Re: [PATCH v3] PCI: hv: Warn when wait_for_response() waits indefinitely

[Sahil Chandna]

On 14-09-2026 04:09, Wei Liu wrote:
> On Fri, Sep 11, 2026 at 12:29:05PM +0200, Manivannan Sadhasivam wrote:
>> On Wed, Sep 09, 2026 at 03:01:22PM +0530, Naman Jain wrote:
>>>
>>>
>>> On 9/8/2026 5:00 PM, Sahil Chandna wrote:
>>>> A guest can wait indefinitely in wait_for_response() for the host to
>>>> send either a rescind message or a packet completion. If the
>>>> host does not send either, the guest can remain blocked with no
>>>> diagnostic indicating a reason.
>>>> This was observed during a guest kernel upgrade in which the
>>>> host-side application handling the PCI channel faulted, causing the
>>>> guest to never receive the completion request.
>>>> Add a warning in wait_for_response() when the wait exceeds
>>>> a timeout so that such a hang is visible in the guest's kernel log
>>>> and can be correlated with host-side state.
>>>>
>>>> Suggested-by: Michael Kelley <mhklinux@xxxxxxxxxxx>
>>>> Signed-off-by: Sahil Chandna <sahilchandna@xxxxxxxxxxxxxxxxxxx>
>>>> ---
>>>> Changes since v2:
>>>> - Add counter based timeout instead of introducing another timer
>>>> Link to v2: https://lore.kernel.org/all/20260902115854.2629164-1-sahilchandna@xxxxxxxxxxxxxxxxxxx/
>>>>
>>>> Changes since v1:
>>>> - Removed periodic warning to one time warning in 2 minutes
>>>> - Include vmbus relid and stuck PCI msg.
>>>> Link to v1: https://lore.kernel.org/all/20260825051850.2438816-1-sahilchandna@xxxxxxxxxxxxxxxxxxx/
>>>>   drivers/pci/controller/pci-hyperv.c | 46 ++++++++++++++++++++++-------
>>>>   1 file changed, 36 insertions(+), 10 deletions(-)
>>>>
>>>> diff --git a/drivers/pci/controller/pci-hyperv.c b/drivers/pci/controller/pci-hyperv.c
>>>> index 89816a2bd7cd..bd07402c47ed 100644
>>>> --- a/drivers/pci/controller/pci-hyperv.c
>>>> +++ b/drivers/pci/controller/pci-hyperv.c
>>>> @@ -1040,19 +1040,38 @@ static void put_pcichild(struct hv_pci_dev *hpdev)
>>>>
>>>>   /*
>>>>    * There is no good way to get notified from vmbus_onoffer_rescind(),
>>>> - * so let's use polling here, since this is not a hot path.
>>>> + * so let's use polling here, since this is not a hot path. If
>>>> + * wait_for_response() has been polling for 2 minutes
>>>> + * without either a rescind or completion, add a warning.
>>>>    */
>>>> +#define PCI_RESPONSE_HANG_TICKS 1200
>>>> +
>>>
>>> Nit, I think this variable name could be misleading, as this represents a
>>> loop counter for a retry logic, but not exactly *ticks* from kernel
>>> terminology.
>>>
>>> I would have preferred something like: PCI_RESPONSE_WARN_POLL_COUNT.
Ack, I have added in v4.
>>>
>>
>> PCI_RESPONSE_POLL_COUNT?
> 
Hi Mani, i am inclining towards PCI_RESPONSE_WARN_POLL_COUNT as
intention of change is to emit a warning when counter has exceeded a set
threshold.

> Sahil, let me know if you will send out another version. If it is only
> changing one macro name, I can do that.
> 
Hi Wei,
I have sent out v4.

Regards,
Sahil
> Wei