linux-kernel - Re: [PATCH v3] PCI: Lock the pci_cfg_wait queue for the consistency of data

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20200624232303.GA2594945@bjorn-Precision-5520>
Date:   Wed, 24 Jun 2020 18:23:03 -0500
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     Xiang Zheng <zhengxiang9@...wei.com>
Cc:     bhelgaas@...gle.com, willy@...radead.org,
        wangxiongfeng2@...wei.com, wanghaibin.wang@...wei.com,
        guoheyi@...wei.com, yebiaoxiang@...wei.com,
        linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org,
        rjw@...ysocki.net, tglx@...utronix.de, guohanjun@...wei.com,
        yangyingliang@...wei.com, Stephane Eranian <eranian@...gle.com>
Subject: Re: [PATCH v3] PCI: Lock the pci_cfg_wait queue for the consistency
 of data

[+cc Stephane]

On Tue, Dec 10, 2019 at 11:15:27AM +0800, Xiang Zheng wrote:
> 7ea7e98fd8d0 ("PCI: Block on access to temporarily unavailable pci
> device") suggests that the "pci_lock" is sufficient, and all the
> callers of pci_wait_cfg() are wrapped with the "pci_lock".
> 
> However, since the commit cdcb33f98244 ("PCI: Avoid possible deadlock on
> pci_lock and p->pi_lock") merged, the accesses to the pci_cfg_wait queue
> are not safe anymore. This would cause kernel panic in a very low chance
> (See more detailed information from the below link). A "pci_lock" is
> insufficient and we need to hold an additional queue lock while read/write
> the wait queue.

<tangent>

  I'm not proud of cdcb33f98244 ("PCI: Avoid possible deadlock on
  pci_lock and p->pi_lock").

  It seems like an ad hoc solution to a problem that shouldn't exist.
  I think what it fixes is reading performance counters from PCI
  config space during a context switch when we're holding the
  task_struct pi_lock.  That doesn't seem like a path that should
  acquire pci_lock.

  I think I should have instead tried to make a lockless PCI config
  accessor that returns failure whenever we aren't allowed to read
  config space, e.g., during the recovery time after a reset or power
  state transition.  We currently *do* use pci_cfg_access_lock() to
  prevent user accesses via /proc or /sys during some of those times,
  but there's nothing that prevents kernel accesses.

  I think we're a little vulnerable there if we read those PCI
  performance counters right after changing the device power state.
  Hopefully it's nothing worse than getting ~0 data back.

</tangent>

> So let's use the add_wait_queue()/remove_wait_queue() instead of
> __add_wait_queue()/__remove_wait_queue(). Also move the wait queue
> functionality around the "schedule()" function to avoid reintroducing
> the deadlock addressed by "cdcb33f98244".
> 
> Signed-off-by: Xiang Zheng <zhengxiang9@...wei.com>
> Cc: Heyi Guo <guoheyi@...wei.com>
> Cc: Biaoxiang Ye <yebiaoxiang@...wei.com>
> Link: https://lore.kernel.org/linux-pci/79827f2f-9b43-4411-1376-b9063b67aee3@huawei.com/
> ---
> 
> v3:
>   Improve the commit subject and message.
> 
> v2:
>   Move the wait queue functionality around the "schedule()".
> 
> ---
>  drivers/pci/access.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/pci/access.c b/drivers/pci/access.c
> index 2fccb5762c76..09342a74e5ea 100644
> --- a/drivers/pci/access.c
> +++ b/drivers/pci/access.c
> @@ -207,14 +207,14 @@ static noinline void pci_wait_cfg(struct pci_dev *dev)
>  {
>  	DECLARE_WAITQUEUE(wait, current);
>  
> -	__add_wait_queue(&pci_cfg_wait, &wait);
>  	do {
>  		set_current_state(TASK_UNINTERRUPTIBLE);
>  		raw_spin_unlock_irq(&pci_lock);
> +		add_wait_queue(&pci_cfg_wait, &wait);
>  		schedule();
> +		remove_wait_queue(&pci_cfg_wait, &wait);
>  		raw_spin_lock_irq(&pci_lock);
>  	} while (dev->block_cfg_access);
> -	__remove_wait_queue(&pci_cfg_wait, &wait);
>  }
>  
>  /* Returns 0 on success, negative values indicate error. */
> -- 
> 2.19.1
> 
>