* [9front] Error compiling libc/amd64/main9.s and/or main9p.s
@ 2024-03-26 16:37 Scott Flowers
2024-03-26 16:46 ` ori
2024-03-26 16:51 ` Jacob Moody
0 siblings, 2 replies; 7+ messages in thread
From: Scott Flowers @ 2024-03-26 16:37 UTC (permalink / raw)
To: 9front
I ran a sysupdate this morning and it won't build on amd64. It's hanging on /sys/src/libc/amd64/main9.s with an error like below.
6a main9.s
6a main9p.s
main9.s:6 main9p.s:6 syntax error, last name: _callmain
syntax error, last name: _callmain
line: LNAME.: $$5 line
inst: LNAME.= expr
line: LNAME.: $$5 line
inst: LNAME.= expr
saw LNAME
saw LNAME
mk: 6a main9.s : exit status=6a 692473: error
mk: for(i in 9sys ... : exit status=rc 691919: mk 692470: error
mk: date for (i ... : exit status=rc 691888: mk 691918: error
I tried doing mk nuke, and then mk libs, and it fails at the same place, like this.
6a main9.s
main9.s:6 6a main9p.s
syntax error, last name: _callmain
line: LNAME.: $$5 line
inst: LNAME.= expr
main9p.s:6 saw LNAME
syntax error, last name: _callmain
mk: 6a main9.s : exit status=6a 703834: errorline: LNAME.: $$5 line
inst: LNAME.= expr
saw LNAME
mk: for(i in 9sys ... : exit status=rc 703070: mk 703825: error
mk: date for (i ... : exit status=rc 702782: rc 702983: mk 703068: error
Any suggestions?
Thanks!
Scott Flowers
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [9front] Error compiling libc/amd64/main9.s and/or main9p.s
2024-03-26 16:37 [9front] Error compiling libc/amd64/main9.s and/or main9p.s Scott Flowers
@ 2024-03-26 16:46 ` ori
2024-03-26 16:51 ` Jacob Moody
1 sibling, 0 replies; 7+ messages in thread
From: ori @ 2024-03-26 16:46 UTC (permalink / raw)
To: 9front
https://inbox.vuxu.org/9front/d9bdf2b5-8da3-4cc7-8f18-40bc3a5d0a84@posixcafe.org/T/#u
note, you can also grab the latest compilers from the nightly builder:
9fs iso
cp /n/iso/amd64/bin/6^(a c l) /amd64/bin
Quoth Scott Flowers <flowerss@cranky.ca>:
> I ran a sysupdate this morning and it won't build on amd64. It's hanging on /sys/src/libc/amd64/main9.s with an error like below.
>
> 6a main9.s
> 6a main9p.s
> main9.s:6 main9p.s:6 syntax error, last name: _callmain
> syntax error, last name: _callmain
> line: LNAME.: $$5 line
> inst: LNAME.= expr
> line: LNAME.: $$5 line
> inst: LNAME.= expr
> saw LNAME
> saw LNAME
> mk: 6a main9.s : exit status=6a 692473: error
> mk: for(i in 9sys ... : exit status=rc 691919: mk 692470: error
> mk: date for (i ... : exit status=rc 691888: mk 691918: error
>
> I tried doing mk nuke, and then mk libs, and it fails at the same place, like this.
>
> 6a main9.s
> main9.s:6 6a main9p.s
> syntax error, last name: _callmain
> line: LNAME.: $$5 line
> inst: LNAME.= expr
> main9p.s:6 saw LNAME
> syntax error, last name: _callmain
> mk: 6a main9.s : exit status=6a 703834: errorline: LNAME.: $$5 line
> inst: LNAME.= expr
>
> saw LNAME
> mk: for(i in 9sys ... : exit status=rc 703070: mk 703825: error
> mk: date for (i ... : exit status=rc 702782: rc 702983: mk 703068: error
>
> Any suggestions?
>
> Thanks!
>
> Scott Flowers
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [9front] Error compiling libc/amd64/main9.s and/or main9p.s
2024-03-26 16:37 [9front] Error compiling libc/amd64/main9.s and/or main9p.s Scott Flowers
2024-03-26 16:46 ` ori
@ 2024-03-26 16:51 ` Jacob Moody
2024-03-26 17:15 ` Scott Flowers
2024-03-27 0:22 ` sl
1 sibling, 2 replies; 7+ messages in thread
From: Jacob Moody @ 2024-03-26 16:51 UTC (permalink / raw)
To: 9front
There was an update that requires an update to the assembler and linker because a new instruction was added.
This was called out on the mailing list when it happened.
If you didn't 'mk nuke' your libc you can build the assembler and linker first:
for(i in (6a 6l))@{ cd /sys/src/cmd/^$i && mk install }
Then you may rebuild the rest of the system.
If you did 'mk nuke':
* if you use cwfs, copy older libc from dump and then do the above
9fs dump cp /n/dump/2024/...
* if you use hjfs, copy updated binaries from the iso box.
9fs iso && cp /n/iso/amd64/bin/^(6a 6c 6l) /amd64/bin/
It would be nice to have a way of these updates notifying people
of such manual steps. But for the time being I would suggest perusing
the mailing lists when doing large updates.
Thanks,
moody
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [9front] Error compiling libc/amd64/main9.s and/or main9p.s
2024-03-26 16:51 ` Jacob Moody
@ 2024-03-26 17:15 ` Scott Flowers
2024-03-27 0:22 ` sl
1 sibling, 0 replies; 7+ messages in thread
From: Scott Flowers @ 2024-03-26 17:15 UTC (permalink / raw)
To: 9front
Thanks for the input.
i got the assembler and linker updated and then the system built normally. I did the same process with the arm64 bits, and got the system files for that arch updated and built a new kernel as well so my pxe boot Pis would be updated too.
I'll make sure to go back through the mailing list next time I want to update.
Scott
On Tue, 26 Mar 2024, at 10:51, Jacob Moody wrote:
> There was an update that requires an update to the assembler and linker because a new instruction was added.
> This was called out on the mailing list when it happened.
>
> If you didn't 'mk nuke' your libc you can build the assembler and linker first:
>
> for(i in (6a 6l))@{ cd /sys/src/cmd/^$i && mk install }
>
> Then you may rebuild the rest of the system.
>
> If you did 'mk nuke':
> * if you use cwfs, copy older libc from dump and then do the above
> 9fs dump cp /n/dump/2024/...
> * if you use hjfs, copy updated binaries from the iso box.
> 9fs iso && cp /n/iso/amd64/bin/^(6a 6c 6l) /amd64/bin/^(6a 6c 6l) /amd64/bin/
>
> It would be nice to have a way of these updates notifying people
> of such manual steps. But for the time being I would suggest perusing
> the mailing lists when doing large updates.
>
> Thanks,
> moody
>
>
>
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [9front] Error compiling libc/amd64/main9.s and/or main9p.s
2024-03-26 16:51 ` Jacob Moody
2024-03-26 17:15 ` Scott Flowers
@ 2024-03-27 0:22 ` sl
2024-03-27 14:55 ` Scott Flowers
1 sibling, 1 reply; 7+ messages in thread
From: sl @ 2024-03-27 0:22 UTC (permalink / raw)
To: 9front
added fqa 5.4.2 to document this:
http://fqa.9front.org/fqa5.html#5.4.2
sl
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [9front] Error compiling libc/amd64/main9.s and/or main9p.s
2024-03-27 0:22 ` sl
@ 2024-03-27 14:55 ` Scott Flowers
2024-03-27 15:16 ` Jacob Moody
0 siblings, 1 reply; 7+ messages in thread
From: Scott Flowers @ 2024-03-27 14:55 UTC (permalink / raw)
To: 9front
I'm so proud that me being dumb has led to an update in the docs.
Scott
flowerss@cranky.ca
On Tue, 26 Mar 2024, at 18:22, sl@stanleylieber.com wrote:
> added fqa 5.4.2 to document this:
>
> http://fqa.9front.org/fqa5.html#5.4.2
>
> sl
>
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [9front] Error compiling libc/amd64/main9.s and/or main9p.s
2024-03-27 14:55 ` Scott Flowers
@ 2024-03-27 15:16 ` Jacob Moody
0 siblings, 0 replies; 7+ messages in thread
From: Jacob Moody @ 2024-03-27 15:16 UTC (permalink / raw)
To: 9front
For what it's worth it's really not you being dumb, this is a flaw
in how we distribute information. Our information channels are well
detached from our update channels.
I think this is the 10th or so time someone has stumbled in to this,
so I am glad sl added it to the fqa.
We've discussed a solution to this a few times but nothing has stuck.
Let me give an idea then and we can find something that sticks.
Maybe /sys/lib/sysupdate (not attached to the path) could have a file
named after a commit with specific update instructions. Sysupdate
could then get the list of commits it is updating through, and cat
out those files from that directory.
Thanks,
moody
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2024-03-27 15:18 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-03-26 16:37 [9front] Error compiling libc/amd64/main9.s and/or main9p.s Scott Flowers
2024-03-26 16:46 ` ori
2024-03-26 16:51 ` Jacob Moody
2024-03-26 17:15 ` Scott Flowers
2024-03-27 0:22 ` sl
2024-03-27 14:55 ` Scott Flowers
2024-03-27 15:16 ` Jacob Moody
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).