linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Dan Carpenter <dan.carpenter@oracle.com>
To: benh@kernel.crashing.org
Cc: linux-pci@vger.kernel.org
Subject: [bug report] PCI: Protect pci_reassign_bridge_resources() against addition/removal
Date: Wed, 4 Sep 2019 13:03:03 +0300	[thread overview]
Message-ID: <20190904100303.GD7007@mwanda> (raw)

Hello Benjamin Herrenschmidt,

The patch 540f62d26f02: "PCI: Protect pci_reassign_bridge_resources()
against addition/removal" from Jun 24, 2019, leads to the following
static checker warning:

	drivers/pci/setup-bus.c:2158 pci_reassign_bridge_resources()
	warn: inconsistent returns 'read_sem:&pci_bus_sem'.

drivers/pci/setup-bus.c
  2066          unsigned int i;
  2067          int ret;
  2068  
  2069          down_read(&pci_bus_sem);

We added some new locking here.

  2070  
  2071          /* Walk to the root hub, releasing bridge BARs when possible */
  2072          next = bridge;
  2073          do {
  2074                  bridge = next;
  2075                  for (i = PCI_BRIDGE_RESOURCES; i < PCI_BRIDGE_RESOURCE_END;
  2076                       i++) {
  2077                          struct resource *res = &bridge->resource[i];
  2078  
  2079                          if ((res->flags ^ type) & PCI_RES_TYPE_MASK)
  2080                                  continue;
  2081  
  2082                          /* Ignore BARs which are still in use */
  2083                          if (res->child)
  2084                                  continue;
  2085  
  2086                          ret = add_to_list(&saved, bridge, res, 0, 0);
  2087                          if (ret)
  2088                                  goto cleanup;
  2089  
  2090                          pci_info(bridge, "BAR %d: releasing %pR\n",
  2091                                   i, res);
  2092  
  2093                          if (res->parent)
  2094                                  release_resource(res);
  2095                          res->start = 0;
  2096                          res->end = 0;
  2097                          break;
  2098                  }
  2099                  if (i == PCI_BRIDGE_RESOURCE_END)
  2100                          break;
  2101  
  2102                  next = bridge->bus ? bridge->bus->self : NULL;
  2103          } while (next);
  2104  
  2105          if (list_empty(&saved))
  2106                  return -ENOENT;

This needs an unlock.  It's not clear to me if any other clean up is
required, but possibly it's worth looking at.

  2107  
  2108          __pci_bus_size_bridges(bridge->subordinate, &added);
  2109          __pci_bridge_assign_resources(bridge, &added, &failed);

regards,
dan carpenter

             reply	other threads:[~2019-09-04 10:05 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-09-04 10:03 Dan Carpenter [this message]
2019-09-09  8:13 ` [bug report] PCI: Protect pci_reassign_bridge_resources() against addition/removal Benjamin Herrenschmidt
2019-09-23 21:50   ` 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=20190904100303.GD7007@mwanda \
    --to=dan.carpenter@oracle.com \
    --cc=benh@kernel.crashing.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).