linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: David Daney <ddaney.cavm@gmail.com>
To: linux-kernel@vger.kernel.org, Bjorn Helgaas <bhelgaas@google.com>,
	linux-pci@vger.kernel.org
Cc: David Daney <david.daney@cavium.com>
Subject: [PATCH] PCI: aer_inject: Don't call config read/write functions with invalid ops structure.
Date: Tue, 22 Dec 2015 13:44:51 -0800	[thread overview]
Message-ID: <1450820691-27588-1-git-send-email-ddaney.cavm@gmail.com> (raw)

From: David Daney <david.daney@cavium.com>

The aer_inject module, when active, replaces the pci_ops structure of
the bus so that it can intercept config space accesses.  If the config
space access needs to be forwarded to the original pci_ops, we must
restore the bus's original ops, as the read()/write() functions may
require use of the data in the pci_ops structure.  An example of such
functions are the pci_generic_config*() functions in
drivers/pci/access.c.  A failure to call these with a valid bus->ops
results in dereferencing an invalid bus->ops->map_bus pointer and
subsequent OOPs and system crash.

So, do as suggested above.  When calling to the original
read()/write() functions, temporarily switch back to the original
pci_ops.  This is safe to do as the pci_lock must already be held, and
thus no other users of the bus->ops are possible.

Signed-off-by: David Daney <david.daney@cavium.com>
---
 drivers/pci/pcie/aer/aer_inject.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)

diff --git a/drivers/pci/pcie/aer/aer_inject.c b/drivers/pci/pcie/aer/aer_inject.c
index 182224a..2d60e0e 100644
--- a/drivers/pci/pcie/aer/aer_inject.c
+++ b/drivers/pci/pcie/aer/aer_inject.c
@@ -188,7 +188,9 @@ static int pci_read_aer(struct pci_bus *bus, unsigned int devfn, int where,
 	struct aer_error *err;
 	unsigned long flags;
 	struct pci_ops *ops;
+	struct pci_ops *my_ops;
 	int domain;
+	int rv;
 
 	spin_lock_irqsave(&inject_lock, flags);
 	if (size != sizeof(u32))
@@ -208,8 +210,19 @@ static int pci_read_aer(struct pci_bus *bus, unsigned int devfn, int where,
 	}
 out:
 	ops = __find_pci_bus_ops(bus);
+	/*
+	 * pci_lock must already be held, so we can directly
+	 * manipulate bus->ops.  Many config access functions,
+	 * including pci_generic_config_read() require the original
+	 * bus->ops be installed to function, so temporarily put them
+	 * back.
+	 */
+	my_ops = bus->ops;
+	bus->ops = ops;
+	rv = ops->read(bus, devfn, where, size, val);
+	bus->ops = my_ops;
 	spin_unlock_irqrestore(&inject_lock, flags);
-	return ops->read(bus, devfn, where, size, val);
+	return rv;
 }
 
 static int pci_write_aer(struct pci_bus *bus, unsigned int devfn, int where,
@@ -220,7 +233,9 @@ static int pci_write_aer(struct pci_bus *bus, unsigned int devfn, int where,
 	unsigned long flags;
 	int rw1cs;
 	struct pci_ops *ops;
+	struct pci_ops *my_ops;
 	int domain;
+	int rv;
 
 	spin_lock_irqsave(&inject_lock, flags);
 	if (size != sizeof(u32))
@@ -243,8 +258,19 @@ static int pci_write_aer(struct pci_bus *bus, unsigned int devfn, int where,
 	}
 out:
 	ops = __find_pci_bus_ops(bus);
+	/*
+	 * pci_lock must already be held, so we can directly
+	 * manipulate bus->ops.  Many config access functions,
+	 * including pci_generic_config_write() require the original
+	 * bus->ops be installed to function, so temporarily put them
+	 * back.
+	 */
+	my_ops = bus->ops;
+	bus->ops = ops;
+	rv = ops->write(bus, devfn, where, size, val);
+	bus->ops = my_ops;
 	spin_unlock_irqrestore(&inject_lock, flags);
-	return ops->write(bus, devfn, where, size, val);
+	return rv;
 }
 
 static struct pci_ops pci_ops_aer = {
-- 
1.8.3.1


             reply	other threads:[~2015-12-22 21:45 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-12-22 21:44 David Daney [this message]
2016-02-04 20:22 ` [PATCH] PCI: aer_inject: Don't call config read/write functions with invalid ops structure Bjorn Helgaas

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=1450820691-27588-1-git-send-email-ddaney.cavm@gmail.com \
    --to=ddaney.cavm@gmail.com \
    --cc=bhelgaas@google.com \
    --cc=david.daney@cavium.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).