Conor Dooley
Posts
475Following
144Followers
157I like cycling, powerlifting, bad video games and metal.
Otherwise, I occupy my time with various bits in RISC-V land.
~useless, placeholder, website: https://www.conchuod.ie/
Conor Dooley
conor
Edited 1 year ago
Today's interesting discovery is that the c908 on the k230 (despite not really advertising it) supports Zicbom and the T-Head custom CMO stuff.
Apparently the Chinese documentation says it supports it but the English documentation just says "RISC-V Base Cache Management Operation ISA Extensions Version 1.0-rc2-2c97b28, 2021-11-05: Frozen.". In comparison, about a third of the user manual talks about the T-Head CMOs.
edit: I realise it is expected that the document would cover the custom stuff in far more detail, but for the other standard extensions it claims to support, it at least mentions that it supports the instruction and what it does.
https://occ-intl-prod.oss-ap-southeast-1.aliyuncs.com/resource//1699268369347/XuanTie-C908-UserManual.pdf
Made for an interesting conversation on IRC with dramforever, revy and cyy this afternoon :)
Apparently the Chinese documentation says it supports it but the English documentation just says "RISC-V Base Cache Management Operation ISA Extensions Version 1.0-rc2-2c97b28, 2021-11-05: Frozen.". In comparison, about a third of the user manual talks about the T-Head CMOs.
edit: I realise it is expected that the document would cover the custom stuff in far more detail, but for the other standard extensions it claims to support, it at least mentions that it supports the instruction and what it does.
https://occ-intl-prod.oss-ap-southeast-1.aliyuncs.com/resource//1699268369347/XuanTie-C908-UserManual.pdf
Made for an interesting conversation on IRC with dramforever, revy and cyy this afternoon :)
1
1
3
Conor Dooley
conor
Edited 1 year ago
Adding U-Boot support for the replacement riscv extension DT properties, come to find out that the extension detection there is completely broken, at least in QEMU.
It expects the ISA string to be 32 characters or less, probably true for all hardware it supports but untrue in QEMU for a long time. I guess there aren't too many ISA string parsers, but I do wonder how many others are broken in the face of the "fancier" strings.
It expects the ISA string to be 32 characters or less, probably true for all hardware it supports but untrue in QEMU for a long time. I guess there aren't too many ISA string parsers, but I do wonder how many others are broken in the face of the "fancier" strings.
1
0
0
Conor Dooley
conor
Ywe Caerlyn is my favourite poster on lkml. I have no idea what this is meant to be, but it is great:
https://www.youtube.com/watch?v=jpyrxj7IVqQ
https://www.youtube.com/watch?v=jpyrxj7IVqQ
1
0
1
Conor Dooley
conor
Whenever I read Nathan Chancellors recaps I always go and look at what
my contributions look like (see https://nathanchance.dev/posts/2023-cbl-retrospective/
for his far more comprehensive retrospective than my own brief look).
In 2023 this is what they looked like:
Authored: 123
Reported: 8
Tested: 29
Acked: 412
Reviewed: 372
In comparison, 2022 looked like:
Authored: 181
Reported: 21
Tested: 28
Acked: 45
Reviewed: 92
Two rather different years in terms of my kernel contributions, that's
for sure. In March or April I started co-maintaining dt-bindings, which
is the reason for the massive growth in terms of ack/review tags.
The number that is there is probably also underselling the amount of
reviewing actually done, since I usually avoid leaving a tag on vN of a
patch when one of the other maintainers already has, even if I spent a
bunch of time reviewing v(N-1).
On another note, the number of reported-by tags dropped considerably - I
like to chalk that up to the riscv port becoming more stable :)
my contributions look like (see https://nathanchance.dev/posts/2023-cbl-retrospective/
for his far more comprehensive retrospective than my own brief look).
In 2023 this is what they looked like:
Authored: 123
Reported: 8
Tested: 29
Acked: 412
Reviewed: 372
In comparison, 2022 looked like:
Authored: 181
Reported: 21
Tested: 28
Acked: 45
Reviewed: 92
Two rather different years in terms of my kernel contributions, that's
for sure. In March or April I started co-maintaining dt-bindings, which
is the reason for the massive growth in terms of ack/review tags.
The number that is there is probably also underselling the amount of
reviewing actually done, since I usually avoid leaving a tag on vN of a
patch when one of the other maintainers already has, even if I spent a
bunch of time reviewing v(N-1).
On another note, the number of reported-by tags dropped considerably - I
like to chalk that up to the riscv port becoming more stable :)
1
1
2
Conor Dooley
repeated
repeated
Drew Fustini
pdp7@fosstodon.orgI'm excited to announce I've joined @olofj's RISC-V Software team at Tenstorrent! If you aren't familiar with Tenstorrent, then check this recent talk from CEO Jim Keller https://www.youtube.com/watch?v=lPX1H3jW8ZQ
2
6
2
Conor Dooley
repeated
repeated
Asahi Linux
AsahiLinux@social.treehouse.systemsWe're proud to announce the official release of Fedora Asahi Remix! 🎉
Learn more:
https://asahilinux.org/fedora/
10
26
4
Conor Dooley
conor
Being in my 20s, I was confused as to why YouTube keeps showing me adverts for toys for young children. I realised all I watch on this particular computer is videos about programming and 10+ year old MMOs and it all started to make a lot of sense.
0
0
3
Conor Dooley
conor
Reading some patch discussion today (between two experienced developers/maintainers) and one cracked out a "I believe the kids these days would say '"Say you don't understand the
code without saying you don't understand the code."'
lmao, good one chief
code without saying you don't understand the code."'
lmao, good one chief
0
1
4
Conor Dooley
repeated
repeated
Krzysztof Kozlowski
krzk
I announced it before on social.kernel.org, mailing lists and finally in the Linux kernel (https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=c25223cba5aa9a536392933782d4a7df71d9093b). So one more time, same announcement:
None of the Samsung and Samsung Foundry platforms can bring any new `dtbs_check W=1` warning.
Contributor can easily test it, so sending code which introduces such warnings is considered close to sending code which does not compile. Fast step to get your maintainer grumpy.
None of the Samsung and Samsung Foundry platforms can bring any new `dtbs_check W=1` warning.
Contributor can easily test it, so sending code which introduces such warnings is considered close to sending code which does not compile. Fast step to get your maintainer grumpy.
0
1
6
Conor Dooley
conor
"Detection of individual cryptography extensions uses the unified software-based RISC-V discovery method.
At the time of writing, these discovery mechanisms are still a work in progress."
...from the documentation for ratified extensions :)
At the time of writing, these discovery mechanisms are still a work in progress."
...from the documentation for ratified extensions :)
1
0
4
Conor Dooley
conor
I see today is iteration N of hating email based workflows. My 0.02 € is that I hate both, clunky aul email and clunky new UIs :)
1
0
2
Conor Dooley
repeated
repeated
Matthew Garrett
mjg59@nondeterministic.computerI'm extremely reluctant to pass much comment on the Israel/Gaza situation because I grew up in an occupied territory where terrorists frequently targeted civilians and arguably the terrorism eventually resulted in a negotiated peace with a vaguely functioning power sharing arrangement, and I do not want that to be the moral of the story
1
1
2
Conor Dooley
repeated
repeated
Marcin Juszkiewicz 🙃
hrw@society.oftrolls.comIt is year 2077.
Computers still use Arm architecture.
Kernel/initramfs image are loaded using TFTP. Still over IPv4.
DeviceTree lives on.
1
2
2
Conor Dooley
conor
My x86 box died yesterday after 10 months of using it just fine*, either CPU or mobo being the problem. Pain in the arse, but also gonna force me to actually configure my newly acquired arm64 laptop. Good a time as any to give sway a shot too I suppose!
* apart from having to run EXPO to boot, /sigh
* apart from having to run EXPO to boot, /sigh
1
0
1
Conor Dooley
conor
Having always flown Ryanair short haul in Europe, my last two flights have been early morning Aer Lingus and Luthansa. It's so strange to me to see flights not only not at maximum capacity, but well under 2/3 of it
0
0
1
Conor Dooley
repeated
repeated
Asahi Linux
AsahiLinux@social.treehouse.systemsAsahi Linux is now shipping the world's first standards conformant GPU driver for Apple M1/M2! 🎉
https://rosenzweig.io/blog/first-conformant-m1-gpu-driver.html
0
0
0