[Openpvrsgx-devgroup] Mailing list openpvrsgx-devgroup
H. Nikolaus Schaller
hns at goldelico.com
Wed May 22 15:01:24 CEST 2024
Hi Julius,
> Am 21.05.2024 um 20:05 schrieb Julius Schwartzenberg <julius.schwartzenberg at gmail.com>:
>
> Hi Nikolaus,
>
> вт, 21 мая 2024 г. в 19:46, H. Nikolaus Schaller <hns at goldelico.com>:
>> Oh, interesting. I think it might be that gmail has enforced to use SPF - but in this case they seem to be the sender and not the recipient who checks for a valid sender.
>> And I had updated the mail server some months ago to correctly handle SPF on outgoing mails (got a lot of bounces from @gmail.com addresses).
>
> I am afraid that problem is still there. I've been trying to make it
> trigger messages, but nothing worked.
Hm. Strange. Unfortunately I can't look into the server logs to check if the mail did even arrive. Or if it did send a response which was then blocked by gmail.
Ah, maybe an idea for testing after my vacation: can you mail to the openpvrsgx-devgroup-join at letux.org address?
Well, there may also be a bug in the SPF record for @letux.org. There also seems to be a new requirement where the server fails (letux.org and goldelico.com run on the same server and share the SPF records): DMARC (https://mxtoolbox.com/emailhealth/goldelico.com/)
>
>> I have checked and it appears that you are not subscribed. Maybe subscription had failed. I have sent out an invitation/subscribe.
>>
>> Please let me know if it works now.
>
> It doesn't seem any message is reaching my gmail mailbox. I'll try it
> with another account.
Yes, please.
>
>
>> Indeed. They try to do useful things to get more customers but it is not always good to use.
>> There was a new beta test robot that suggests pull requests for known security vulnerabilities.
>> This morning it suggested something we had copied from upstream git.kernel.org. So nothing
>> to care about.
>
> Yeah I saw those PRs :)
>
>
>> In any case you can check the archive:
>>
>> https://lists.goldelico.com/pipermail/openpvrsgx-devgroup/2024-May/thread.html
>
> Yes indeed, I already found it. I am mainly doubting about whether it
> makes sense to write messages to both the mailing list and GitHub. I
> guess the mailing list may have a bigger audience.
>
>
>> I am on vacation the next days, but after I am back I'll try to work on the new DDK version.
>
> That would be great! Enjoy your vacation first!!
And afterwards we can find out what is wrong with the mailing list. The server needs an update
anyways (which will be a big step since it runs on a quite old OS and there are incompatiblities
to expect for PHP, Python, Apache, Let'sEncrypt configurations etc.)
>
>
>> Unfortunately I have no x86 device with an PVR/SGX GPU inside...
>
> Regarding x86, my idea was to try to add the missing parts from the
> Cedarview driver to your branch with the same version (and dropping
> the gma500 driver for now). If we could bring up an 5.x kernel on x86,
> that would be a huge leap forward. It would then at least work with
> the old binaries from the Cedarview package.
That looks like a good plan!
> A further step might be
> to have those changes in the other branches as well which might be
> able to give accelerated Android support. On x86 the only other option
> is Windows, but the versions that work with the chip are not updated
> anymore by Microsoft, so from a security point of view there is really
> nothing adequate available at this point.
Yes, that is the unfortunate lack of user-space source code...
I am not sure if the confidential part of the 1.8 repository contains
enough information, but back 10 years ago when it was leaked there
was an intensive discussion that anyone who opens the files and
reads them can no longer sign a copyright assignment for an
open source project. This may depend on the country someone
is living in, but this stopped people from reasoning how to write
a new user-space from scratch.
On the other hand I think since Imagination hasn't asked github
for apparently 10 years to delete the repo they don't care about
it. But without legal advice it is difficult to decide.
And, in the 10 years the community interested in GPU and SGX
drivers did shrink significantly. Maybe, because hardware using
it has been obsoleted almost (AFAIK AM3358 and some other
TI SoC are still in mass production).
Best regards,
Nikolaus
More information about the openpvrsgx-devgroup
mailing list