Re: [PATCH v7 01/12] vpci: introduce per-domain lock to protect vpci structure

[Volodymyr Babchuk <Volodymyr_Babchuk@xxxxxxxx>]

Hi Roger,

Roger Pau Monné <roger.pau@xxxxxxxxxx> writes:

> On Wed, Jun 21, 2023 at 10:07:20PM +0000, Volodymyr Babchuk wrote:
>> 
>> Hi Roger,
>> 
>> Roger Pau Monné <roger.pau@xxxxxxxxxx> writes:
>> 
>> > On Tue, Jun 13, 2023 at 10:32:26AM +0000, Volodymyr Babchuk wrote:
>> >> From: Oleksandr Andrushchenko <oleksandr_andrushchenko@xxxxxxxx>
>> >> 
>> >> Introduce a per-domain read/write lock to check whether vpci is present,
>> >> so we are sure there are no accesses to the contents of the vpci struct
>> >> if not. This lock can be used (and in a few cases is used right away)
>> >> so that vpci removal can be performed while holding the lock in write
>> >> mode. Previously such removal could race with vpci_read for example.
>> >> 
>> >> 1. Per-domain's vpci_rwlock is used to protect pdev->vpci structure
>> >> from being removed.
>> >> 
>> >> 2. Writing the command register and ROM BAR register may trigger
>> >> modify_bars to run, which in turn may access multiple pdevs while
>> >> checking for the existing BAR's overlap. The overlapping check, if done
>> >> under the read lock, requires vpci->lock to be acquired on both devices
>> >> being compared, which may produce a deadlock. It is not possible to
>> >> upgrade read lock to write lock in such a case. So, in order to prevent
>> >> the deadlock, check which registers are going to be written and acquire
>> >> the lock in the appropriate mode from the beginning.
>> >> 
>> >> All other code, which doesn't lead to pdev->vpci destruction and does not
>> >> access multiple pdevs at the same time, can still use a combination of the
>> >> read lock and pdev->vpci->lock.
>> >> 
>> >> 3. Optimize if ROM BAR write lock required detection by caching offset
>> >> of the ROM BAR register in vpci->header->rom_reg which depends on
>> >> header's type.
>> >> 
>> >> 4. Reduce locked region in vpci_remove_device as it is now possible
>> >> to set pdev->vpci to NULL early right after the write lock is acquired.
>> >> 
>> >> 5. Reduce locked region in vpci_add_handlers as it is possible to
>> >> initialize many more fields of the struct vpci before assigning it to
>> >> pdev->vpci.
>> >> 
>> >> 6. vpci_{add|remove}_register are required to be called with the write 
>> >> lock
>> >> held, but it is not feasible to add an assert there as it requires
>> >> struct domain to be passed for that. So, add a comment about this 
>> >> requirement
>> >> to these and other functions with the equivalent constraints.
>> >> 
>> >> 7. Drop const qualifier where the new rwlock is used and this is 
>> >> appropriate.
>> >> 
>> >> 8. Do not call process_pending_softirqs with any locks held. For that 
>> >> unlock
>> >> prior the call and re-acquire the locks after. After re-acquiring the
>> >> lock there is no need to check if pdev->vpci exists:
>> >>  - in apply_map because of the context it is called (no race condition
>> >>    possible)
>> >>  - for MSI/MSI-X debug code because it is called at the end of
>> >>    pdev->vpci access and no further access to pdev->vpci is made
>> >> 
>> >> 9. Check for !pdev->vpci in vpci_{read|write} after acquiring the lock
>> >> and if so, allow reading or writing the hardware register directly. This 
>> >> is
>> >> acceptable as we only deal with Dom0 as of now. Once DomU support is
>> >> added the write will need to be ignored and read return all 0's for the
>> >> guests, while Dom0 can still access the registers directly.
>> >> 
>> >> 10. Introduce pcidevs_trylock, so there is a possibility to try locking
>> >> the pcidev's lock.
>> >> 
>> >> 11. Use pcidev's lock around for_each_pdev and pci_get_pdev_by_domain
>> >> while accessing pdevs in vpci code.
>> >> 
>> >> 12. This is based on the discussion at [1].
>> >> 
>> >> [1] 
>> >> https://urldefense.com/v3/__https://lore.kernel.org/all/20220204063459.680961-4-andr2000@xxxxxxxxx/__;!!GF_29dbcQIUBPA!zPy31CWFWlyC0xhEHiSj6rOPe7RDSjLranI9KZqhG4ssmChJMWvsPLJPQGTcVsnnowZpP8-LaKJkIWIzb8ue0DoYhg$
>> >>  [lore[.]kernel[.]org]
>> >> 
>> >> Suggested-by: Roger Pau Monné <roger.pau@xxxxxxxxxx>
>> >> Suggested-by: Jan Beulich <jbeulich@xxxxxxxx>
>> >> Signed-off-by: Oleksandr Andrushchenko <oleksandr_andrushchenko@xxxxxxxx>
>> >
>> > Thanks.
>> >
>> > I haven't looked in full detail, but I'm afraid there's an ABBA
>> > deadlock with the per-domain vpci lock and the pcidevs lock in
>> > modify_bars() vs  vpci_add_handlers() and vpci_remove_device().
>> >
>> > I've made some comments below.
>> 
>> Thank you for the review. I believe that it is a good idea to have a
>> per-domain pdev_list lock. See my answers below.
>
> I think it's important that the lock that protects domain->pdev_list
> must be the same that also protects pdev->vpci, or else you might run
> into similar ABBA deadlock situations.
>
> The problem then could be that in vpci_{read,write} you will take the
> per-domain pdev lock in read mode in order to get the pdev, and for
> writes to the command register or the ROM BAR you would have to
> upgrade such lock to a write lock without dropping it, and we don't
> have such functionality for rw locks ATM.
>
> Maybe just re-starting the function knowing that the lock must be
> taken in write mode would be a good solution: writes to the command
> register will already be slow since they are likely to involve
> modifications to the p2m.

Looks like modify_bars() is the only cause for this extended lock. I
know that this was discussed earlier, but can we rework modify_bars to
not iterate over all the pdevs? We can store copy of all enabled BARs in
a domain structure, protected by domain->vpci_lock. Something akin to

struct vpci_bar {
        list_head list;
        struct vpci *vpci;
        unsigned long start;
        unsigned long end;
        bool is_rom;
};


>> >> @@ -213,6 +227,7 @@ static void defer_map(struct domain *d, struct 
>> >> pci_dev *pdev,
>> >>      raise_softirq(SCHEDULE_SOFTIRQ);
>> >>  }
>> >>  
>> >> +/* This must hold domain's vpci_rwlock in write mode. */
>> >
>> > Why it must be in write mode?
>> >
>> > Is this to avoid ABBA deadlocks as not taking the per-domain lock in
>> > write mode would then force us to take each pdev->vpci->lock in order
>> > to prevent concurrent modifications of remote devices?
>> 
>> Yes, exactly this. This is mentioned in the commit message:
>> 
>>     2. Writing the command register and ROM BAR register may trigger
>>     modify_bars to run, which in turn may access multiple pdevs while
>>     checking for the existing BAR's overlap. The overlapping check, if done
>>     under the read lock, requires vpci->lock to be acquired on both devices
>>     being compared, which may produce a deadlock. It is not possible to
>>     upgrade read lock to write lock in such a case. So, in order to prevent
>>     the deadlock, check which registers are going to be written and acquire
>>     the lock in the appropriate mode from the beginning.
>
> Might be good to put part of the commit message in the code comment,
> just saying that the lock must be taken in write mode is not very
> informative.
>
> /*
>  * The <lock-name> per domain lock must be taken in write mode in
>  * order to prevent changes to the vPCI data of devices assigned to
>  * the domain, since the function parses such data.
>  */
>

Agree. Also, I'll add a corresponding ASSERT()

-- 
WBR, Volodymyr