Trying to figure out why your attestation is failing

Trying to figure out why your attestation is failing

[Konstantin Ryabitsev]

Neil:

I am keeping an eye on attestation emails sent to the signatures list 
(https://lore.kernel.org/signatures/) and I noticed that b4 isn't 
properly finding your attestation data. I poked at this series:

- thread: https://lore.kernel.org/linux-amlogic/20200907081825.1654-1-narmstrong@baylibre.com
- attestation: https://lore.kernel.org/signatures/159946699508.19944.11132773619713429568@kernel.org

When we generate attestation hashes from the patches you submitted, we 
get a different value for the "i" hash:

$ git mailinfo m p < 0001-dt-bindings-display-amlogic-meson-vpu-add-bindings-f.patch | egrep '^(Author|Email|Subject)' | sha256sum
7158b7d79fbd5342923a63ea14701fa75061a856f6eb30c9dfe83672894113ba -

Apparently, when you run it on your system during attestation generation 
step, you get a different hash for i:

3753d76c-4b385c46-aa69b83c:
  i: 3753d76ca01efcc78b682dc0f9feb582d67a1e9a6a89b0519cdf54a36eee5e0b
  m: 4b385c46651635915366eba13de55a445ae3d41afd2d0e5da8bad270a08c0e70
  p: aa69b83c2b5d484b33deec0c59135fdfe7290b658f29527590261abab8a4ac51

This is why when we're trying to find your attestation document, we are 
not finding it:

  Cache says looking up 7158b7d7-4b385c46-aa69b83c = 404

I am curious why this happens -- either the patches generated on your 
system have different Author/Email/Subject values until git-send-email actually
runs, or git mailinfo output has changed between revisions. I would like to
understand this better so we can properly address it, if you're willing to lend
a hand.

Best,
Konstantin

Re: Trying to figure out why your attestation is failing

[Neil Armstrong]

Hi,

On 08/09/2020 19:58, Konstantin Ryabitsev wrote:
> Neil:
> 
> I am keeping an eye on attestation emails sent to the signatures list 
> (https://lore.kernel.org/signatures/) and I noticed that b4 isn't 
> properly finding your attestation data. I poked at this series:
> 
> - thread: https://lore.kernel.org/linux-amlogic/20200907081825.1654-1-narmstrong@baylibre.com
> - attestation: https://lore.kernel.org/signatures/159946699508.19944.11132773619713429568@kernel.org
> 
> When we generate attestation hashes from the patches you submitted, we 
> get a different value for the "i" hash:
> 
> $ git mailinfo m p < 0001-dt-bindings-display-amlogic-meson-vpu-add-bindings-f.patch | egrep '^(Author|Email|Subject)' | sha256sum
> 7158b7d79fbd5342923a63ea14701fa75061a856f6eb30c9dfe83672894113ba -
> 
> Apparently, when you run it on your system during attestation generation 
> step, you get a different hash for i:
> 
> 3753d76c-4b385c46-aa69b83c:
>   i: 3753d76ca01efcc78b682dc0f9feb582d67a1e9a6a89b0519cdf54a36eee5e0b
>   m: 4b385c46651635915366eba13de55a445ae3d41afd2d0e5da8bad270a08c0e70
>   p: aa69b83c2b5d484b33deec0c59135fdfe7290b658f29527590261abab8a4ac51
> 
> This is why when we're trying to find your attestation document, we are 
> not finding it:
> 
>   Cache says looking up 7158b7d7-4b385c46-aa69b83c = 404
> 
> I am curious why this happens -- either the patches generated on your 
> system have different Author/Email/Subject values until git-send-email actually
> runs, or git mailinfo output has changed between revisions. I would like to
> understand this better so we can properly address it, if you're willing to lend
> a hand.

Indeed, I usually only add partial To/Cc directly in the patches and the patchset-wise
to git send-email, would it be the issue ?

Neil

> 
> Best,
> Konstantin
>