[PATCH v5 11/15] PCI: Obey iomem restrictions for procfs mmap

[PATCH v5 11/15] PCI: Obey iomem restrictions for procfs mmap

[Daniel Vetter]

There's three ways to access PCI BARs from userspace: /dev/mem, sysfs
files, and the old proc interface. Two check against
iomem_is_exclusive, proc never did. And with CONFIG_IO_STRICT_DEVMEM,
this starts to matter, since we don't want random userspace having
access to PCI BARs while a driver is loaded and using it.

Fix this by adding the same iomem_is_exclusive() check we already have
on the sysfs side in pci_mmap_resource().

References: 90a545e98126 ("restrict /dev/mem to idle io memory ranges")
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Kees Cook <keescook@chromium.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: linux-mm@kvack.org
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-samsung-soc@vger.kernel.org
Cc: linux-media@vger.kernel.org
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
--
v2: Improve commit message (Bjorn)
---
 drivers/pci/proc.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/pci/proc.c b/drivers/pci/proc.c
index d35186b01d98..3a2f90beb4cb 100644
--- a/drivers/pci/proc.c
+++ b/drivers/pci/proc.c
@@ -274,6 +274,11 @@ static int proc_bus_pci_mmap(struct file *file, struct vm_area_struct *vma)
 		else
 			return -EINVAL;
 	}
+
+	if (dev->resource[i].flags & IORESOURCE_MEM &&
+	    iomem_is_exclusive(dev->resource[i].start))
+		return -EINVAL;
+
 	ret = pci_mmap_page_range(dev, i, vma,
 				  fpriv->mmap_state, write_combine);
 	if (ret < 0)
-- 
2.28.0

Re: [PATCH v5 11/15] PCI: Obey iomem restrictions for procfs mmap

[Bjorn Helgaas]

On Fri, Oct 30, 2020 at 11:08:11AM +0100, Daniel Vetter wrote:
> There's three ways to access PCI BARs from userspace: /dev/mem, sysfs
> files, and the old proc interface. Two check against
> iomem_is_exclusive, proc never did. And with CONFIG_IO_STRICT_DEVMEM,
> this starts to matter, since we don't want random userspace having
> access to PCI BARs while a driver is loaded and using it.
> 
> Fix this by adding the same iomem_is_exclusive() check we already have
> on the sysfs side in pci_mmap_resource().
> 
> References: 90a545e98126 ("restrict /dev/mem to idle io memory ranges")
> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

This is OK with me but it looks like IORESOURCE_EXCLUSIVE is currently
only used in a few places:

  e1000_probe() calls pci_request_selected_regions_exclusive(),
  ne_pci_probe() calls pci_request_regions_exclusive(),
  vmbus_allocate_mmio() calls request_mem_region_exclusive()

which raises the question of whether it's worth keeping
IORESOURCE_EXCLUSIVE at all.  I'm totally fine with removing it
completely.

But if you want it,

Acked-by: Bjorn Helgaas <bhelgaas@google.com>

> Cc: Jason Gunthorpe <jgg@ziepe.ca>
> Cc: Kees Cook <keescook@chromium.org>
> Cc: Dan Williams <dan.j.williams@intel.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: John Hubbard <jhubbard@nvidia.com>
> Cc: Jérôme Glisse <jglisse@redhat.com>
> Cc: Jan Kara <jack@suse.cz>
> Cc: Dan Williams <dan.j.williams@intel.com>
> Cc: linux-mm@kvack.org
> Cc: linux-arm-kernel@lists.infradead.org
> Cc: linux-samsung-soc@vger.kernel.org
> Cc: linux-media@vger.kernel.org
> Cc: Bjorn Helgaas <bhelgaas@google.com>
> Cc: linux-pci@vger.kernel.org
> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> --
> v2: Improve commit message (Bjorn)
> ---
>  drivers/pci/proc.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/pci/proc.c b/drivers/pci/proc.c
> index d35186b01d98..3a2f90beb4cb 100644
> --- a/drivers/pci/proc.c
> +++ b/drivers/pci/proc.c
> @@ -274,6 +274,11 @@ static int proc_bus_pci_mmap(struct file *file, struct vm_area_struct *vma)
>  		else
>  			return -EINVAL;
>  	}
> +
> +	if (dev->resource[i].flags & IORESOURCE_MEM &&
> +	    iomem_is_exclusive(dev->resource[i].start))
> +		return -EINVAL;
> +
>  	ret = pci_mmap_page_range(dev, i, vma,
>  				  fpriv->mmap_state, write_combine);
>  	if (ret < 0)
> -- 
> 2.28.0
> 

Re: [PATCH v5 11/15] PCI: Obey iomem restrictions for procfs mmap

[Dan Williams]

On Tue, Nov 3, 2020 at 1:28 PM Bjorn Helgaas <helgaas@kernel.org> wrote:
>
> On Fri, Oct 30, 2020 at 11:08:11AM +0100, Daniel Vetter wrote:
> > There's three ways to access PCI BARs from userspace: /dev/mem, sysfs
> > files, and the old proc interface. Two check against
> > iomem_is_exclusive, proc never did. And with CONFIG_IO_STRICT_DEVMEM,
> > this starts to matter, since we don't want random userspace having
> > access to PCI BARs while a driver is loaded and using it.
> >
> > Fix this by adding the same iomem_is_exclusive() check we already have
> > on the sysfs side in pci_mmap_resource().
> >
> > References: 90a545e98126 ("restrict /dev/mem to idle io memory ranges")
> > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
>
> This is OK with me but it looks like IORESOURCE_EXCLUSIVE is currently
> only used in a few places:
>
>   e1000_probe() calls pci_request_selected_regions_exclusive(),
>   ne_pci_probe() calls pci_request_regions_exclusive(),
>   vmbus_allocate_mmio() calls request_mem_region_exclusive()
>
> which raises the question of whether it's worth keeping
> IORESOURCE_EXCLUSIVE at all.  I'm totally fine with removing it
> completely.

Now that CONFIG_IO_STRICT_DEVMEM upgrades IORESOURCE_BUSY to
IORESOURCE_EXCLUSIVE semantics the latter has lost its meaning so I'd
be in favor of removing it as well.

Re: [PATCH v5 11/15] PCI: Obey iomem restrictions for procfs mmap

[Daniel Vetter]

On Tue, Nov 3, 2020 at 11:09 PM Dan Williams <dan.j.williams@intel.com> wrote:
> On Tue, Nov 3, 2020 at 1:28 PM Bjorn Helgaas <helgaas@kernel.org> wrote:
> > On Fri, Oct 30, 2020 at 11:08:11AM +0100, Daniel Vetter wrote:
> > > There's three ways to access PCI BARs from userspace: /dev/mem, sysfs
> > > files, and the old proc interface. Two check against
> > > iomem_is_exclusive, proc never did. And with CONFIG_IO_STRICT_DEVMEM,
> > > this starts to matter, since we don't want random userspace having
> > > access to PCI BARs while a driver is loaded and using it.
> > >
> > > Fix this by adding the same iomem_is_exclusive() check we already have
> > > on the sysfs side in pci_mmap_resource().
> > >
> > > References: 90a545e98126 ("restrict /dev/mem to idle io memory ranges")
> > > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
> >
> > This is OK with me but it looks like IORESOURCE_EXCLUSIVE is currently
> > only used in a few places:
> >
> >   e1000_probe() calls pci_request_selected_regions_exclusive(),
> >   ne_pci_probe() calls pci_request_regions_exclusive(),
> >   vmbus_allocate_mmio() calls request_mem_region_exclusive()
> >
> > which raises the question of whether it's worth keeping
> > IORESOURCE_EXCLUSIVE at all.  I'm totally fine with removing it
> > completely.
>
> Now that CONFIG_IO_STRICT_DEVMEM upgrades IORESOURCE_BUSY to
> IORESOURCE_EXCLUSIVE semantics the latter has lost its meaning so I'd
> be in favor of removing it as well.

Still has some value since it enforces exclusive access even if the
config isn't enabled, and iirc e1000 had some fun with userspace tools
clobbering the firmware and bricking the chip.

Another thing I kinda wondered, since pci maintainer is here: At least
in drivers/gpu I see very few drivers explicitly requestion regions
(this might be a historical artifact due to the shadow attach stuff
before we had real modesetting drivers). And pci core doesn't do that
either, even when a driver is bound. Is this intentional, or
should/could we do better? Since drivers work happily without
reserving regions I don't think "the drivers need to remember to do
this" will ever really work out well.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

Re: [PATCH v5 11/15] PCI: Obey iomem restrictions for procfs mmap

[Bjorn Helgaas]

On Wed, Nov 04, 2020 at 09:44:04AM +0100, Daniel Vetter wrote:
> On Tue, Nov 3, 2020 at 11:09 PM Dan Williams <dan.j.williams@intel.com> wrote:
> > On Tue, Nov 3, 2020 at 1:28 PM Bjorn Helgaas <helgaas@kernel.org> wrote:
> > > On Fri, Oct 30, 2020 at 11:08:11AM +0100, Daniel Vetter wrote:
> > > > There's three ways to access PCI BARs from userspace: /dev/mem, sysfs
> > > > files, and the old proc interface. Two check against
> > > > iomem_is_exclusive, proc never did. And with CONFIG_IO_STRICT_DEVMEM,
> > > > this starts to matter, since we don't want random userspace having
> > > > access to PCI BARs while a driver is loaded and using it.
> > > >
> > > > Fix this by adding the same iomem_is_exclusive() check we already have
> > > > on the sysfs side in pci_mmap_resource().
> > > >
> > > > References: 90a545e98126 ("restrict /dev/mem to idle io memory ranges")
> > > > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
> > >
> > > This is OK with me but it looks like IORESOURCE_EXCLUSIVE is currently
> > > only used in a few places:
> > >
> > >   e1000_probe() calls pci_request_selected_regions_exclusive(),
> > >   ne_pci_probe() calls pci_request_regions_exclusive(),
> > >   vmbus_allocate_mmio() calls request_mem_region_exclusive()
> > >
> > > which raises the question of whether it's worth keeping
> > > IORESOURCE_EXCLUSIVE at all.  I'm totally fine with removing it
> > > completely.
> >
> > Now that CONFIG_IO_STRICT_DEVMEM upgrades IORESOURCE_BUSY to
> > IORESOURCE_EXCLUSIVE semantics the latter has lost its meaning so I'd
> > be in favor of removing it as well.
> 
> Still has some value since it enforces exclusive access even if the
> config isn't enabled, and iirc e1000 had some fun with userspace tools
> clobbering the firmware and bricking the chip.

There's *some* value; I'm just skeptical since only three drivers use
it.

IORESOURCE_EXCLUSIVE is from e8de1481fd71 ("resource: allow MMIO
exclusivity for device drivers"), and the commit message says this is
only active when CONFIG_STRICT_DEVMEM is set.  I didn't check to see
whether that's still true.

That commit adds a bunch of wrappers and "__"-prefixed functions to
pass the IORESOURCE_EXCLUSIVE flag around.  That's a fair bit of
uglification for three drivers.

> Another thing I kinda wondered, since pci maintainer is here: At least
> in drivers/gpu I see very few drivers explicitly requestion regions
> (this might be a historical artifact due to the shadow attach stuff
> before we had real modesetting drivers). And pci core doesn't do that
> either, even when a driver is bound. Is this intentional, or
> should/could we do better? Since drivers work happily without
> reserving regions I don't think "the drivers need to remember to do
> this" will ever really work out well.

You're right, many drivers don't call pci_request_regions().  Maybe we
could do better, but I haven't looked into that recently.  There is a
related note in Documentation/PCI/pci.rst that's been there for a long
time (it refers to "pci_request_resources()", which has never existed
AFAICT).  I'm certainly open to proposals.

Re: [PATCH v5 11/15] PCI: Obey iomem restrictions for procfs mmap

[Dan Williams]

On Wed, Nov 4, 2020 at 8:50 AM Bjorn Helgaas <helgaas@kernel.org> wrote:
>
> On Wed, Nov 04, 2020 at 09:44:04AM +0100, Daniel Vetter wrote:
> > On Tue, Nov 3, 2020 at 11:09 PM Dan Williams <dan.j.williams@intel.com> wrote:
> > > On Tue, Nov 3, 2020 at 1:28 PM Bjorn Helgaas <helgaas@kernel.org> wrote:
> > > > On Fri, Oct 30, 2020 at 11:08:11AM +0100, Daniel Vetter wrote:
> > > > > There's three ways to access PCI BARs from userspace: /dev/mem, sysfs
> > > > > files, and the old proc interface. Two check against
> > > > > iomem_is_exclusive, proc never did. And with CONFIG_IO_STRICT_DEVMEM,
> > > > > this starts to matter, since we don't want random userspace having
> > > > > access to PCI BARs while a driver is loaded and using it.
> > > > >
> > > > > Fix this by adding the same iomem_is_exclusive() check we already have
> > > > > on the sysfs side in pci_mmap_resource().
> > > > >
> > > > > References: 90a545e98126 ("restrict /dev/mem to idle io memory ranges")
> > > > > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
> > > >
> > > > This is OK with me but it looks like IORESOURCE_EXCLUSIVE is currently
> > > > only used in a few places:
> > > >
> > > >   e1000_probe() calls pci_request_selected_regions_exclusive(),
> > > >   ne_pci_probe() calls pci_request_regions_exclusive(),
> > > >   vmbus_allocate_mmio() calls request_mem_region_exclusive()
> > > >
> > > > which raises the question of whether it's worth keeping
> > > > IORESOURCE_EXCLUSIVE at all.  I'm totally fine with removing it
> > > > completely.
> > >
> > > Now that CONFIG_IO_STRICT_DEVMEM upgrades IORESOURCE_BUSY to
> > > IORESOURCE_EXCLUSIVE semantics the latter has lost its meaning so I'd
> > > be in favor of removing it as well.
> >
> > Still has some value since it enforces exclusive access even if the
> > config isn't enabled, and iirc e1000 had some fun with userspace tools
> > clobbering the firmware and bricking the chip.
>
> There's *some* value; I'm just skeptical since only three drivers use
> it.
>
> IORESOURCE_EXCLUSIVE is from e8de1481fd71 ("resource: allow MMIO
> exclusivity for device drivers"), and the commit message says this is
> only active when CONFIG_STRICT_DEVMEM is set.  I didn't check to see
> whether that's still true.
>
> That commit adds a bunch of wrappers and "__"-prefixed functions to
> pass the IORESOURCE_EXCLUSIVE flag around.  That's a fair bit of
> uglification for three drivers.
>
> > Another thing I kinda wondered, since pci maintainer is here: At least
> > in drivers/gpu I see very few drivers explicitly requestion regions
> > (this might be a historical artifact due to the shadow attach stuff
> > before we had real modesetting drivers). And pci core doesn't do that
> > either, even when a driver is bound. Is this intentional, or
> > should/could we do better? Since drivers work happily without
> > reserving regions I don't think "the drivers need to remember to do
> > this" will ever really work out well.
>
> You're right, many drivers don't call pci_request_regions().  Maybe we
> could do better, but I haven't looked into that recently.  There is a
> related note in Documentation/PCI/pci.rst that's been there for a long
> time (it refers to "pci_request_resources()", which has never existed
> AFAICT).  I'm certainly open to proposals.

It seems a bug that the kernel permits MMIO regions with side effects
to be ioremap()'ed without request_mem_region() on the resource. I
wonder how much log spam would happen if ioremap() reported whenever a
non-IORESOURE_BUSY range was passed to it? The current state of
affairs to trust *remap users to have claimed their remap target seems
too ingrained to unwind now.

Re: [PATCH v5 11/15] PCI: Obey iomem restrictions for procfs mmap

[Daniel Vetter]

On Wed, Nov 04, 2020 at 12:12:15PM -0800, Dan Williams wrote:
> On Wed, Nov 4, 2020 at 8:50 AM Bjorn Helgaas <helgaas@kernel.org> wrote:
> >
> > On Wed, Nov 04, 2020 at 09:44:04AM +0100, Daniel Vetter wrote:
> > > On Tue, Nov 3, 2020 at 11:09 PM Dan Williams <dan.j.williams@intel.com> wrote:
> > > > On Tue, Nov 3, 2020 at 1:28 PM Bjorn Helgaas <helgaas@kernel.org> wrote:
> > > > > On Fri, Oct 30, 2020 at 11:08:11AM +0100, Daniel Vetter wrote:
> > > > > > There's three ways to access PCI BARs from userspace: /dev/mem, sysfs
> > > > > > files, and the old proc interface. Two check against
> > > > > > iomem_is_exclusive, proc never did. And with CONFIG_IO_STRICT_DEVMEM,
> > > > > > this starts to matter, since we don't want random userspace having
> > > > > > access to PCI BARs while a driver is loaded and using it.
> > > > > >
> > > > > > Fix this by adding the same iomem_is_exclusive() check we already have
> > > > > > on the sysfs side in pci_mmap_resource().
> > > > > >
> > > > > > References: 90a545e98126 ("restrict /dev/mem to idle io memory ranges")
> > > > > > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
> > > > >
> > > > > This is OK with me but it looks like IORESOURCE_EXCLUSIVE is currently
> > > > > only used in a few places:
> > > > >
> > > > >   e1000_probe() calls pci_request_selected_regions_exclusive(),
> > > > >   ne_pci_probe() calls pci_request_regions_exclusive(),
> > > > >   vmbus_allocate_mmio() calls request_mem_region_exclusive()
> > > > >
> > > > > which raises the question of whether it's worth keeping
> > > > > IORESOURCE_EXCLUSIVE at all.  I'm totally fine with removing it
> > > > > completely.
> > > >
> > > > Now that CONFIG_IO_STRICT_DEVMEM upgrades IORESOURCE_BUSY to
> > > > IORESOURCE_EXCLUSIVE semantics the latter has lost its meaning so I'd
> > > > be in favor of removing it as well.
> > >
> > > Still has some value since it enforces exclusive access even if the
> > > config isn't enabled, and iirc e1000 had some fun with userspace tools
> > > clobbering the firmware and bricking the chip.
> >
> > There's *some* value; I'm just skeptical since only three drivers use
> > it.
> >
> > IORESOURCE_EXCLUSIVE is from e8de1481fd71 ("resource: allow MMIO
> > exclusivity for device drivers"), and the commit message says this is
> > only active when CONFIG_STRICT_DEVMEM is set.  I didn't check to see
> > whether that's still true.
> >
> > That commit adds a bunch of wrappers and "__"-prefixed functions to
> > pass the IORESOURCE_EXCLUSIVE flag around.  That's a fair bit of
> > uglification for three drivers.
> >
> > > Another thing I kinda wondered, since pci maintainer is here: At least
> > > in drivers/gpu I see very few drivers explicitly requestion regions
> > > (this might be a historical artifact due to the shadow attach stuff
> > > before we had real modesetting drivers). And pci core doesn't do that
> > > either, even when a driver is bound. Is this intentional, or
> > > should/could we do better? Since drivers work happily without
> > > reserving regions I don't think "the drivers need to remember to do
> > > this" will ever really work out well.
> >
> > You're right, many drivers don't call pci_request_regions().  Maybe we
> > could do better, but I haven't looked into that recently.  There is a
> > related note in Documentation/PCI/pci.rst that's been there for a long
> > time (it refers to "pci_request_resources()", which has never existed
> > AFAICT).  I'm certainly open to proposals.
> 
> It seems a bug that the kernel permits MMIO regions with side effects
> to be ioremap()'ed without request_mem_region() on the resource. I
> wonder how much log spam would happen if ioremap() reported whenever a
> non-IORESOURE_BUSY range was passed to it? The current state of
> affairs to trust *remap users to have claimed their remap target seems
> too ingrained to unwind now.

Yeah I think that's hopeless. I think the only feasible approach is if bus
drivers claim resources by default when a driver is bound (it should nest,
so if the driver claims again, I think that should all keep working), just
using the driver name. Probably with some special casing for legacy io
(vgaarb.c should claim these I guess). Still probably tons of fallout.

Once that's rolled out to all bus drivers we could perhaps add the ioremap
check without drowning in log spam. Still a multi-year project I think :-/
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

[PATCH v5 14/15] sysfs: Support zapping of binary attr mmaps

[Daniel Vetter]

We want to be able to revoke pci mmaps so that the same access rules
applies as for /dev/kmem. Revoke support for devmem was added in
3234ac664a87 ("/dev/mem: Revoke mappings when a driver claims the
region").

The simplest way to achieve this is by having the same filp->f_mapping
for all mappings, so that unmap_mapping_range can find them all, no
matter through which file they've been created. Since this must be set
at open time we need sysfs support for this.

Add an optional mapping parameter bin_attr, which is only consulted
when there's also an mmap callback, since without mmap support
allowing to adjust the ->f_mapping makes no sense.

Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Kees Cook <keescook@chromium.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: linux-mm@kvack.org
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-samsung-soc@vger.kernel.org
Cc: linux-media@vger.kernel.org
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linux-pci@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Sourabh Jain <sourabhjain@linux.ibm.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Cc: Nayna Jain <nayna@linux.ibm.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 fs/sysfs/file.c       | 11 +++++++++++
 include/linux/sysfs.h |  2 ++
 2 files changed, 13 insertions(+)

diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c
index 96d0da65e088..9aefa7779b29 100644
--- a/fs/sysfs/file.c
+++ b/fs/sysfs/file.c
@@ -170,6 +170,16 @@ static int sysfs_kf_bin_mmap(struct kernfs_open_file *of,
 	return battr->mmap(of->file, kobj, battr, vma);
 }
 
+static int sysfs_kf_bin_open(struct kernfs_open_file *of)
+{
+	struct bin_attribute *battr = of->kn->priv;
+
+	if (battr->mapping)
+		of->file->f_mapping = battr->mapping;
+
+	return 0;
+}
+
 void sysfs_notify(struct kobject *kobj, const char *dir, const char *attr)
 {
 	struct kernfs_node *kn = kobj->sd, *tmp;
@@ -241,6 +251,7 @@ static const struct kernfs_ops sysfs_bin_kfops_mmap = {
 	.read		= sysfs_kf_bin_read,
 	.write		= sysfs_kf_bin_write,
 	.mmap		= sysfs_kf_bin_mmap,
+	.open		= sysfs_kf_bin_open,
 };
 
 int sysfs_add_file_mode_ns(struct kernfs_node *parent,
diff --git a/include/linux/sysfs.h b/include/linux/sysfs.h
index 2caa34c1ca1a..d76a1ddf83a3 100644
--- a/include/linux/sysfs.h
+++ b/include/linux/sysfs.h
@@ -164,11 +164,13 @@ __ATTRIBUTE_GROUPS(_name)
 
 struct file;
 struct vm_area_struct;
+struct address_space;
 
 struct bin_attribute {
 	struct attribute	attr;
 	size_t			size;
 	void			*private;
+	struct address_space	*mapping;
 	ssize_t (*read)(struct file *, struct kobject *, struct bin_attribute *,
 			char *, loff_t, size_t);
 	ssize_t (*write)(struct file *, struct kobject *, struct bin_attribute *,
-- 
2.28.0

[PATCH v5 15/15] PCI: Revoke mappings like devmem

[Daniel Vetter]

Since 3234ac664a87 ("/dev/mem: Revoke mappings when a driver claims
the region") /dev/kmem zaps ptes when the kernel requests exclusive
acccess to an iomem region. And with CONFIG_IO_STRICT_DEVMEM, this is
the default for all driver uses.

Except there's two more ways to access PCI BARs: sysfs and proc mmap
support. Let's plug that hole.

For revoke_devmem() to work we need to link our vma into the same
address_space, with consistent vma->vm_pgoff. ->pgoff is already
adjusted, because that's how (io_)remap_pfn_range works, but for the
mapping we need to adjust vma->vm_file->f_mapping. The cleanest way is
to adjust this at at ->open time:

- for sysfs this is easy, now that binary attributes support this. We
  just set bin_attr->mapping when mmap is supported
- for procfs it's a bit more tricky, since procfs pci access has only
  one file per device, and access to a specific resources first needs
  to be set up with some ioctl calls. But mmap is only supported for
  the same resources as sysfs exposes with mmap support, and otherwise
  rejected, so we can set the mapping unconditionally at open time
  without harm.

A special consideration is for arch_can_pci_mmap_io() - we need to
make sure that the ->f_mapping doesn't alias between ioport and iomem
space. There's only 2 ways in-tree to support mmap of ioports: generic
pci mmap (ARCH_GENERIC_PCI_MMAP_RESOURCE), and sparc as the single
architecture hand-rolling. Both approach support ioport mmap through a
special pfn range and not through magic pte attributes. Aliasing is
therefore not a problem.

The only difference in access checks left is that sysfs PCI mmap does
not check for CAP_RAWIO. I'm not really sure whether that should be
added or not.

Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Kees Cook <keescook@chromium.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linux-mm@kvack.org
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-samsung-soc@vger.kernel.org
Cc: linux-media@vger.kernel.org
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
--
v2:
- Totally new approach: Adjust filp->f_mapping at open time. Note that
  this now works on all architectures, not just those support
  ARCH_GENERIC_PCI_MMAP_RESOURCE
---
 drivers/pci/pci-sysfs.c | 4 ++++
 drivers/pci/proc.c      | 1 +
 2 files changed, 5 insertions(+)

diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c
index d15c881e2e7e..3f1c31bc0b7c 100644
--- a/drivers/pci/pci-sysfs.c
+++ b/drivers/pci/pci-sysfs.c
@@ -929,6 +929,7 @@ void pci_create_legacy_files(struct pci_bus *b)
 	b->legacy_io->read = pci_read_legacy_io;
 	b->legacy_io->write = pci_write_legacy_io;
 	b->legacy_io->mmap = pci_mmap_legacy_io;
+	b->legacy_io->mapping = iomem_get_mapping();
 	pci_adjust_legacy_attr(b, pci_mmap_io);
 	error = device_create_bin_file(&b->dev, b->legacy_io);
 	if (error)
@@ -941,6 +942,7 @@ void pci_create_legacy_files(struct pci_bus *b)
 	b->legacy_mem->size = 1024*1024;
 	b->legacy_mem->attr.mode = 0600;
 	b->legacy_mem->mmap = pci_mmap_legacy_mem;
+	b->legacy_io->mapping = iomem_get_mapping();
 	pci_adjust_legacy_attr(b, pci_mmap_mem);
 	error = device_create_bin_file(&b->dev, b->legacy_mem);
 	if (error)
@@ -1156,6 +1158,8 @@ static int pci_create_attr(struct pci_dev *pdev, int num, int write_combine)
 			res_attr->mmap = pci_mmap_resource_uc;
 		}
 	}
+	if (res_attr->mmap)
+		res_attr->mapping = iomem_get_mapping();
 	res_attr->attr.name = res_attr_name;
 	res_attr->attr.mode = 0600;
 	res_attr->size = pci_resource_len(pdev, num);
diff --git a/drivers/pci/proc.c b/drivers/pci/proc.c
index 3a2f90beb4cb..9bab07302bbf 100644
--- a/drivers/pci/proc.c
+++ b/drivers/pci/proc.c
@@ -298,6 +298,7 @@ static int proc_bus_pci_open(struct inode *inode, struct file *file)
 	fpriv->write_combine = 0;
 
 	file->private_data = fpriv;
+	file->f_mapping = iomem_get_mapping();
 
 	return 0;
 }
-- 
2.28.0

Re: [PATCH v5 15/15] PCI: Revoke mappings like devmem

[Dan Williams]

On Fri, Oct 30, 2020 at 3:09 AM Daniel Vetter <daniel.vetter@ffwll.ch> wrote:
>
> Since 3234ac664a87 ("/dev/mem: Revoke mappings when a driver claims
> the region") /dev/kmem zaps ptes when the kernel requests exclusive
> acccess to an iomem region. And with CONFIG_IO_STRICT_DEVMEM, this is
> the default for all driver uses.
>
> Except there's two more ways to access PCI BARs: sysfs and proc mmap
> support. Let's plug that hole.
>
> For revoke_devmem() to work we need to link our vma into the same
> address_space, with consistent vma->vm_pgoff. ->pgoff is already
> adjusted, because that's how (io_)remap_pfn_range works, but for the
> mapping we need to adjust vma->vm_file->f_mapping. The cleanest way is
> to adjust this at at ->open time:
>
> - for sysfs this is easy, now that binary attributes support this. We
>   just set bin_attr->mapping when mmap is supported
> - for procfs it's a bit more tricky, since procfs pci access has only
>   one file per device, and access to a specific resources first needs
>   to be set up with some ioctl calls. But mmap is only supported for
>   the same resources as sysfs exposes with mmap support, and otherwise
>   rejected, so we can set the mapping unconditionally at open time
>   without harm.
>
> A special consideration is for arch_can_pci_mmap_io() - we need to
> make sure that the ->f_mapping doesn't alias between ioport and iomem
> space. There's only 2 ways in-tree to support mmap of ioports: generic
> pci mmap (ARCH_GENERIC_PCI_MMAP_RESOURCE), and sparc as the single
> architecture hand-rolling. Both approach support ioport mmap through a
> special pfn range and not through magic pte attributes. Aliasing is
> therefore not a problem.
>
> The only difference in access checks left is that sysfs PCI mmap does
> not check for CAP_RAWIO. I'm not really sure whether that should be
> added or not.
>
> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
> Cc: Jason Gunthorpe <jgg@ziepe.ca>
> Cc: Kees Cook <keescook@chromium.org>
> Cc: Dan Williams <dan.j.williams@intel.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: John Hubbard <jhubbard@nvidia.com>
> Cc: Jérôme Glisse <jglisse@redhat.com>
> Cc: Jan Kara <jack@suse.cz>
> Cc: Dan Williams <dan.j.williams@intel.com>

Looks good to me:

Reviewed-by: Dan Williams <dan.j.williams@intel.com>

Re: [PATCH v5 15/15] PCI: Revoke mappings like devmem

[Bjorn Helgaas]

On Fri, Oct 30, 2020 at 11:08:15AM +0100, Daniel Vetter wrote:
> Since 3234ac664a87 ("/dev/mem: Revoke mappings when a driver claims
> the region") /dev/kmem zaps ptes when the kernel requests exclusive
> acccess to an iomem region. And with CONFIG_IO_STRICT_DEVMEM, this is
> the default for all driver uses.
> 
> Except there's two more ways to access PCI BARs: sysfs and proc mmap
> support. Let's plug that hole.
> 
> For revoke_devmem() to work we need to link our vma into the same
> address_space, with consistent vma->vm_pgoff. ->pgoff is already
> adjusted, because that's how (io_)remap_pfn_range works, but for the
> mapping we need to adjust vma->vm_file->f_mapping. The cleanest way is
> to adjust this at at ->open time:
> 
> - for sysfs this is easy, now that binary attributes support this. We
>   just set bin_attr->mapping when mmap is supported
> - for procfs it's a bit more tricky, since procfs pci access has only
>   one file per device, and access to a specific resources first needs
>   to be set up with some ioctl calls. But mmap is only supported for
>   the same resources as sysfs exposes with mmap support, and otherwise
>   rejected, so we can set the mapping unconditionally at open time
>   without harm.
> 
> A special consideration is for arch_can_pci_mmap_io() - we need to
> make sure that the ->f_mapping doesn't alias between ioport and iomem
> space. There's only 2 ways in-tree to support mmap of ioports: generic
> pci mmap (ARCH_GENERIC_PCI_MMAP_RESOURCE), and sparc as the single
> architecture hand-rolling. Both approach support ioport mmap through a
> special pfn range and not through magic pte attributes. Aliasing is
> therefore not a problem.
> 
> The only difference in access checks left is that sysfs PCI mmap does
> not check for CAP_RAWIO. I'm not really sure whether that should be
> added or not.
> 
> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
> Cc: Jason Gunthorpe <jgg@ziepe.ca>
> Cc: Kees Cook <keescook@chromium.org>
> Cc: Dan Williams <dan.j.williams@intel.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: John Hubbard <jhubbard@nvidia.com>
> Cc: Jérôme Glisse <jglisse@redhat.com>
> Cc: Jan Kara <jack@suse.cz>
> Cc: Dan Williams <dan.j.williams@intel.com>
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Cc: linux-mm@kvack.org
> Cc: linux-arm-kernel@lists.infradead.org
> Cc: linux-samsung-soc@vger.kernel.org
> Cc: linux-media@vger.kernel.org
> Cc: Bjorn Helgaas <bhelgaas@google.com>
> Cc: linux-pci@vger.kernel.org
> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Acked-by: Bjorn Helgaas <bhelgaas@google.com>

> --
> v2:
> - Totally new approach: Adjust filp->f_mapping at open time. Note that
>   this now works on all architectures, not just those support
>   ARCH_GENERIC_PCI_MMAP_RESOURCE
> ---
>  drivers/pci/pci-sysfs.c | 4 ++++
>  drivers/pci/proc.c      | 1 +
>  2 files changed, 5 insertions(+)
> 
> diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c
> index d15c881e2e7e..3f1c31bc0b7c 100644
> --- a/drivers/pci/pci-sysfs.c
> +++ b/drivers/pci/pci-sysfs.c
> @@ -929,6 +929,7 @@ void pci_create_legacy_files(struct pci_bus *b)
>  	b->legacy_io->read = pci_read_legacy_io;
>  	b->legacy_io->write = pci_write_legacy_io;
>  	b->legacy_io->mmap = pci_mmap_legacy_io;
> +	b->legacy_io->mapping = iomem_get_mapping();
>  	pci_adjust_legacy_attr(b, pci_mmap_io);
>  	error = device_create_bin_file(&b->dev, b->legacy_io);
>  	if (error)
> @@ -941,6 +942,7 @@ void pci_create_legacy_files(struct pci_bus *b)
>  	b->legacy_mem->size = 1024*1024;
>  	b->legacy_mem->attr.mode = 0600;
>  	b->legacy_mem->mmap = pci_mmap_legacy_mem;
> +	b->legacy_io->mapping = iomem_get_mapping();
>  	pci_adjust_legacy_attr(b, pci_mmap_mem);
>  	error = device_create_bin_file(&b->dev, b->legacy_mem);
>  	if (error)
> @@ -1156,6 +1158,8 @@ static int pci_create_attr(struct pci_dev *pdev, int num, int write_combine)
>  			res_attr->mmap = pci_mmap_resource_uc;
>  		}
>  	}
> +	if (res_attr->mmap)
> +		res_attr->mapping = iomem_get_mapping();
>  	res_attr->attr.name = res_attr_name;
>  	res_attr->attr.mode = 0600;
>  	res_attr->size = pci_resource_len(pdev, num);
> diff --git a/drivers/pci/proc.c b/drivers/pci/proc.c
> index 3a2f90beb4cb..9bab07302bbf 100644
> --- a/drivers/pci/proc.c
> +++ b/drivers/pci/proc.c
> @@ -298,6 +298,7 @@ static int proc_bus_pci_open(struct inode *inode, struct file *file)
>  	fpriv->write_combine = 0;
>  
>  	file->private_data = fpriv;
> +	file->f_mapping = iomem_get_mapping();
>  
>  	return 0;
>  }
> -- 
> 2.28.0
>