodrigo.vivi@intel.com, Bjorn Helgaas <bhelgaas@google.com>,
	Claire Chang <tientzu@chromium.org>,
	Dan Williams <dan.j.williams@intel.com>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	boris.ostrovsky@oracle.com, airlied@linux.ie,
	Greg KH <gregkh@linuxfoundation.org>,
	Randy Dunlap <rdunlap@infradead.org>,
	Qian Cai <quic_qiancai@quicinc.com>,
	lkml <linux-kernel@vger.kernel.org>,
	"list@263.net:IOMMU DRIVERS" <iommu@lists.linux-foundation.org>,
	Daniel Vetter <daniel@ffwll.ch>,
	xypron.glpk@gmx.de, Tom Lendacky <thomas.lendacky@amd.com>,
	linuxppc-dev@lists.ozlabs.org, bauerman@linux.ibm.com
Subject: Re: [PATCH v15 06/12] swiotlb: Use is_swiotlb_force_bounce for swiotlb data bouncing
Date: Tue, 6 Jul 2021 15:01:04 +0100	[thread overview]
Message-ID: <a59f771f-3289-62f0-ca50-8f3675d9b166@arm.com> (raw)
In-Reply-To: <20210706132422.GA20327@willie-the-truck>

On 2021-07-06 14:24, Will Deacon wrote:
> On Tue, Jul 06, 2021 at 06:48:48AM +0200, Christoph Hellwig wrote:
>> On Mon, Jul 05, 2021 at 08:03:52PM +0100, Will Deacon wrote:
>>> So at this point, the AMD IOMMU driver does:
>>>
>>> 	swiotlb        = (iommu_default_passthrough() || sme_me_mask) ? 1 : 0;
>>>
>>> where 'swiotlb' is a global variable indicating whether or not swiotlb
>>> is in use. It's picked up a bit later on by pci_swiotlb_late_init(), which
>>> will call swiotlb_exit() if 'swiotlb' is false.
>>>
>>> Now, that used to work fine, because swiotlb_exit() clears
>>> 'io_tlb_default_mem' to NULL, but now with the restricted DMA changes, I
>>> think that all the devices which have successfully probed beforehand will
>>> have stale pointers to the freed structure in their 'dev->dma_io_tlb_mem'
>>> field.
>>
>> Yeah.  I don't think we can do that anymore, and I also think it is
>> a bad idea to start with.
> 
> I've had a crack at reworking things along the following lines:
> 
>    - io_tlb_default_mem now lives in the BSS, the flexible array member
>      is now a pointer and that part is allocated dynamically (downside of
>      this is an extra indirection to get at the slots).
> 
>    - io_tlb_default_mem.nslabs tells you whether the thing is valid
> 
>    - swiotlb_exit() frees the slots array and clears the rest of the
>      structure to 0. I also extended it to free the actual slabs, but I'm
>      not sure why it wasn't doing that before.
> 
> So a non-NULL dev->dma_io_tlb_mem should always be valid to follow.

FWIW I was pondering the question of whether to do something along those 
lines or just scrap the default assignment entirely, so since I hadn't 
got round to saying that I've gone ahead and hacked up the alternative 
(similarly untested) for comparison :)

TBH I'm still not sure which one I prefer...

Robin.

----->8-----
diff --git a/drivers/base/core.c b/drivers/base/core.c
index ea5b85354526..394abf184c1a 100644
--- a/drivers/base/core.c
+++ b/drivers/base/core.c
@@ -2847,9 +2847,6 @@ void device_initialize(struct device *dev)
      defined(CONFIG_ARCH_HAS_SYNC_DMA_FOR_CPU_ALL)
  	dev->dma_coherent = dma_default_coherent;
  #endif
-#ifdef CONFIG_SWIOTLB
-	dev->dma_io_tlb_mem = io_tlb_default_mem;
-#endif
  }
  EXPORT_SYMBOL_GPL(device_initialize);

diff --git a/include/linux/swiotlb.h b/include/linux/swiotlb.h
index 39284ff2a6cd..620f16d89a98 100644
--- a/include/linux/swiotlb.h
+++ b/include/linux/swiotlb.h
@@ -107,16 +107,21 @@ struct io_tlb_mem {
  };
  extern struct io_tlb_mem *io_tlb_default_mem;

+static inline struct io_tlb_mem *dev_iotlb_mem(struct device *dev)
+{
+	return dev->dma_io_tlb_mem ?: io_tlb_default_mem;
+}
+
  static inline bool is_swiotlb_buffer(struct device *dev, phys_addr_t 
paddr)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);

  	return mem && paddr >= mem->start && paddr < mem->end;
  }

  static inline bool is_swiotlb_force_bounce(struct device *dev)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);

  	return mem && mem->force_bounce;
  }
@@ -167,7 +172,7 @@ bool swiotlb_free(struct device *dev, struct page 
*page, size_t size);

  static inline bool is_swiotlb_for_alloc(struct device *dev)
  {
-	return dev->dma_io_tlb_mem->for_alloc;
+	return dev_iotlb_mem(dev)->for_alloc;
  }
  #else
  static inline struct page *swiotlb_alloc(struct device *dev, size_t size)
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index b7f76bca89bf..f4942149f87d 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -359,7 +359,7 @@ static unsigned int swiotlb_align_offset(struct 
device *dev, u64 addr)
  static void swiotlb_bounce(struct device *dev, phys_addr_t tlb_addr, 
size_t size,
  			   enum dma_data_direction dir)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	int index = (tlb_addr - mem->start) >> IO_TLB_SHIFT;
  	phys_addr_t orig_addr = mem->slots[index].orig_addr;
  	size_t alloc_size = mem->slots[index].alloc_size;
@@ -440,7 +440,7 @@ static unsigned int wrap_index(struct io_tlb_mem 
*mem, unsigned int index)
  static int swiotlb_find_slots(struct device *dev, phys_addr_t orig_addr,
  			      size_t alloc_size)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned long boundary_mask = dma_get_seg_boundary(dev);
  	dma_addr_t tbl_dma_addr =
  		phys_to_dma_unencrypted(dev, mem->start) & boundary_mask;
@@ -522,7 +522,7 @@ phys_addr_t swiotlb_tbl_map_single(struct device 
*dev, phys_addr_t orig_addr,
  		size_t mapping_size, size_t alloc_size,
  		enum dma_data_direction dir, unsigned long attrs)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned int offset = swiotlb_align_offset(dev, orig_addr);
  	unsigned int i;
  	int index;
@@ -565,7 +565,7 @@ phys_addr_t swiotlb_tbl_map_single(struct device 
*dev, phys_addr_t orig_addr,

  static void swiotlb_release_slots(struct device *dev, phys_addr_t 
tlb_addr)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned long flags;
  	unsigned int offset = swiotlb_align_offset(dev, tlb_addr);
  	int index = (tlb_addr - offset - mem->start) >> IO_TLB_SHIFT;
@@ -682,7 +682,7 @@ size_t swiotlb_max_mapping_size(struct device *dev)

  bool is_swiotlb_active(struct device *dev)
  {
-	return dev->dma_io_tlb_mem != NULL;
+	return dev_iotlb_mem(dev) != NULL;
  }
  EXPORT_SYMBOL_GPL(is_swiotlb_active);

@@ -729,7 +729,7 @@ static void rmem_swiotlb_debugfs_init(struct 
reserved_mem *rmem)

  struct page *swiotlb_alloc(struct device *dev, size_t size)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	phys_addr_t tlb_addr;
  	int index;

@@ -792,7 +792,7 @@ static int rmem_swiotlb_device_init(struct 
reserved_mem *rmem,
  static void rmem_swiotlb_device_release(struct reserved_mem *rmem,
  					struct device *dev)
  {
-	dev->dma_io_tlb_mem = io_tlb_default_mem;
+	dev->dma_io_tlb_mem = NULL;
  }

  static const struct reserved_mem_ops rmem_swiotlb_ops = {

WARNING: multiple messages have this Message-ID (diff)
From: Robin Murphy <robin.murphy@arm.com>
To: Will Deacon <will@kernel.org>, Christoph Hellwig <hch@lst.de>
Cc: Jim Quinlan <james.quinlan@broadcom.com>,
	heikki.krogerus@linux.intel.com,
	linux-devicetree <devicetree@vger.kernel.org>,
	peterz@infradead.org, benh@kernel.crashing.org,
	joonas.lahtinen@linux.intel.com, dri-devel@lists.freedesktop.org,
	chris@chris-wilson.co.uk, grant.likely@arm.com, paulus@samba.org,
	Frank Rowand <frowand.list@gmail.com>,
	mingo@kernel.org, Jianxiong Gao <jxgao@google.com>,
	Stefano Stabellini <sstabellini@kernel.org>,
	Saravana Kannan <saravanak@google.com>,
	mpe@ellerman.id.au,
	"Rafael J . Wysocki" <rafael.j.wysocki@intel.com>,
	Bartosz Golaszewski <bgolaszewski@baylibre.com>,
	matthew.auld@intel.com, linux-pci@vger.kernel.org,
	xen-devel@lists.xenproject.org,
	Thierry Reding <treding@nvidia.com>,
	bskeggs@redhat.com, Nicolas Boichat <drinkcat@chromium.org>,
	thomas.hellstrom@linux.intel.com, jgross@suse.com,
	Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>,
	intel-gfx@lists.freedesktop.org,
	maarten.lankhorst@linux.intel.com, jani.nikula@linux.intel.com,
	Nathan Chancellor <nathan@kernel.org>,
	Rob Herring <robh+dt@kernel.org>,
	rodrigo.vivi@intel.com, Bjorn Helgaas <bhelgaas@google.com>,
	Claire Chang <tientzu@chromium.org>,
	Dan Williams <dan.j.williams@intel.com>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	boris.ostrovsky@oracle.com, airlied@linux.ie,
	Greg KH <gregkh@linuxfoundation.org>,
	Randy Dunlap <rdunlap@infradead.org>,
	Qian Cai <quic_qiancai@quicinc.com>,
	lkml <linux-kernel@vger.kernel.org>,
	"list@263.net:IOMMU DRIVERS" <iommu@lists.linux-foundation.org>,
	Daniel Vetter <daniel@ffwll.ch>,
	xypron.glpk@gmx.de, Tom Lendacky <thomas.lendacky@amd.com>,
	linuxppc-dev@lists.ozlabs.org, bauerman@linux.ibm.com
Subject: Re: [PATCH v15 06/12] swiotlb: Use is_swiotlb_force_bounce for swiotlb data bouncing
Date: Tue, 6 Jul 2021 15:01:04 +0100	[thread overview]
Message-ID: <a59f771f-3289-62f0-ca50-8f3675d9b166@arm.com> (raw)
In-Reply-To: <20210706132422.GA20327@willie-the-truck>

On 2021-07-06 14:24, Will Deacon wrote:
> On Tue, Jul 06, 2021 at 06:48:48AM +0200, Christoph Hellwig wrote:
>> On Mon, Jul 05, 2021 at 08:03:52PM +0100, Will Deacon wrote:
>>> So at this point, the AMD IOMMU driver does:
>>>
>>> 	swiotlb        = (iommu_default_passthrough() || sme_me_mask) ? 1 : 0;
>>>
>>> where 'swiotlb' is a global variable indicating whether or not swiotlb
>>> is in use. It's picked up a bit later on by pci_swiotlb_late_init(), which
>>> will call swiotlb_exit() if 'swiotlb' is false.
>>>
>>> Now, that used to work fine, because swiotlb_exit() clears
>>> 'io_tlb_default_mem' to NULL, but now with the restricted DMA changes, I
>>> think that all the devices which have successfully probed beforehand will
>>> have stale pointers to the freed structure in their 'dev->dma_io_tlb_mem'
>>> field.
>>
>> Yeah.  I don't think we can do that anymore, and I also think it is
>> a bad idea to start with.
> 
> I've had a crack at reworking things along the following lines:
> 
>    - io_tlb_default_mem now lives in the BSS, the flexible array member
>      is now a pointer and that part is allocated dynamically (downside of
>      this is an extra indirection to get at the slots).
> 
>    - io_tlb_default_mem.nslabs tells you whether the thing is valid
> 
>    - swiotlb_exit() frees the slots array and clears the rest of the
>      structure to 0. I also extended it to free the actual slabs, but I'm
>      not sure why it wasn't doing that before.
> 
> So a non-NULL dev->dma_io_tlb_mem should always be valid to follow.

FWIW I was pondering the question of whether to do something along those 
lines or just scrap the default assignment entirely, so since I hadn't 
got round to saying that I've gone ahead and hacked up the alternative 
(similarly untested) for comparison :)

TBH I'm still not sure which one I prefer...

Robin.

----->8-----
diff --git a/drivers/base/core.c b/drivers/base/core.c
index ea5b85354526..394abf184c1a 100644
--- a/drivers/base/core.c
+++ b/drivers/base/core.c
@@ -2847,9 +2847,6 @@ void device_initialize(struct device *dev)
      defined(CONFIG_ARCH_HAS_SYNC_DMA_FOR_CPU_ALL)
  	dev->dma_coherent = dma_default_coherent;
  #endif
-#ifdef CONFIG_SWIOTLB
-	dev->dma_io_tlb_mem = io_tlb_default_mem;
-#endif
  }
  EXPORT_SYMBOL_GPL(device_initialize);

diff --git a/include/linux/swiotlb.h b/include/linux/swiotlb.h
index 39284ff2a6cd..620f16d89a98 100644
--- a/include/linux/swiotlb.h
+++ b/include/linux/swiotlb.h
@@ -107,16 +107,21 @@ struct io_tlb_mem {
  };
  extern struct io_tlb_mem *io_tlb_default_mem;

+static inline struct io_tlb_mem *dev_iotlb_mem(struct device *dev)
+{
+	return dev->dma_io_tlb_mem ?: io_tlb_default_mem;
+}
+
  static inline bool is_swiotlb_buffer(struct device *dev, phys_addr_t 
paddr)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);

  	return mem && paddr >= mem->start && paddr < mem->end;
  }

  static inline bool is_swiotlb_force_bounce(struct device *dev)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);

  	return mem && mem->force_bounce;
  }
@@ -167,7 +172,7 @@ bool swiotlb_free(struct device *dev, struct page 
*page, size_t size);

  static inline bool is_swiotlb_for_alloc(struct device *dev)
  {
-	return dev->dma_io_tlb_mem->for_alloc;
+	return dev_iotlb_mem(dev)->for_alloc;
  }
  #else
  static inline struct page *swiotlb_alloc(struct device *dev, size_t size)
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index b7f76bca89bf..f4942149f87d 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -359,7 +359,7 @@ static unsigned int swiotlb_align_offset(struct 
device *dev, u64 addr)
  static void swiotlb_bounce(struct device *dev, phys_addr_t tlb_addr, 
size_t size,
  			   enum dma_data_direction dir)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	int index = (tlb_addr - mem->start) >> IO_TLB_SHIFT;
  	phys_addr_t orig_addr = mem->slots[index].orig_addr;
  	size_t alloc_size = mem->slots[index].alloc_size;
@@ -440,7 +440,7 @@ static unsigned int wrap_index(struct io_tlb_mem 
*mem, unsigned int index)
  static int swiotlb_find_slots(struct device *dev, phys_addr_t orig_addr,
  			      size_t alloc_size)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned long boundary_mask = dma_get_seg_boundary(dev);
  	dma_addr_t tbl_dma_addr =
  		phys_to_dma_unencrypted(dev, mem->start) & boundary_mask;
@@ -522,7 +522,7 @@ phys_addr_t swiotlb_tbl_map_single(struct device 
*dev, phys_addr_t orig_addr,
  		size_t mapping_size, size_t alloc_size,
  		enum dma_data_direction dir, unsigned long attrs)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned int offset = swiotlb_align_offset(dev, orig_addr);
  	unsigned int i;
  	int index;
@@ -565,7 +565,7 @@ phys_addr_t swiotlb_tbl_map_single(struct device 
*dev, phys_addr_t orig_addr,

  static void swiotlb_release_slots(struct device *dev, phys_addr_t 
tlb_addr)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned long flags;
  	unsigned int offset = swiotlb_align_offset(dev, tlb_addr);
  	int index = (tlb_addr - offset - mem->start) >> IO_TLB_SHIFT;
@@ -682,7 +682,7 @@ size_t swiotlb_max_mapping_size(struct device *dev)

  bool is_swiotlb_active(struct device *dev)
  {
-	return dev->dma_io_tlb_mem != NULL;
+	return dev_iotlb_mem(dev) != NULL;
  }
  EXPORT_SYMBOL_GPL(is_swiotlb_active);

@@ -729,7 +729,7 @@ static void rmem_swiotlb_debugfs_init(struct 
reserved_mem *rmem)

  struct page *swiotlb_alloc(struct device *dev, size_t size)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	phys_addr_t tlb_addr;
  	int index;

@@ -792,7 +792,7 @@ static int rmem_swiotlb_device_init(struct 
reserved_mem *rmem,
  static void rmem_swiotlb_device_release(struct reserved_mem *rmem,
  					struct device *dev)
  {
-	dev->dma_io_tlb_mem = io_tlb_default_mem;
+	dev->dma_io_tlb_mem = NULL;
  }

  static const struct reserved_mem_ops rmem_swiotlb_ops = {
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu

WARNING: multiple messages have this Message-ID (diff)
From: Robin Murphy <robin.murphy@arm.com>
To: Will Deacon <will@kernel.org>, Christoph Hellwig <hch@lst.de>
Cc: Jim Quinlan <james.quinlan@broadcom.com>,
	heikki.krogerus@linux.intel.com,
	linux-devicetree <devicetree@vger.kernel.org>,
	peterz@infradead.org, dri-devel@lists.freedesktop.org,
	chris@chris-wilson.co.uk, grant.likely@arm.com, paulus@samba.org,
	Frank Rowand <frowand.list@gmail.com>,
	mingo@kernel.org, Jianxiong Gao <jxgao@google.com>,
	Stefano Stabellini <sstabellini@kernel.org>,
	Saravana Kannan <saravanak@google.com>,
	mpe@ellerman.id.au,
	"Rafael J . Wysocki" <rafael.j.wysocki@intel.com>,
	Bartosz Golaszewski <bgolaszewski@baylibre.com>,
	matthew.auld@intel.com, linux-pci@vger.kernel.org,
	xen-devel@lists.xenproject.org,
	Thierry Reding <treding@nvidia.com>,
	bskeggs@redhat.com, Nicolas Boichat <drinkcat@chromium.org>,
	thomas.hellstrom@linux.intel.com, jgross@suse.com,
	Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>,
	intel-gfx@lists.freedesktop.org,
	Nathan Chancellor <nathan@kernel.org>,
	Rob Herring <robh+dt@kernel.org>,
	rodrigo.vivi@intel.com, Bjorn Helgaas <bhelgaas@google.com>,
	Claire Chang <tientzu@chromium.org>,
	Dan Williams <dan.j.williams@intel.com>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	boris.ostrovsky@oracle.com, airlied@linux.ie,
	Greg KH <gregkh@linuxfoundation.org>,
	Randy Dunlap <rdunlap@infradead.org>,
	Qian Cai <quic_qiancai@quicinc.com>,
	lkml <linux-kernel@vger.kernel.org>,
	"list@263.net:IOMMU DRIVERS" <iommu@lists.linux-foundation.org>,
	xypron.glpk@gmx.de, Tom Lendacky <thomas.lendacky@amd.com>,
	linuxppc-dev@lists.ozlabs.org, bauerman@linux.ibm.com
Subject: Re: [PATCH v15 06/12] swiotlb: Use is_swiotlb_force_bounce for swiotlb data bouncing
Date: Tue, 6 Jul 2021 15:01:04 +0100	[thread overview]
Message-ID: <a59f771f-3289-62f0-ca50-8f3675d9b166@arm.com> (raw)
In-Reply-To: <20210706132422.GA20327@willie-the-truck>

On 2021-07-06 14:24, Will Deacon wrote:
> On Tue, Jul 06, 2021 at 06:48:48AM +0200, Christoph Hellwig wrote:
>> On Mon, Jul 05, 2021 at 08:03:52PM +0100, Will Deacon wrote:
>>> So at this point, the AMD IOMMU driver does:
>>>
>>> 	swiotlb        = (iommu_default_passthrough() || sme_me_mask) ? 1 : 0;
>>>
>>> where 'swiotlb' is a global variable indicating whether or not swiotlb
>>> is in use. It's picked up a bit later on by pci_swiotlb_late_init(), which
>>> will call swiotlb_exit() if 'swiotlb' is false.
>>>
>>> Now, that used to work fine, because swiotlb_exit() clears
>>> 'io_tlb_default_mem' to NULL, but now with the restricted DMA changes, I
>>> think that all the devices which have successfully probed beforehand will
>>> have stale pointers to the freed structure in their 'dev->dma_io_tlb_mem'
>>> field.
>>
>> Yeah.  I don't think we can do that anymore, and I also think it is
>> a bad idea to start with.
> 
> I've had a crack at reworking things along the following lines:
> 
>    - io_tlb_default_mem now lives in the BSS, the flexible array member
>      is now a pointer and that part is allocated dynamically (downside of
>      this is an extra indirection to get at the slots).
> 
>    - io_tlb_default_mem.nslabs tells you whether the thing is valid
> 
>    - swiotlb_exit() frees the slots array and clears the rest of the
>      structure to 0. I also extended it to free the actual slabs, but I'm
>      not sure why it wasn't doing that before.
> 
> So a non-NULL dev->dma_io_tlb_mem should always be valid to follow.

FWIW I was pondering the question of whether to do something along those 
lines or just scrap the default assignment entirely, so since I hadn't 
got round to saying that I've gone ahead and hacked up the alternative 
(similarly untested) for comparison :)

TBH I'm still not sure which one I prefer...

Robin.

----->8-----
diff --git a/drivers/base/core.c b/drivers/base/core.c
index ea5b85354526..394abf184c1a 100644
--- a/drivers/base/core.c
+++ b/drivers/base/core.c
@@ -2847,9 +2847,6 @@ void device_initialize(struct device *dev)
      defined(CONFIG_ARCH_HAS_SYNC_DMA_FOR_CPU_ALL)
  	dev->dma_coherent = dma_default_coherent;
  #endif
-#ifdef CONFIG_SWIOTLB
-	dev->dma_io_tlb_mem = io_tlb_default_mem;
-#endif
  }
  EXPORT_SYMBOL_GPL(device_initialize);

diff --git a/include/linux/swiotlb.h b/include/linux/swiotlb.h
index 39284ff2a6cd..620f16d89a98 100644
--- a/include/linux/swiotlb.h
+++ b/include/linux/swiotlb.h
@@ -107,16 +107,21 @@ struct io_tlb_mem {
  };
  extern struct io_tlb_mem *io_tlb_default_mem;

+static inline struct io_tlb_mem *dev_iotlb_mem(struct device *dev)
+{
+	return dev->dma_io_tlb_mem ?: io_tlb_default_mem;
+}
+
  static inline bool is_swiotlb_buffer(struct device *dev, phys_addr_t 
paddr)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);

  	return mem && paddr >= mem->start && paddr < mem->end;
  }

  static inline bool is_swiotlb_force_bounce(struct device *dev)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);

  	return mem && mem->force_bounce;
  }
@@ -167,7 +172,7 @@ bool swiotlb_free(struct device *dev, struct page 
*page, size_t size);

  static inline bool is_swiotlb_for_alloc(struct device *dev)
  {
-	return dev->dma_io_tlb_mem->for_alloc;
+	return dev_iotlb_mem(dev)->for_alloc;
  }
  #else
  static inline struct page *swiotlb_alloc(struct device *dev, size_t size)
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index b7f76bca89bf..f4942149f87d 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -359,7 +359,7 @@ static unsigned int swiotlb_align_offset(struct 
device *dev, u64 addr)
  static void swiotlb_bounce(struct device *dev, phys_addr_t tlb_addr, 
size_t size,
  			   enum dma_data_direction dir)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	int index = (tlb_addr - mem->start) >> IO_TLB_SHIFT;
  	phys_addr_t orig_addr = mem->slots[index].orig_addr;
  	size_t alloc_size = mem->slots[index].alloc_size;
@@ -440,7 +440,7 @@ static unsigned int wrap_index(struct io_tlb_mem 
*mem, unsigned int index)
  static int swiotlb_find_slots(struct device *dev, phys_addr_t orig_addr,
  			      size_t alloc_size)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned long boundary_mask = dma_get_seg_boundary(dev);
  	dma_addr_t tbl_dma_addr =
  		phys_to_dma_unencrypted(dev, mem->start) & boundary_mask;
@@ -522,7 +522,7 @@ phys_addr_t swiotlb_tbl_map_single(struct device 
*dev, phys_addr_t orig_addr,
  		size_t mapping_size, size_t alloc_size,
  		enum dma_data_direction dir, unsigned long attrs)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned int offset = swiotlb_align_offset(dev, orig_addr);
  	unsigned int i;
  	int index;
@@ -565,7 +565,7 @@ phys_addr_t swiotlb_tbl_map_single(struct device 
*dev, phys_addr_t orig_addr,

  static void swiotlb_release_slots(struct device *dev, phys_addr_t 
tlb_addr)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned long flags;
  	unsigned int offset = swiotlb_align_offset(dev, tlb_addr);
  	int index = (tlb_addr - offset - mem->start) >> IO_TLB_SHIFT;
@@ -682,7 +682,7 @@ size_t swiotlb_max_mapping_size(struct device *dev)

  bool is_swiotlb_active(struct device *dev)
  {
-	return dev->dma_io_tlb_mem != NULL;
+	return dev_iotlb_mem(dev) != NULL;
  }
  EXPORT_SYMBOL_GPL(is_swiotlb_active);

@@ -729,7 +729,7 @@ static void rmem_swiotlb_debugfs_init(struct 
reserved_mem *rmem)

  struct page *swiotlb_alloc(struct device *dev, size_t size)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	phys_addr_t tlb_addr;
  	int index;

@@ -792,7 +792,7 @@ static int rmem_swiotlb_device_init(struct 
reserved_mem *rmem,
  static void rmem_swiotlb_device_release(struct reserved_mem *rmem,
  					struct device *dev)
  {
-	dev->dma_io_tlb_mem = io_tlb_default_mem;
+	dev->dma_io_tlb_mem = NULL;
  }

  static const struct reserved_mem_ops rmem_swiotlb_ops = {

WARNING: multiple messages have this Message-ID (diff)
From: Robin Murphy <robin.murphy@arm.com>
To: Will Deacon <will@kernel.org>, Christoph Hellwig <hch@lst.de>
Cc: Jim Quinlan <james.quinlan@broadcom.com>,
	heikki.krogerus@linux.intel.com,
	linux-devicetree <devicetree@vger.kernel.org>,
	peterz@infradead.org, benh@kernel.crashing.org,
	dri-devel@lists.freedesktop.org, chris@chris-wilson.co.uk,
	grant.likely@arm.com, paulus@samba.org,
	Frank Rowand <frowand.list@gmail.com>,
	mingo@kernel.org, Jianxiong Gao <jxgao@google.com>,
	Stefano Stabellini <sstabellini@kernel.org>,
	Saravana Kannan <saravanak@google.com>,
	mpe@ellerman.id.au,
	"Rafael J . Wysocki" <rafael.j.wysocki@intel.com>,
	Bartosz Golaszewski <bgolaszewski@baylibre.com>,
	matthew.auld@intel.com, linux-pci@vger.kernel.org,
	xen-devel@lists.xenproject.org,
	Thierry Reding <treding@nvidia.com>,
	bskeggs@redhat.com, Nicolas Boichat <drinkcat@chromium.org>,
	thomas.hellstrom@linux.intel.com, jgross@suse.com,
	Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>,
	intel-gfx@lists.freedesktop.org,
	Nathan Chancellor <nathan@kernel.org>,
	Rob Herring <robh+dt@kernel.org>,
	Bjorn Helgaas <bhelgaas@google.com>,
	Claire Chang <tientzu@chromium.org>,
	Dan Williams <dan.j.williams@intel.com>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	boris.ostrovsky@oracle.com, airlied@linux.ie,
	Greg KH <gregkh@linuxfoundation.org>,
	Randy Dunlap <rdunlap@infradead.org>,
	Qian Cai <quic_qiancai@quicinc.com>,
	lkml <linux-kernel@vger.kernel.org>,
	"list@263.net:IOMMU DRIVERS" <iommu@lists.linux-foundation.org>,
	xypron.glpk@gmx.de, Tom Lendacky <thomas.lendacky@amd.com>,
	linuxppc-dev@lists.ozlabs.org, bauerman@linux.ibm.com
Subject: Re: [Intel-gfx] [PATCH v15 06/12] swiotlb: Use is_swiotlb_force_bounce for swiotlb data bouncing
Date: Tue, 6 Jul 2021 15:01:04 +0100	[thread overview]
Message-ID: <a59f771f-3289-62f0-ca50-8f3675d9b166@arm.com> (raw)
In-Reply-To: <20210706132422.GA20327@willie-the-truck>

On 2021-07-06 14:24, Will Deacon wrote:
> On Tue, Jul 06, 2021 at 06:48:48AM +0200, Christoph Hellwig wrote:
>> On Mon, Jul 05, 2021 at 08:03:52PM +0100, Will Deacon wrote:
>>> So at this point, the AMD IOMMU driver does:
>>>
>>> 	swiotlb        = (iommu_default_passthrough() || sme_me_mask) ? 1 : 0;
>>>
>>> where 'swiotlb' is a global variable indicating whether or not swiotlb
>>> is in use. It's picked up a bit later on by pci_swiotlb_late_init(), which
>>> will call swiotlb_exit() if 'swiotlb' is false.
>>>
>>> Now, that used to work fine, because swiotlb_exit() clears
>>> 'io_tlb_default_mem' to NULL, but now with the restricted DMA changes, I
>>> think that all the devices which have successfully probed beforehand will
>>> have stale pointers to the freed structure in their 'dev->dma_io_tlb_mem'
>>> field.
>>
>> Yeah.  I don't think we can do that anymore, and I also think it is
>> a bad idea to start with.
> 
> I've had a crack at reworking things along the following lines:
> 
>    - io_tlb_default_mem now lives in the BSS, the flexible array member
>      is now a pointer and that part is allocated dynamically (downside of
>      this is an extra indirection to get at the slots).
> 
>    - io_tlb_default_mem.nslabs tells you whether the thing is valid
> 
>    - swiotlb_exit() frees the slots array and clears the rest of the
>      structure to 0. I also extended it to free the actual slabs, but I'm
>      not sure why it wasn't doing that before.
> 
> So a non-NULL dev->dma_io_tlb_mem should always be valid to follow.

FWIW I was pondering the question of whether to do something along those 
lines or just scrap the default assignment entirely, so since I hadn't 
got round to saying that I've gone ahead and hacked up the alternative 
(similarly untested) for comparison :)

TBH I'm still not sure which one I prefer...

Robin.

----->8-----
diff --git a/drivers/base/core.c b/drivers/base/core.c
index ea5b85354526..394abf184c1a 100644
--- a/drivers/base/core.c
+++ b/drivers/base/core.c
@@ -2847,9 +2847,6 @@ void device_initialize(struct device *dev)
      defined(CONFIG_ARCH_HAS_SYNC_DMA_FOR_CPU_ALL)
  	dev->dma_coherent = dma_default_coherent;
  #endif
-#ifdef CONFIG_SWIOTLB
-	dev->dma_io_tlb_mem = io_tlb_default_mem;
-#endif
  }
  EXPORT_SYMBOL_GPL(device_initialize);

diff --git a/include/linux/swiotlb.h b/include/linux/swiotlb.h
index 39284ff2a6cd..620f16d89a98 100644
--- a/include/linux/swiotlb.h
+++ b/include/linux/swiotlb.h
@@ -107,16 +107,21 @@ struct io_tlb_mem {
  };
  extern struct io_tlb_mem *io_tlb_default_mem;

+static inline struct io_tlb_mem *dev_iotlb_mem(struct device *dev)
+{
+	return dev->dma_io_tlb_mem ?: io_tlb_default_mem;
+}
+
  static inline bool is_swiotlb_buffer(struct device *dev, phys_addr_t 
paddr)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);

  	return mem && paddr >= mem->start && paddr < mem->end;
  }

  static inline bool is_swiotlb_force_bounce(struct device *dev)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);

  	return mem && mem->force_bounce;
  }
@@ -167,7 +172,7 @@ bool swiotlb_free(struct device *dev, struct page 
*page, size_t size);

  static inline bool is_swiotlb_for_alloc(struct device *dev)
  {
-	return dev->dma_io_tlb_mem->for_alloc;
+	return dev_iotlb_mem(dev)->for_alloc;
  }
  #else
  static inline struct page *swiotlb_alloc(struct device *dev, size_t size)
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index b7f76bca89bf..f4942149f87d 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -359,7 +359,7 @@ static unsigned int swiotlb_align_offset(struct 
device *dev, u64 addr)
  static void swiotlb_bounce(struct device *dev, phys_addr_t tlb_addr, 
size_t size,
  			   enum dma_data_direction dir)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	int index = (tlb_addr - mem->start) >> IO_TLB_SHIFT;
  	phys_addr_t orig_addr = mem->slots[index].orig_addr;
  	size_t alloc_size = mem->slots[index].alloc_size;
@@ -440,7 +440,7 @@ static unsigned int wrap_index(struct io_tlb_mem 
*mem, unsigned int index)
  static int swiotlb_find_slots(struct device *dev, phys_addr_t orig_addr,
  			      size_t alloc_size)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned long boundary_mask = dma_get_seg_boundary(dev);
  	dma_addr_t tbl_dma_addr =
  		phys_to_dma_unencrypted(dev, mem->start) & boundary_mask;
@@ -522,7 +522,7 @@ phys_addr_t swiotlb_tbl_map_single(struct device 
*dev, phys_addr_t orig_addr,
  		size_t mapping_size, size_t alloc_size,
  		enum dma_data_direction dir, unsigned long attrs)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned int offset = swiotlb_align_offset(dev, orig_addr);
  	unsigned int i;
  	int index;
@@ -565,7 +565,7 @@ phys_addr_t swiotlb_tbl_map_single(struct device 
*dev, phys_addr_t orig_addr,

  static void swiotlb_release_slots(struct device *dev, phys_addr_t 
tlb_addr)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	unsigned long flags;
  	unsigned int offset = swiotlb_align_offset(dev, tlb_addr);
  	int index = (tlb_addr - offset - mem->start) >> IO_TLB_SHIFT;
@@ -682,7 +682,7 @@ size_t swiotlb_max_mapping_size(struct device *dev)

  bool is_swiotlb_active(struct device *dev)
  {
-	return dev->dma_io_tlb_mem != NULL;
+	return dev_iotlb_mem(dev) != NULL;
  }
  EXPORT_SYMBOL_GPL(is_swiotlb_active);

@@ -729,7 +729,7 @@ static void rmem_swiotlb_debugfs_init(struct 
reserved_mem *rmem)

  struct page *swiotlb_alloc(struct device *dev, size_t size)
  {
-	struct io_tlb_mem *mem = dev->dma_io_tlb_mem;
+	struct io_tlb_mem *mem = dev_iotlb_mem(dev);
  	phys_addr_t tlb_addr;
  	int index;

@@ -792,7 +792,7 @@ static int rmem_swiotlb_device_init(struct 
reserved_mem *rmem,
  static void rmem_swiotlb_device_release(struct reserved_mem *rmem,
  					struct device *dev)
  {
-	dev->dma_io_tlb_mem = io_tlb_default_mem;
+	dev->dma_io_tlb_mem = NULL;
  }

  static const struct reserved_mem_ops rmem_swiotlb_ops = {
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  reply	other threads:[~2021-07-06 14:40 UTC|newest]

Thread overview: 245+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-24 15:55 [PATCH v15 00/12] Restricted DMA Claire Chang
2021-06-24 15:55 ` [Intel-gfx] " Claire Chang
2021-06-24 15:55 ` Claire Chang
2021-06-24 15:55 ` Claire Chang
2021-06-24 15:55 ` Claire Chang
2021-06-24 15:55 ` [PATCH v15 01/12] swiotlb: Refactor swiotlb init functions Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55 ` [PATCH v15 02/12] swiotlb: Refactor swiotlb_create_debugfs Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55 ` [PATCH v15 03/12] swiotlb: Set dev->dma_io_tlb_mem to the swiotlb pool used Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55 ` [PATCH v15 04/12] swiotlb: Update is_swiotlb_buffer to add a struct device argument Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55 ` [PATCH v15 05/12] swiotlb: Update is_swiotlb_active " Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55 ` [PATCH v15 06/12] swiotlb: Use is_swiotlb_force_bounce for swiotlb data bouncing Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-30  1:43   ` Nathan Chancellor
2021-06-30  1:43     ` [Intel-gfx] " Nathan Chancellor
2021-06-30  1:43     ` Nathan Chancellor
2021-06-30  1:43     ` Nathan Chancellor
2021-06-30  1:43     ` Nathan Chancellor
2021-06-30  9:17     ` Claire Chang
2021-06-30  9:17       ` Claire Chang
2021-06-30  9:17       ` [Intel-gfx] " Claire Chang
2021-06-30  9:17       ` Claire Chang
2021-06-30  9:17       ` Claire Chang
2021-06-30  9:17       ` Claire Chang
2021-06-30 11:43       ` Will Deacon
2021-06-30 11:43         ` [Intel-gfx] " Will Deacon
2021-06-30 11:43         ` Will Deacon
2021-06-30 11:43         ` Will Deacon
2021-06-30 11:43         ` Will Deacon
2021-06-30 15:56         ` Nathan Chancellor
2021-06-30 15:56           ` [Intel-gfx] " Nathan Chancellor
2021-06-30 15:56           ` Nathan Chancellor
2021-06-30 15:56           ` Nathan Chancellor
2021-06-30 15:56           ` Nathan Chancellor
2021-07-01  7:40           ` Will Deacon
2021-07-01  7:40             ` [Intel-gfx] " Will Deacon
2021-07-01  7:40             ` Will Deacon
2021-07-01  7:40             ` Will Deacon
2021-07-01  7:40             ` Will Deacon
2021-07-01  7:52             ` Nathan Chancellor
2021-07-01  7:52               ` [Intel-gfx] " Nathan Chancellor
2021-07-01  7:52               ` Nathan Chancellor
2021-07-01  7:52               ` Nathan Chancellor
2021-07-01  7:52               ` Nathan Chancellor
2021-07-02 13:58               ` Will Deacon
2021-07-02 13:58                 ` [Intel-gfx] " Will Deacon
2021-07-02 13:58                 ` Will Deacon
2021-07-02 13:58                 ` Will Deacon
2021-07-02 13:58                 ` Will Deacon
2021-07-02 15:13                 ` Robin Murphy
2021-07-02 15:13                   ` [Intel-gfx] " Robin Murphy
2021-07-02 15:13                   ` Robin Murphy
2021-07-02 15:13                   ` Robin Murphy
2021-07-02 15:13                   ` Robin Murphy
2021-07-03  5:55                   ` Nathan Chancellor
2021-07-03  5:55                     ` [Intel-gfx] " Nathan Chancellor
2021-07-03  5:55                     ` Nathan Chancellor
2021-07-03  5:55                     ` Nathan Chancellor
2021-07-03  5:55                     ` Nathan Chancellor
2021-07-05  7:29                     ` Claire Chang
2021-07-05  7:29                       ` Claire Chang
2021-07-05  7:29                       ` [Intel-gfx] " Claire Chang
2021-07-05  7:29                       ` Claire Chang
2021-07-05  7:29                       ` Claire Chang
2021-07-05  7:29                       ` Claire Chang
2021-07-05 18:25                       ` Nathan Chancellor
2021-07-05 18:25                         ` [Intel-gfx] " Nathan Chancellor
2021-07-05 18:25                         ` Nathan Chancellor
2021-07-05 18:25                         ` Nathan Chancellor
2021-07-05 18:25                         ` Nathan Chancellor
2021-07-05 19:03                     ` Will Deacon
2021-07-05 19:03                       ` [Intel-gfx] " Will Deacon
2021-07-05 19:03                       ` Will Deacon
2021-07-05 19:03                       ` Will Deacon
2021-07-05 19:03                       ` Will Deacon
2021-07-06  4:48                       ` Christoph Hellwig
2021-07-06  4:48                         ` [Intel-gfx] " Christoph Hellwig
2021-07-06  4:48                         ` Christoph Hellwig
2021-07-06  4:48                         ` Christoph Hellwig
2021-07-06 13:24                         ` Will Deacon
2021-07-06 13:24                           ` [Intel-gfx] " Will Deacon
2021-07-06 13:24                           ` Will Deacon
2021-07-06 13:24                           ` Will Deacon
2021-07-06 13:24                           ` Will Deacon
2021-07-06 14:01                           ` Robin Murphy [this message]
2021-07-06 14:01                             ` [Intel-gfx] " Robin Murphy
2021-07-06 14:01                             ` Robin Murphy
2021-07-06 14:01                             ` Robin Murphy
2021-07-06 14:01                             ` Robin Murphy
2021-07-06 14:05                             ` Christoph Hellwig
2021-07-06 14:05                               ` [Intel-gfx] " Christoph Hellwig
2021-07-06 14:05                               ` Christoph Hellwig
2021-07-06 14:05                               ` Christoph Hellwig
2021-07-06 14:46                               ` Konrad Rzeszutek Wilk
2021-07-06 14:46                                 ` [Intel-gfx] " Konrad Rzeszutek Wilk
2021-07-06 14:46                                 ` Konrad Rzeszutek Wilk
2021-07-06 14:46                                 ` Konrad Rzeszutek Wilk
2021-07-06 14:46                                 ` Konrad Rzeszutek Wilk
2021-07-06 16:57                                 ` Will Deacon
2021-07-06 16:57                                   ` [Intel-gfx] " Will Deacon
2021-07-06 16:57                                   ` Will Deacon
2021-07-06 16:57                                   ` Will Deacon
2021-07-06 16:57                                   ` Will Deacon
2021-07-06 16:59                                   ` Konrad Rzeszutek Wilk
2021-07-06 16:59                                     ` [Intel-gfx] " Konrad Rzeszutek Wilk
2021-07-06 16:59                                     ` Konrad Rzeszutek Wilk
2021-07-06 16:59                                     ` Konrad Rzeszutek Wilk
2021-07-06 16:59                                     ` Konrad Rzeszutek Wilk
2021-07-12 13:56                                     ` Will Deacon
2021-07-12 13:56                                       ` [Intel-gfx] " Will Deacon
2021-07-12 13:56                                       ` Will Deacon
2021-07-12 13:56                                       ` Will Deacon
2021-07-12 13:56                                       ` Will Deacon
2021-07-14  0:06                                       ` Konrad Rzeszutek Wilk
2021-07-14  0:06                                         ` [Intel-gfx] " Konrad Rzeszutek Wilk
2021-07-14  0:06                                         ` Konrad Rzeszutek Wilk
2021-07-14  0:06                                         ` Konrad Rzeszutek Wilk
2021-07-14  0:06                                         ` Konrad Rzeszutek Wilk
2021-07-06 15:39                               ` Robin Murphy
2021-07-06 15:39                                 ` [Intel-gfx] " Robin Murphy
2021-07-06 15:39                                 ` Robin Murphy
2021-07-06 15:39                                 ` Robin Murphy
2021-07-06 15:39                                 ` Robin Murphy
2021-07-06 17:06                                 ` Will Deacon
2021-07-06 17:06                                   ` [Intel-gfx] " Will Deacon
2021-07-06 17:06                                   ` Will Deacon
2021-07-06 17:06                                   ` Will Deacon
2021-07-06 17:06                                   ` Will Deacon
2021-07-06 19:14                                   ` Nathan Chancellor
2021-07-06 19:14                                     ` [Intel-gfx] " Nathan Chancellor
2021-07-06 19:14                                     ` Nathan Chancellor
2021-07-06 19:14                                     ` Nathan Chancellor
2021-07-06 19:14                                     ` Nathan Chancellor
2021-07-08 16:44                                     ` Will Deacon
2021-07-08 16:44                                       ` [Intel-gfx] " Will Deacon
2021-07-08 16:44                                       ` Will Deacon
2021-07-08 16:44                                       ` Will Deacon
2021-07-08 16:44                                       ` Will Deacon
2021-06-24 15:55 ` [PATCH v15 07/12] swiotlb: Move alloc_size to swiotlb_find_slots Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55 ` [PATCH v15 08/12] swiotlb: Refactor swiotlb_tbl_unmap_single Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55 ` [PATCH v15 09/12] swiotlb: Add restricted DMA alloc/free support Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55 ` [PATCH v15 10/12] swiotlb: Add restricted DMA pool initialization Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-08-24 14:26   ` Guenter Roeck
2021-08-24 14:26     ` Guenter Roeck
2021-08-24 14:26     ` [Intel-gfx] " Guenter Roeck
2021-08-24 14:26     ` Guenter Roeck
2021-08-27  3:50     ` Claire Chang
2021-08-27  3:50       ` Claire Chang
2021-08-27  3:50       ` [Intel-gfx] " Claire Chang
2021-08-27  3:50       ` Claire Chang
2021-08-27  3:50       ` Claire Chang
2021-08-27  3:50       ` Claire Chang
2021-08-27  6:58   ` Andy Shevchenko
2021-08-27  6:58     ` Andy Shevchenko
2021-08-27  6:58     ` [Intel-gfx] " Andy Shevchenko
2021-08-27  6:58     ` Andy Shevchenko
2021-08-27  6:58     ` Andy Shevchenko
2021-06-24 15:55 ` [PATCH v15 11/12] dt-bindings: of: Add restricted DMA pool Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55 ` [PATCH v15 12/12] of: Add plumbing for " Claire Chang
2021-06-24 15:55   ` [Intel-gfx] " Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-06-24 15:55   ` Claire Chang
2021-07-02  3:08   ` Guenter Roeck
2021-07-02  3:08     ` [Intel-gfx] " Guenter Roeck
2021-07-02  3:08     ` Guenter Roeck
2021-07-02  3:08     ` Guenter Roeck
2021-07-02  3:08     ` Guenter Roeck
2021-07-02 11:39     ` Robin Murphy
2021-07-02 11:39       ` [Intel-gfx] " Robin Murphy
2021-07-02 11:39       ` Robin Murphy
2021-07-02 11:39       ` Robin Murphy
2021-07-02 11:39       ` Robin Murphy
2021-07-02 13:18       ` Will Deacon
2021-07-02 13:18         ` [Intel-gfx] " Will Deacon
2021-07-02 13:18         ` Will Deacon
2021-07-02 13:18         ` Will Deacon
2021-07-02 13:18         ` Will Deacon
2021-07-02 13:48         ` Guenter Roeck
2021-07-02 13:48           ` [Intel-gfx] " Guenter Roeck
2021-07-02 13:48           ` Guenter Roeck
2021-07-02 13:48           ` Guenter Roeck
2021-07-02 13:48           ` Guenter Roeck
2021-06-24 16:42 ` [Intel-gfx] ✗ Fi.CI.BUILD: failure for Restricted DMA Patchwork
2021-06-24 19:19 ` [PATCH v15 00/12] " Konrad Rzeszutek Wilk
2021-06-24 19:19   ` [Intel-gfx] " Konrad Rzeszutek Wilk
2021-06-24 19:19   ` Konrad Rzeszutek Wilk
2021-06-24 19:19   ` Konrad Rzeszutek Wilk
2021-06-24 19:19   ` Konrad Rzeszutek Wilk
2021-06-25  0:41   ` Claire Chang
2021-06-25  0:41     ` Claire Chang
2021-06-25  0:41     ` [Intel-gfx] " Claire Chang
2021-06-25  0:41     ` Claire Chang
2021-06-25  0:41     ` Claire Chang
2021-06-25  0:41     ` Claire Chang
2021-06-25 12:30   ` Will Deacon
2021-06-25 12:30     ` [Intel-gfx] " Will Deacon
2021-06-25 12:30     ` Will Deacon
2021-06-25 12:30     ` Will Deacon
2021-06-25 12:30     ` Will Deacon
2021-07-02 15:48 ` [Intel-gfx] ✗ Fi.CI.BUILD: failure for Restricted DMA (rev2) Patchwork
2021-07-05  7:57 ` [Intel-gfx] ✗ Fi.CI.BUILD: failure for Restricted DMA (rev3) Patchwork
2021-07-06 18:57 ` [Intel-gfx] ✗ Fi.CI.BUILD: failure for Restricted DMA (rev5) Patchwork

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=a59f771f-3289-62f0-ca50-8f3675d9b166@arm.com \
    --to=robin.murphy@arm.com \
    --cc=airlied@linux.ie \
    --cc=andriy.shevchenko@linux.intel.com \
    --cc=bauerman@linux.ibm.com \
    --cc=benh@kernel.crashing.org \
    --cc=bgolaszewski@baylibre.com \
    --cc=bhelgaas@google.com \
    --cc=boris.ostrovsky@oracle.com \
    --cc=bskeggs@redhat.com \
    --cc=chris@chris-wilson.co.uk \
    --cc=dan.j.williams@intel.com \
    --cc=daniel@ffwll.ch \
    --cc=devicetree@vger.kernel.org \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=drinkcat@chromium.org \
    --cc=frowand.list@gmail.com \
    --cc=grant.likely@arm.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=hch@lst.de \
    --cc=heikki.krogerus@linux.intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=iommu@lists.linux-foundation.org \
    --cc=james.quinlan@broadcom.com \
    --cc=jani.nikula@linux.intel.com \
    --cc=jgross@suse.com \
    --cc=joonas.lahtinen@linux.intel.com \
    --cc=jxgao@google.com \
    --cc=konrad.wilk@oracle.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=maarten.lankhorst@linux.intel.com \
    --cc=matthew.auld@intel.com \
    --cc=mingo@kernel.org \
    --cc=mpe@ellerman.id.au \
    --cc=nathan@kernel.org \
    --cc=paulus@samba.org \
    --cc=peterz@infradead.org \
    --cc=quic_qiancai@quicinc.com \
    --cc=rafael.j.wysocki@intel.com \
    --cc=rdunlap@infradead.org \
    --cc=robh+dt@kernel.org \
    --cc=rodrigo.vivi@intel.com \
    --cc=saravanak@google.com \
    --cc=sstabellini@kernel.org \
    --cc=thomas.hellstrom@linux.intel.com \
    --cc=thomas.lendacky@amd.com \
    --cc=tientzu@chromium.org \
    --cc=treding@nvidia.com \
    --cc=will@kernel.org \
    --cc=xen-devel@lists.xenproject.org \
    --cc=xypron.glpk@gmx.de \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.