RE: [PATCH v2] Fix failure cleanup in EvtchnFifoAcquire

["Durrant, Paul" <pdurrant@xxxxxxxxxxxx>]

> -----Original Message-----
> From: Tu Dinh <ngoc-tu.dinh@xxxxxxxxxx>
> Sent: Friday, July 18, 2025 11:54 AM
> To: Durrant, Paul <pdurrant@xxxxxxxxxxxx>; win-pv-
> devel@xxxxxxxxxxxxxxxxxxxx
> Cc: Owen Smith <owen.smith@xxxxxxxxx>
> Subject: RE: [EXTERNAL] [PATCH v2] Fix failure cleanup in
> EvtchnFifoAcquire
> 
> CAUTION: This email originated from outside of the organization. Do not
> click links or open attachments unless you can confirm the sender and know
> the content is safe.
> 
> 
> 
> On 18/07/2025 12:16, Durrant, Paul wrote:
> >> -----Original Message-----
> >> From: win-pv-devel <win-pv-devel-bounces@xxxxxxxxxxxxxxxxxxxx> On
> Behalf
> >> Of Tu Dinh
> >> Sent: Friday, July 18, 2025 10:59 AM
> >> To: win-pv-devel@xxxxxxxxxxxxxxxxxxxx
> >> Cc: Tu Dinh <ngoc-tu.dinh@xxxxxxxxxx>; Owen Smith
> <owen.smith@xxxxxxxxx>;
> >> Durrant, Paul <pdurrant@xxxxxxxxxxxx>
> >> Subject: [EXTERNAL] [PATCH v2] Fix failure cleanup in EvtchnFifoAcquire
> >>
> >> CAUTION: This email originated from outside of the organization. Do not
> >> click links or open attachments unless you can confirm the sender and
> know
> >> the content is safe.
> >>
> >>
> >>
> >> The current index is incremented before checking for failure:
> >>
> >>      while (Index < ProcessorCount) {
> >>          Index++;
> >>          [...]
> >>          if (!NT_SUCCESS(status))
> >>              goto fail1;
> >>          Context->ControlBlockMdl[vcpu_id] = Mdl;
> >>      }
> >>
> >> Make the main loop a normal for loop to avoid calling __FreePage on
> >> invalid PMDLs.
> >>
> >> Signed-off-by: Tu Dinh <ngoc-tu.dinh@xxxxxxxxxx>
> >> ---
> >>   src/xenbus/evtchn_fifo.c | 5 +----
> >>   1 file changed, 1 insertion(+), 4 deletions(-)
> >>
> >> diff --git a/src/xenbus/evtchn_fifo.c b/src/xenbus/evtchn_fifo.c
> >> index 1137dea..fd06ef4 100644
> >> --- a/src/xenbus/evtchn_fifo.c
> >> +++ b/src/xenbus/evtchn_fifo.c
> >> @@ -502,16 +502,13 @@ EvtchnFifoAcquire(
> >>
> >>       ProcessorCount =
> >> KeQueryMaximumProcessorCountEx(ALL_PROCESSOR_GROUPS);
> >>
> >> -    Index = 0;
> >> -    while (Index < ProcessorCount) {
> >> +    for (Index = 0; Index < ProcessorCount; Index++) {
> >
> > Prevailing style uses while loops I think so stick with that for
> consistency, just move the increment down.
> 
> I don't see how the increment could be moved down without breaking the
> index skip when SystemProcessorVcpuId() == STATUS_NOT_SUPPORTED.
> 

That's true, it would need a separate increment in that case.

> I also find the while loop usage in EvtchnFifoAcquire very confusing,
> and it was the cause of the cleanup bug in the first place. So I'd
> prefer to make it more idiomatic this way.
> 

More idiomatic but less consistent. Anyway...

Reviewed-by: Paul Durrant <paul@xxxxxxx>

> >
> >>           unsigned int        vcpu_id;
> >>           PFN_NUMBER          Pfn;
> >>           PHYSICAL_ADDRESS    Address;
> >>
> >>           status = SystemProcessorVcpuId(Index, &vcpu_id);
> >>
> >> -        Index++;
> >> -
> >>           if (status == STATUS_NOT_SUPPORTED)
> >>               continue;
> >>
> >> --
> >> 2.50.1.windows.1
> >>
> >>
> >>
> >> Ngoc Tu Dinh | Vates XCP-ng Developer
> >>
> >> XCP-ng & Xen Orchestra - Vates solutions
> >>
> >> web: https://vates.tech
> >>
> >
> 
> 
> 
> Ngoc Tu Dinh | Vates XCP-ng Developer
> 
> XCP-ng & Xen Orchestra - Vates solutions
> 
> web: https://vates.tech