ChanServ changed the topic of #armlinux to: ARM kernel talk [Upstream kernel, find your vendor forums for questions about their kernels] | https://libera.irclog.whitequark.org/armlinux
Misotauros has quit [Ping timeout: 264 seconds]
apritzel has quit [Ping timeout: 255 seconds]
Misotauros has joined #armlinux
DarkL0rd has quit [Remote host closed the connection]
heat has quit [Ping timeout: 255 seconds]
alpernebbi has quit [Ping timeout: 252 seconds]
alpernebbi has joined #armlinux
DarkL0rd has joined #armlinux
DarkL0rd has quit [Ping timeout: 260 seconds]
naoki has quit [Quit: naoki]
naoki has joined #armlinux
altf2o has quit [Remote host closed the connection]
altf2o has joined #armlinux
prabhakarlad has quit [Ping timeout: 244 seconds]
iivanov has joined #armlinux
hanetzer has quit [Ping timeout: 260 seconds]
hanetzer has joined #armlinux
altf2o has quit [Remote host closed the connection]
altf2o has joined #armlinux
DarkL0rd has joined #armlinux
DarkL0rd has quit [Ping timeout: 260 seconds]
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #armlinux
cbeznea has joined #armlinux
chipxxx has joined #armlinux
ajfriesen has quit [Ping timeout: 264 seconds]
audgirka has joined #armlinux
ajfriesen has joined #armlinux
prabhakarlad has joined #armlinux
altf2o has quit [Remote host closed the connection]
altf2o has joined #armlinux
sszy has joined #armlinux
<ardb>
linusw_: so i can see that your patches create a vmalloc region mapping for the kernel image, but where do you actually start executin from it??
<ardb>
i'd expect the kernel vmalloc mapping to be created first from asm, without creating the linear map
<ardb>
and then jump to it and do the rest from C code
<linusw_>
ardb: yeah I think I do that :D
<linusw_>
The identity map at asm execution should end up at 0xf1000000
<linusw_>
Hm what am I saying ... too early in the morning
<linusw_>
commits a91da54570856e3d3af4ba2884db71fbce06f70b and b78f63f4439bbfd02bfc628114ed0f63460e5570 alter the kernel virtual start execution address to KERNEL_OFFSET
<linusw_>
Then in this series I exploit that, so the kernel does not execute from PAGE_OFFSET but from KERNEL_OFFSET
<linusw_>
I think is the answer to the question, correct me if I'm wrong.
<linusw_>
It seems yesterday I confused myself by fixing a bug in mm/init.c and then fooling myself that it was still not working ... I fixed the bug actually mea culpa.
<linusw_>
Now I am onto 4G/4G splitting again, something much harder than kernel-in-vmalloc.
<ardb>
linusw_: ah ok i see
<linusw_>
I did some groundwork like a year+ ago to keep track of the physical sections for the kernel in the assembly, nothing goes quickly on my end :P
<ardb>
linusw_: just make sure to finish it before arch/arm gets removed :-)
<linusw_>
haha yeah my execuse is a new baby in the family
<linusw_>
kernel-in-vmalloc is the next stepping stone but I can't see a reason to merge that without also the 4G/4G split so now trying to fix both.
<linusw_>
at least I want some prototype up so we can see if this thing will even fly
apritzel has joined #armlinux
headless has joined #armlinux
altf2o has quit [Remote host closed the connection]
altf2o has joined #armlinux
Amit_T has joined #armlinux
Misotauros has quit [Ping timeout: 246 seconds]
<geertu>
linusw_: Congrats
<geertu>
ardb: That'll happen long after 2038, I assume?
<ardb>
geertu: right after arch/x86 :-)
<Xogium>
;)
<maz>
as long as arch/alpha stays around, we can remove anything.
<geertu>
maz: cannot?
<Xogium>
I like arm, tbh
<Xogium>
it is nice to work with and is sane, unlike x86 :D
<maz>
geertu: no, can. that's the only architecture I'd really want to keep! :)
<ardb>
let's start with arch/ia64 then
<geertu>
maz: My alpha broke ca. 20y ago
<maz>
geertu: I can probably ship you one if you really want.
<geertu>
it used to be my router
<maz>
yeah, I used to have a Multia as a router too.
<maz>
(4 network interfaces, including 2 on PCMCIA...)
<Xogium>
never played with other arch yet, must say. Very little with mips... Trying to figure out if I should attempt to recompile a kernel for mu ubiquiti managed switch ;)
<geertu>
maz: Multia was nice
<Xogium>
I mean, realistically speaking I definitely should, but have no clue if mainline would play nice
<Xogium>
Linux USW-Lite-8-PoE 3.18.24 #0 Thu Aug 30 12:10:54 2018 mips GNU/Linux
<geertu>
Xogium: does OpenWRT support that switch? I run OpenWRT on my edgerouter-x
<Xogium>
I've actually no idea
<Xogium>
let me check
<Xogium>
doesn't sound like it
haritz has quit [Remote host closed the connection]
<Xogium>
I mean I'm not afraid of making my own system using buildroot, certainly, but I don't even know how I'd flash it :p
<Xogium>
plus if things go wrong I'm a bit screwed because unlike on the edgerouter there's no serial over ethernet thingy
Misotauros has joined #armlinux
prabhakarlad has quit [Ping timeout: 244 seconds]
Misotauros has quit [Ping timeout: 246 seconds]
haritz has joined #armlinux
haritz has quit [Changing host]
haritz has joined #armlinux
Misotauros has joined #armlinux
headless has quit [Quit: Konversation terminated!]
<georgem>
hmm. now I'm tempted to flash my edgerouter-x
<arnd>
no wait, basic mainline support is there, just lacking the dts files for anything other than cisco-sg220
Misotauros has quit [Ping timeout: 246 seconds]
Misotauros has joined #armlinux
DarkL0rd has joined #armlinux
DarkL0rd has quit [Ping timeout: 255 seconds]
DarkL0rd has joined #armlinux
DarkL0rd has quit [Read error: Connection reset by peer]
DarkL0rd has joined #armlinux
DarkLord_ has joined #armlinux
DarkL0rd has quit [Ping timeout: 252 seconds]
<arnd>
maz: regarding alpha: mattst88 hasn't merged any bugfix patches through his tree since v5.15, you might need to get involved yourself if you care deeply about it working
<arnd>
on other obsolete architecture, I just noticed that ppc32 has dropped off NXP's product longevity program this year, they previously guaranteed availability until 2025, but not it's just whatever is left of old stock. Some qoriq products are still left (P5, T1/T2/T3/T4), but they are all 64-bit E5500/E6500 cores
<arnd>
geertu: OTOH NXP mcf5441x and Renesas SH7750 still list their respective 2025 EOL dates
<maz>
arnd: damn. I need to power on my DS20L and see if it is still alive...
<geertu>
arnd: Quoting someone from the AU navy: "Given how long we used the m68k VME boards, the PPC MVME boards will defiitely be in use after 2038"
Misotauros has quit [Ping timeout: 246 seconds]
sudeepholla has quit [Remote host closed the connection]
torez has joined #armlinux
torez has quit [Read error: Connection reset by peer]
torez has joined #armlinux
sudeepholla has joined #armlinux
<arnd>
geertu: there is probably enough old stock around to keep them going for a while, but at some point they may need to move to mvme8100, which I think is the only one in production that has a non-EOL CPU
DarkL0rd has joined #armlinux
DarkLord_ has quit [Ping timeout: 252 seconds]
DarkLord_ has joined #armlinux
DarkLord_ has quit [Remote host closed the connection]
DarkL0rd has quit [Ping timeout: 246 seconds]
DarkLord_ has joined #armlinux
<linusw_>
geertu: actually my son is like 20 months old so not so small :)
<linusw_>
arnd: I rebased my TTBR switching hacks on v6.1-rc1. Funny things happened: boot goes all the way to executing init on v6.0-rc1 before any fatal exception occurs.
<arnd>
nice!
<linusw_>
on v6.1-rc1 I have to disable MTD and network drivers otherwise something exceptionizes earlier and hangs boot... I wonder what!
<linusw_>
also why it happened in v6.1-rc1...
<linusw_>
I'm a bit puzzled why the kernel would fault in userspace stuff before userspace is even started but I guess there are things I don't understand.
<arnd>
linusw_: maybe mtd has some code that assumes the kernel is part of the linear map, and it breaks with kernel-in-vmalloc?
<linusw_>
arnd: no I tested with just kernel-in-vmalloc and that works fine...
<linusw_>
It's when I start switching TTBR0 around it breaks.
<linusw_>
I guess I'll figure it out.
viorel has joined #armlinux
audgirka has quit [Remote host closed the connection]
atorgue has joined #armlinux
DarkLord_ has quit [Remote host closed the connection]
DarkL0rd has joined #armlinux
Misotauros has joined #armlinux
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #armlinux
heat has joined #armlinux
DarkL0rd has quit [Remote host closed the connection]