On 3/28/2024 6:53 PM, Chang S. Bae wrote:
>
> Then, the following is a summary of changes per patch since v8 [6]:
>
> PATCH7-8:
> * Invoke the setup code via arch_initcall() due to upstream changes
> delaying the FPU setup.
>
> PATCH9-11:
> * Add new patches for security and hotplug support clarification
I've recently made updates to a few patches, primarily related to the
mitigation parts. While the series is still under review, Eric's VAES
patches have been merged into the crypto tree and are currently being
sorted out. Once things settle down, I will make a few adjustments on the
crypto side. Then, another revision will be necessary thereafter.
Thanks,
Chang
Hi,
On Sun, Apr 07, 2024 at 04:24:18PM -0700, Chang S. Bae wrote:
> On 3/28/2024 6:53 PM, Chang S. Bae wrote:
> >
> > Then, the following is a summary of changes per patch since v8 [6]:
> >
> > PATCH7-8:
> > * Invoke the setup code via arch_initcall() due to upstream changes
> > delaying the FPU setup.
> >
> > PATCH9-11:
> > * Add new patches for security and hotplug support clarification
>
> I've recently made updates to a few patches, primarily related to the
> mitigation parts. While the series is still under review, Eric's VAES
> patches have been merged into the crypto tree and are currently being
> sorted out. Once things settle down, I will make a few adjustments on the
> crypto side. Then, another revision will be necessary thereafter.
>
> Thanks,
> Chang
Thanks for the updated patchset!
Do you have a plan for how this will be merged? Which trees will the patches go
through? I think that the actual AES-XTS implementation could still use a bit
more polishing; see my comments below. However, patches 1-12 don't need to wait
for that. Perhaps the x86 maintainers would like to take patches 1-12 for
v6.10? Then the AES-XTS support can go through the crypto tree afterwards.
As you noticed, this cycle I've been optimizing AES-XTS for x86_64 by adding new
VAES and AES-NI + AVX implementations. I have some ideas for the Key Locker
based implementation of AES-XTS:
First, surely it's the case that in practice, all CPUs that support Key Locker
also support AVX? If so, then there's no need for the Key Locker assembly to
use legacy SSE instructions. It should instead target AVX and use VEX-coded
instructions. This would save some instructions and improve performance.
Since the Key Locker assembly only supports 64-bit mode, it should also feel
free to use registers xmm8-xmm15 for purposes such as caching the XTS tweaks.
This would improve performance.
Since the Key Locker assembly advances a large number of XTS tweaks at a time
(8), I'm also wondering if it would be faster to multiply by x^8 directly
instead of multiplying by x sequentially eight times. This can be done using
the pclmulqdq instruction; see aes-xts-avx-x86_64.S which implements this
optimization. Probably all CPUs that support Key Locker also support PCLMULQDQ.
I'm also trying to think of the best way to organize the Key Locker AES-XTS glue
code. I see that you're proposing to share the glue code with the existing
AES-XTS implementations. Unfortunately I don't think this ends up working very
well, due to the facts that the Key Locker code can return errors and uses a
different key type. I think that for now, I'd prefer that you simply copied the
XTS glue code into aeskl-intel_glue.c and modified it as needed. (But make sure
to use the new version of the glue code, which is faster.)
For falling back to AES-NI, I think the cleanest solution is to call the
top-level setkey, encrypt, and decrypt functions (the ones that are set in the
xts-aes-aesni skcipher_alg), instead of calling lower-level functions as your
current patchset does.
If you could keep the Key Locker assembly roughly stylistically consistent with
the new aes-xts-avx-x86_64.S, that would be great too.
Do you happen to know if there's any way to test the Key Locker AES-XTS code
without having access to a bare metal machine with a CPU that supports Key
Locker? I tried a Sapphire Rapids based VM in Google Compute Engine, but it
doesn't enumerate Key Locker. I also don't see anything in QEMU related to Key
Locker. So I don't currently have an easy way to test this patchset.
Finally, a high level question. Key Locker has been reported to be
substantially slower than AES-NI. At the same time, VAES has recently doubled
performance over AES-NI. I'd guess this leaves Key Locker even further behind.
Given that, how useful is this patchset? I'm a bit concerned that this might be
something that sounds good in theory but won't be used in practice. Are
performance improvements for Key Locker on the horizon? (Well, there are the
improvements I suggested above, which should help, but it sounds like main issue
is the Key Locker instructions themselves which are just fundamentally slower.)
- Eric
Hi Eric,
Sorry for my delay. I just found your message in my cluttered junk
folder today after returning from my week off.
On 4/7/2024 6:48 PM, Eric Biggers wrote:
>
> Do you have a plan for how this will be merged? Which trees will the patches go
> through? I think that the actual AES-XTS implementation could still use a bit
> more polishing; see my comments below. However, patches 1-12 don't need to wait
> for that. Perhaps the x86 maintainers would like to take patches 1-12 for
> v6.10? Then the AES-XTS support can go through the crypto tree afterwards.
Yeah, this series spans both x86 code and the crypto driver. I believe
the decision should be made by maintainers. But, I suspect they'll want
to see a well-established use case code before moving forward.
> As you noticed, this cycle I've been optimizing AES-XTS for x86_64 by adding new
> VAES and AES-NI + AVX implementations. I have some ideas for the Key Locker
> based implementation of AES-XTS:
Thanks for the effort! I agree that the code could be beneficial for
daily disk encryption needs.
> First, surely it's the case that in practice, all CPUs that support Key Locker
> also support AVX? If so, then there's no need for the Key Locker assembly to
> use legacy SSE instructions. It should instead target AVX and use VEX-coded
> instructions. This would save some instructions and improve performance.
Unfortunately, the Key Locker instructions using the AVX states were
never implemented.
> Since the Key Locker assembly only supports 64-bit mode, it should also feel
> free to use registers xmm8-xmm15 for purposes such as caching the XTS tweaks.
> This would improve performance.
>
> Since the Key Locker assembly advances a large number of XTS tweaks at a time
> (8), I'm also wondering if it would be faster to multiply by x^8 directly
> instead of multiplying by x sequentially eight times. This can be done using
> the pclmulqdq instruction; see aes-xts-avx-x86_64.S which implements this
> optimization. Probably all CPUs that support Key Locker also support PCLMULQDQ.
I'll revisit the assembly code to incorporate your suggestions.
> I'm also trying to think of the best way to organize the Key Locker AES-XTS glue
> code. I see that you're proposing to share the glue code with the existing
> AES-XTS implementations. Unfortunately I don't think this ends up working very
> well, due to the facts that the Key Locker code can return errors and uses a
> different key type. I think that for now, I'd prefer that you simply copied the
> XTS glue code into aeskl-intel_glue.c and modified it as needed. (But make sure
> to use the new version of the glue code, which is faster.)
I agree; the proposed glue code looks messy due to the different return
errors and key types. Ard made a point earlier about establishing a
shared common code as they're logically quite similar. But I suppose it
is more practical to pursue a separate glue code at this point.
> For falling back to AES-NI, I think the cleanest solution is to call the
> top-level setkey, encrypt, and decrypt functions (the ones that are set in the
> xts-aes-aesni skcipher_alg), instead of calling lower-level functions as your
> current patchset does.
Yes, falling back is indeed one of the ugly parts of this series. Let me
retry this as you suggested.
> If you could keep the Key Locker assembly roughly stylistically consistent with
> the new aes-xts-avx-x86_64.S, that would be great too.
Okay.
> Do you happen to know if there's any way to test the Key Locker AES-XTS code
> without having access to a bare metal machine with a CPU that supports Key
> Locker? I tried a Sapphire Rapids based VM in Google Compute Engine, but it
> doesn't enumerate Key Locker. I also don't see anything in QEMU related to Key
> Locker. So I don't currently have an easy way to test this patchset.
No, there isn't currently an emulation option available to the public
that I'm aware of. this feature has been available on client systems
since the Tiger Lake generation.
> Finally, a high level question. Key Locker has been reported to be
> substantially slower than AES-NI. At the same time, VAES has recently doubled
> performance over AES-NI. I'd guess this leaves Key Locker even further behind.
> Given that, how useful is this patchset? I'm a bit concerned that this might be
> something that sounds good in theory but won't be used in practice. Are
> performance improvements for Key Locker on the horizon? (Well, there are the
> improvements I suggested above, which should help, but it sounds like main issue
> is the Key Locker instructions themselves which are just fundamentally slower.)
On our latest implementations, we've observed the Key Locker performance
on cryptsetup seems to be roughly the same as what we posted earlier
[2]. Yes, this sounds like a fair analogy to me, especially given your
vAES code.
Thanks,
Chang
[1]
https://lore.kernel.org/lkml/CAMj1kXGa4f21eH0mdxd1pQsZMUjUr1Btq+Dgw-gC=O-yYft7xw@mail.gmail.com/
[2]
https://lore.kernel.org/lkml/[email protected]/
On Mon, Apr 15, 2024 at 03:16:18PM -0700, Chang S. Bae wrote:
> > First, surely it's the case that in practice, all CPUs that support Key Locker
> > also support AVX? If so, then there's no need for the Key Locker assembly to
> > use legacy SSE instructions. It should instead target AVX and use VEX-coded
> > instructions. This would save some instructions and improve performance.
>
> Unfortunately, the Key Locker instructions using the AVX states were never
> implemented.
Sure, you could still use VEX-coded 128-bit instructions for everything other
than the actual AES (for example, the XTS tweak computation) though, right?
They're a bit more convenient to work with since they are non-destructive.
- Eric