Re: [PATCH v2] MdeModulePkg/NonDiscoverablePciDeviceDxe: Allow partial FreeBuffer


Ard Biesheuvel
 

On Fri, 5 Aug 2022 at 18:56, Jeff Brasen via groups.io
<jbrasen@...> wrote:



-----Original Message-----
From: Ard Biesheuvel <ardb@...>
Sent: Tuesday, August 2, 2022 10:51 AM
To: Jeff Brasen <jbrasen@...>
Cc: devel@edk2.groups.io; hao.a.wu@...; ray.ni@...;
quic_llindhol@...; ardb+tianocore@...
Subject: Re: [PATCH v2] MdeModulePkg/NonDiscoverablePciDeviceDxe:
Allow partial FreeBuffer

External email: Use caution opening links or attachments


On Tue, 2 Aug 2022 at 17:32, Jeff Brasen <jbrasen@...> wrote:



-----Original Message-----
From: Ard Biesheuvel <ardb@...>
Sent: Friday, July 29, 2022 9:48 AM
To: Jeff Brasen <jbrasen@...>
Cc: devel@edk2.groups.io; hao.a.wu@...; ray.ni@...;
quic_llindhol@...; ardb+tianocore@...
Subject: Re: [PATCH v2] MdeModulePkg/NonDiscoverablePciDeviceDxe:
Allow partial FreeBuffer

External email: Use caution opening links or attachments


On Thu, 28 Jul 2022 at 13:25, Jeff Brasen <jbrasen@...> wrote:


Adding Leif/Ard to CC incase they have any comments on this patch.
This generally looks ok to me. I just wonder if it wouldn't be
simpler to reuse the existing allocation descriptor if it is not
being freed entirely. Given the [presumably] the most common case is
to allocate and then free some pages at the end, lowering the page
count on the existing descriptor would cover most cases, and we'd
only need to allocate new ones if pages are being freed at the start or in
the middle.

There is often freeing at the beginning as well as this is being used to create
a 64K aligned section of memory in the case. So it over allocates and the
free's some at the beginning and the end. I could probably make it detect
and use that but figured this code would support all cases and required less
case specific detection.
Ah interesting. Would it help if the allocate routine aligned allocations to their
size?
The PciIo->AllocateBuffer function doesn't support passing the request in so we would need to know that info beforehand. The current calling in the XHCI driver does a free at the beginning and then the end of the buffer so we could the existing allocation tracker but figured it would be better to correct the function just in case someone called it to free in the middle.
What I was wondering is whether such allocations are themselves
multiples of 64k. This is perhaps orthogonal to the issue this patch
addresses, as we'' still need to deal with partial free calls
regardless. But I was curious whether XHCI in particular, and perhaps
more generally, we could streamline this by aligning all allocations
to a log2 upper bound of their sizes.

Join devel@edk2.groups.io to automatically receive all group messages.