klange changed the topic of #osdev to: Operating System Development || Don't ask to ask---just ask! || For 3+ LoC, use a pastebin (for example https://gist.github.com/) || Stats + Old logs: http://osdev-logs.qzx.com New Logs: https://libera.irclog.whitequark.org/osdev || Visit https://wiki.osdev.org and https://forum.osdev.org || Books: https://wiki.osdev.org/Books
LittleFox has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
LittleFox has joined #osdev
<heat> ah i fixed my gcc build bug, i think
<heat> my ext2 symlink reading code was looking at the wrong filesystem struct, which was basically making it race with inode writeback. so all long symlinks read garbage right after being created
bauen1 has joined #osdev
hwpplayer1 has joined #osdev
<Ermine> Seems like bionic has no posix_spawn
<heat> posix purists in shambles
<Ermine> given bionic's purpose they can probably afford not caring about posix. But I was surprised
<heat> anyone can afford not caring about posix, it's fucking posix after all
<heat> "oh no now my linux program cant be ported to dragonflybsd!!!!"
<Ermine> that's what program author would say
<Ermine> not libc author
<heat> no, dalias would praise posix because it's so useful!!!!
<heat> glibc would not
<heat> all behold GNU OPERATING SYSTEM SUPREMACY
<heat> RE_SEARCH MY OBSTACK
<Ermine> and yet they try to be compliant-ish?
<heat> sure
<heat> for various reasons it's beneficial to be compatible-ish
<heat> but they dont go out of their way to be POSIX (see musl ioctl debacle)
<Ermine> when did that happen?
<heat> uhhhh always
<heat> basically the glibc ioctl signature is int ioctl(int fd, unsigned long cmd, ...); the musl ioctl signature is int ioctl(int fd, int cmd, ...);
<heat> guess why? because STREAMS mandated that signature
<heat> this causes very subtle incompatibility from time to time
<Ermine> but dalias chose to int...
ThinkT510 has quit [Ping timeout: 276 seconds]
<kof673> obstack is one thing i usually have to patch for gcc 2.x or binutils or so because new compilers don't like how it does arithmetic on void * or something, don't recall :D
<heat> GNU C allows for void * arithmetic
<kof673> i would never have heard of it except for that lol
<kof673> well it is a macro, don't recall details
<kof673> i think use memcpy() where it copies a pointer or similar
<kof673> * i use
<Ermine> well, int vs unsigned long is pedantism moment
<Ermine> while no posix_spawn in bionic is likely because android is not your linux distro and user apps are not expected to execute programs like this
<heat> like what?
<heat> android has more than user apps
<Ermine> that 'more than' part is pretty much controlled and probably gets away with execve/clone*
<Ermine> (and google play now has a policy on executing third party programs, so termux is no longer there)
<vin> Do you a reliable way to get column/bank group/bank/row from a physical address? I am currently doing this `static DDR4Address decode_address(uint64_t phys_addr) { DDR4Address addr; addr.column = (phys_addr >> 6) & 0xF; addr.bank = (phys_addr >> 10) & 0x3; addr.bank_group = (phys_addr >> 12) & 0x3; addr.row = (phys_addr >> 14) & 0xFFFF; return addr; }` But am doubtful if its correct.
<Mutabah> for what purpose?
<vin> Mutabah: I am measuring the latency/throughput impact of global row buffer misses
<Mutabah> Huh, interesting
<vin> per bank. So got to make sure I am selective about bank accesses
<Mutabah> Find some documentation on the ram chips in question?
<vin> Yes https://download.datasheets.com/pdfs/2016/4/11/3/23/30/408/hyn_/manual/591772121422088591724984739574computing_ds_8gb_ddr4m-verbased_rdimmsrev.1.3.pdf.pdf from the address table for 16 GB is how I wrote the above decode function but I would like to verify some other way that this is how its actually mapped
hwpplayer1 has quit [Quit: bbl]
blockhead has joined #osdev
goliath has quit [Quit: SIGSEGV]
gog has quit [Quit: byee]
agent314 has quit [Ping timeout: 276 seconds]
nadja has quit [Ping timeout: 260 seconds]
nadja has joined #osdev
<kof673> <replies to ghost self> # sleep # sleep! could also go for rpg-style, sleep sleep2 sleep3
agent314 has joined #osdev
<geist> vin: that's almost certainly not reliable, though it may be for a particular vendor/family/etc of cpu and controller
<vin> geist: I have been trying to get https://github.com/helchr/reMap working on my server -- even with porting to right perf events it still doesn't seem to work.
<bslsk05> ​helchr/reMap - (2 forks/7 stargazers)
* geist nods
terrorjack4 has quit [Quit: The Lounge - https://thelounge.chat]
<zid`> heat: that's fine? cherry pick out the one patch into master, then rebase it out of your feature
<geist> thr reason i say that is for example on a threadripper machine ihave there are config options in the bios to adjust how addresses are mapped
<zid`> or not
<geist> ie, is it interleaved between controllers, and at what granularity, etc
<geist> then i would assume within a single controller depending on how many banks are populated, etc
<geist> i'm going to guess though that in *general* there's a pattern that most things follow, based on variables of the particular configuration
<geist> ie, maybe column first then bank then controller, etc
<heat> zid`, yeah but my branch's topic is atm background noise compared to the unrelated patches in it
<geist> i do that sometimes and then i end up going to the main branch and cherry picking stuff over to it
<geist> like little fixes or whatnot i did on a topic branch
terrorjack4 has joined #osdev
<geist> if you dont let things get too out of hand it works reasonably well. keeps you from needing to keep switching branch until there's a good stopping point
<geist> side ntoe someone at work was talking about git branchless as a useful tool. i should look into it
<heat> i started doing a tcp rewrite back in august, then started package management, unrelated fixes, a bunch of builds uncovered a bunch of bugs, improved my vt emulator, and now i've been doing gcc builds for 3 or 4 days
<heat> im now fighting memory fragmentation
<geist> yep, i have a bunch of wip branches for that sort of thing indeed
<heat> and the tcp rewrite _still isn't committed_
<geist> though i'm pretty good about not stacking them in one topic branch
<geist> at least make WIP comits and push them
<geist> you dont want to lose your work
<geist> i'm religious about that
<heat> git diff --stat | wc -l
<heat> 63
<heat> welp
<geist> well that's your problem. you are playing with fire bigtime
<heat> oh definitely
<geist> stop doing that *right now*
<zid`> Just cherry pick everything interesting :P
<geist> that's the beauty of git, you can commit often
<geist> and just fix it in post
<geist> theres no reason to not make WIP commits at least once a day
<geist> if not just 'oh this works right now, snapshot' so that if you break it in the next 30 minutes you have something to diff against
<heat> everything's interesting but some bits are not quite stable or quite finished
<heat> and i tend to not commit until i'm reasonably happy
<zid`> git rebase -i @{1970}
<geist> zid`: none of this applies to heat because he's not making commits
<zid`> what does he have then?
<heat> oh i have commits
<geist> that's kinda my point. make commits early and often then you have something to work with
<heat> and a lot of uncommitted stuff too
<zid`> he said he was dumping random commits into branches that weren't related to the fixes
<zid`> I told him to cherry pick it out into a diff branch then rebase it away in the original
<heat> i'll need to suck it up and git add -p the things i haven't been committing
<geist> oh the diff above he was pointing out that he has 63 modified files
<geist> and hasn't been committing them
<geist> yah i literally feel icky having uncommitted files. it feels so good to get it committed somewhere
<zid`> I need to stop stashing shit
<zid`> and just commit it
<zid`> stashes are much easier to forget about or lose
<geist> yah also i switch between computers a lot, so having commits helps me move stuff between
<heat> ok dad i'll commit things
<geist> good. and take out the damn trash
<vin> geist: From remap I modified main.cpp https://ctxt.io/2/AAB41gdqEQ to determine which bits of the physical address is responsible for channel/bank/bank group. However, It seems to only capture channel 0, bank group 0, bank 0 -- so hard to tell which bits are for what.
<bslsk05> ​ctxt.io: Context – share whatever you see with others in seconds
<heat> git diff --stat | wc -l
<heat> 28
<heat> solid.
heat has quit [Ping timeout: 252 seconds]
Arthuria has quit [Ping timeout: 264 seconds]
HavenE has joined #osdev
HavenE has quit [Quit: cya]
truebad0ur has quit [Read error: Connection reset by peer]
truebad0ur has joined #osdev
vdamewood has joined #osdev
ThinkT510 has joined #osdev
GeDaMo has joined #osdev
spare has joined #osdev
karenw has joined #osdev
<the_oz> I use sublime text as my editor because it never loses editing history, so if some retard os decides it's time, or florida lightning storm, or etc. I need to integrate git and make THAT never lose history too
<kof673> i don't know about history, but vim just leaves .swp files around as long as a file is open, so "recovers" using those
<the_oz> I hate vim, emacs. NOMODES
<kof673> not arguing that :D just never had a problem with that feature
xenos1984 has quit [Ping timeout: 248 seconds]
xenos1984 has joined #osdev
<the_oz> yeah
<the_oz> is all personal taste, and not a "you should" etc.
<the_oz> nor comparative definitions, others definitely have that, more like me pondering if their git integration can even do it. auto save to a specific branch to the point I don't need to worry about it
<the_oz> have like a branch autobot committing...
<the_oz> to ease the mental load
<kof673> that sounds like continuous integration so agile it hurts :D
<the_oz> IT IS CURTIME, I DID WORK, PLAN MORE WORK, NEXT
<the_oz> 10,000 megachecks per sond
<the_oz> second
xenos1984 has quit [Ping timeout: 264 seconds]
aejsmith has quit [Quit: Lost terminal]
aejsmith has joined #osdev
xenos1984 has joined #osdev
netbsduser` has joined #osdev
<geist> my experience is an individual computer is very fallible
<geist> so do not depend on it not completely losing everything on it on any given day
<Ermine> android got lags while it was building cmake
<Ermine> but.. I've got build errors
Left_Turn has joined #osdev
goliath has joined #osdev
<johnjaye> geist: that is true but setting up a backup solution seems to have a same issue
<johnjaye> meaning. if i backup everything to a drive then that drive could fail. at least that's what i think mentally
<johnjaye> more practically i don't know how to setup a network share so i can backup to it from any pc on my LAN. at least that's what i would be using a share for
the_oz has quit [Remote host closed the connection]
the_oz has joined #osdev
<the_oz> the only way is to assume failures will happen, like my own stupidity does constantly
benlyn has joined #osdev
kneskade has joined #osdev
benlyn has quit [Ping timeout: 246 seconds]
benlyn has joined #osdev
kneskade has quit [Ping timeout: 246 seconds]
kneskade has joined #osdev
benlyn has quit [Ping timeout: 252 seconds]
benlyn has joined #osdev
kneskade has quit [Ping timeout: 252 seconds]
kneskade has joined #osdev
craigo has joined #osdev
craigo has quit [Read error: Connection reset by peer]
benlyn has quit [Ping timeout: 252 seconds]
craigo has joined #osdev
benlyn has joined #osdev
kneskade has quit [Ping timeout: 272 seconds]
kneskade has joined #osdev
benlyn has quit [Ping timeout: 272 seconds]
goliath has quit [Quit: SIGSEGV]
benlyn has joined #osdev
kneskade has quit [Ping timeout: 248 seconds]
pabs3 has quit [Ping timeout: 248 seconds]
ramenu has quit [Ping timeout: 276 seconds]
kneskade has joined #osdev
agent314 has quit [Ping timeout: 272 seconds]
benlyn has quit [Ping timeout: 272 seconds]
pabs3 has joined #osdev
gog has joined #osdev
spare has quit [Remote host closed the connection]
jedesa has joined #osdev
Turn_Left has joined #osdev
Left_Turn has quit [Ping timeout: 276 seconds]
Left_Turn has joined #osdev
Turn_Left has quit [Ping timeout: 244 seconds]
heat has joined #osdev
Turn_Left has joined #osdev
Left_Turn has quit [Ping timeout: 246 seconds]
housemate has joined #osdev
housemate has quit [Remote host closed the connection]
air has joined #osdev
blockhead has quit []
karenw has quit [Ping timeout: 244 seconds]
melnary has quit [Read error: Connection reset by peer]
melnary has joined #osdev
_ngn has quit [Remote host closed the connection]
_ngn has joined #osdev
housemate has joined #osdev
goliath has joined #osdev
raphaelsc has joined #osdev
Left_Turn has joined #osdev
Turn_Left has quit [Ping timeout: 252 seconds]
dostoyevsky2 has quit [Quit: leaving]
dostoyevsky2 has joined #osdev
_ngn has quit [Ping timeout: 260 seconds]
_ngn has joined #osdev
npc has joined #osdev
<heat> "People are hung up on Reiser having killed someone. I say, who cares? Maybe the bitch deserved it? We’ll never know."
<heat> thanks hacker news
jedesa has quit [Quit: jedesa]
<gog> yikes
<johnjaye> ok all the tutorials i'm finding on setting up a network share say to use samba
<johnjaye> i'm guessing mac and android can access those. maybe with an app or something. network share only makes sense to me if my pc and phone can both access it
<kof673> i don't follow modern stuff, but: samba does the ms stuff, or there is nfs if it is just filesystem, or sshfs even (fuse clients, and e.g. gnome vfs i think has it built in)
Turn_Left has joined #osdev
<kof673> i mean, not sure people use samba unless they specifically have windows clients
housemate has quit [Quit: "I saw it in a tiktok video and thought that it was the most smartest answer ever." ~ AnonOps Radio [some time some place] | I AM THE DERIVATIVE I AM GOING TANGENT TO THE CURVE!]
Left_Turn has quit [Ping timeout: 245 seconds]
<kof673> or if it is just to push/retrieve backups somewhere, scp is just ftp-like interface, can be scripted surely too, etc.
<heat> rsync AFAIK is the gold standard for this kind of stuff
<zid`> samba's dirt simple and works at least
<zid`> saves you having to use winscp or whatever on windows end, can just mount it, etc
<johnjaye> windows has built in nfs support now but you have to enable it
<Ermine> heat: you reading hackernews comments is like me reading phoronix comments
<heat> i read all of the comments
<heat> it's my way to avoid doing actually productive things
<Ermine> heh, can't blame
<heat> i'm also just building gcc in a loop and finding operational issues soo
<heat> not too interesting
<heat> found a bug in my buddy allocator that stopped the merging of free pages into higher order ones, found a bug where I GFP_ATOMIC'd a 512KB chunk of pages (which obviously isn't promptly available)
<Ermine> bruh allocator
<nikolar> brullocator
<zid`> I prefer a mate manager
<zid`> just because more instances of 'mm' are better
<zid`> vmm_mm_mm()
<nikolar> kek
<heat> kekllocator
<johnjaye> heat: have you deep dived into the gcc source code or just had to use the manual a bit?
<johnjaye> to test the allocator i mean
<heat> hm?
<heat> the build itself is the stress test
<johnjaye> ah ok. i think i tried building gcc once but ran out of memory
<johnjaye> and nothing really said how much memory you need. curiously
<heat> well it depends on the toolchain you use, libc, allocator(s), the kernel, your number of threads, etc
Left_Turn has joined #osdev
Left_Turn has quit [Read error: Connection reset by peer]
Turn_Left has quit [Ping timeout: 244 seconds]
<kof673> older versions like 2.x 3.x is less. and possibly depends what languages perhaps :D in any case, after those i believe there is a file or three that needs "lots" of ram, even if the majority does not, certain files do (one time i ran into that i just copied the source tree and build dir to another similar machine, built the one object, then copy back)
<kof673> it may just be a few files anyways...
<kof673> *copied the single object back to local machine
<kof673> also the "meta" stuff that generates source, possibly needs "lots" too
zenmov has joined #osdev
Left_Turn has joined #osdev
Left_Turn has quit [Remote host closed the connection]
Left_Turn has joined #osdev
Left_Turn has quit [Max SendQ exceeded]
xenos1984 has quit [Ping timeout: 272 seconds]
xenos1984 has joined #osdev
guideX has quit [Ping timeout: 260 seconds]
guideX has joined #osdev
xenos1984 has quit [Ping timeout: 248 seconds]
<heat> woooo gcc built
<sortie> heat congrats
<heat> it takes a little over an hour with all the decompressing and compressing around
<heat> but like 50min of actual build which is honestly a solid start
guideX has quit [Ping timeout: 252 seconds]
housemate has joined #osdev
<zid`> does it work? :P
xenos1984 has joined #osdev
<heat> yep it do be workin
<nikolar> No compile onyx
<nikolar> On onyx
<heat> what
<nikolar> Erm, *now
<heat> compile onyx on onyx?
<nikolar> Yes
<heat> we'll get there, Eventually
<heat> gcc is a much larger build and more technically impressive
<heat> even cmake
<nikolar> Just try and see how far it goes heh
<heat> i've built the kernel
<heat> it's ez
<heat> problem is that 1) i'm missing gn + ninja for userspace 2) i suspect some bits of the build system assume you're cross compiling
<nikolar> Interesting
<heat> my os is generally a very quick build
<heat> very lightweight code
<heat> even an LTO'd kernel takes just a couple of seconds (maybe 20 or 30) to LTO it up
<nikolar> Not horrible
<Ermine> is gn that hard to build
<netbsduser`> i finally got around to figuring out how my iterative layered async i/o dispatcher works and rewriting it cleanly and with more comments and a full coverage test suite
<Ermine> lightweight vs minimal
<netbsduser`> it is the signature function of my kernel and deserved the attention. i don't know of any others that go quite as hard as it
<heat> Ermine, nah it's pretty easy
<Ermine> "lightweight vs minimal" was about onyx
<heat> my code is categorically not minimal
<Ermine> yeah
<Ermine> hence "vs"
<heat> it's just that the worse C++ creation i can pull off is still 10x faster to compile than the average LLVM file
nur has quit [Ping timeout: 245 seconds]
<netbsduser`> congrats on the gcc build
<heat> thanks
<heat> it truly was a really nice stress test on the whole system
<netbsduser`> bloody c++ will do that
<netbsduser`> it stress tests my desktop hard enough
<heat> not just C++
<netbsduser`> the hundred thousand forks probably help too
<heat> at the tail ends of the build you get a whole bunch of page cache all over memory
<heat> and some really nice half-meg argv invocations
zenmov has quit [Quit: leaving]
housemate has quit [Quit: "I saw it in a tiktok video and thought that it was the most smartest answer ever." ~ AnonOps Radio [some time some place] | I AM THE DERIVATIVE I AM GOING TANGENT TO THE CURVE!]
Dead_Bush_Sanpai has quit [Read error: Connection reset by peer]
Dead_Bush_Sanpa1 has joined #osdev
jjuran has quit [Ping timeout: 246 seconds]
Dead_Bush_Sanpa1 is now known as Dead_Bush_Sanpai
jjuran has joined #osdev
guideX has joined #osdev
npc has quit [Read error: Connection reset by peer]
op has joined #osdev
<johnjaye> heat: what's the issue with llvm? you mean in general?
<heat> llvm is veeeeeeeeery slow to compile
<johnjaye> eh. that's bad considering it's supposed to be more modern than gcc
<johnjaye> or rather you'd think they would prioritize speed and compact code more than gcc
<nikolar> Lol
<nikolar> You clearly don't know what "modern" c++ is
<johnjaye> i thought about learning it
<johnjaye> as best i gather there's a different "C++" depending on the decade
<johnjaye> there's pre standard 90s c++. there's standardized c++ in the 2000s. then c++11 in 2010s then whatever we have now
khrbtxyz has joined #osdev
<kof673> i would think opposite, modern stuff is less concerned about time to compile itself. unless something is jitting, compile-time is a cost someone else pays, and the end-user doesn't necessarily see
<kof673> likewise, focussing on the generated code, not the time it takes to compile the compiler
hwpplayer1 has joined #osdev
<kof673> i wouldn't even say modern stuff, any non-jittish thing is going to be like that IMO
<kof673> unless you have some situation you expect hardware to get slower, and number of cores etc. to go down :D
Dead_Bush_Sanpai has quit [Read error: Connection reset by peer]
<kof673> a depression maybe :D
Dead_Bush_Sanpai has joined #osdev
<kof673> or <insert new cpu here that is really nice but too slow ATM>
<Matt|home> hello. im not entirely sure if this belongs here or in ##hardware or something, but im 99% sure this is a kernel-level thing
<Matt|home> hm. .. i should type this into pastebin or something
Arthuria has joined #osdev
<zid`> johnjaye: C++ has n! compilation speed, as you add more C++ features
<Matt|home> https://pastebin.com/4zVAt8vL <-- line 6 and line 19 have my questions, the rest of the text is just explaining why im confused if anyone has any input on the topic. im sorry if it's off-topic, i legit don't know how low level this goes
bslsk05 has quit [Quit: :)]
puck has quit [Remote host closed the connection]
<kof673> > possible to maximize a window to fit across multiple monitors that works fine in unixland, windows won't let you manually resize? or show 1/2 and 1/2 in the middle? everything else is "window manager" for x. there used to be tweakui for windows that let you set "focus on hover [of pointer]" this is just terminology, "I originally thought that "focus" just meant "the window currently receiving input"" i would agree w
<kof673> ith that
zenmov has joined #osdev
<Matt|home> well lemme try
<Matt|home> nah, if i click and drag a window in extended mode between A and B and hit the maximize button it'll only maximize in a given monitor
<Matt|home> not both
<kof673> yes, i am saying manually resize though, maximize may be screwed though :D
<kof673> s/yes/nah/
<Matt|home> i mean i totally can manually resize it, it looks awful but i can
<Matt|home> but yeah im not sure how input is determined here
<kof673> well very briefly i have 2 monitors and different resolutions and different sizes
<Matt|home> oh right, i forgot to mention one of the most important parts rofl
<kof673> so when i use mplayer (the unix thing, but also had windows port, not the ms one) ...to play full screen [across both]....it will chop off part of the picture on one monitor, unless i set both monitors to same resolution first :D
bslsk05 has joined #osdev
<kof673> not the same, just saying those types of issues can happen for whatever reasons, someone else knows more :D
housemate has joined #osdev
<Matt|home> the taskbar on the bottom with the start menu is displayed twice, one on both monitors :p it's not an extension, it seems to suggest the computer is well aware that these are two separate monitors and treats them as such
<kof673> well all of this is going to be os-specific, windowing system specific, window manager specific, etc.
<kof673> and then if you have 2 say separate video cards...how to display video across both of them? that will get into drivers then perhaps
<kof673> not much else to say
<kof673> even openstep had multihead support for certain matrix cards, don't recall specifics though
puck has joined #osdev
<kof673> old mac os i think had it for a long time too, x window for a very long time
<kof673> > with the mouse cursor being able to send a signal to a non-focused window you might find a "macro" program or something that lets you simulate sending pointer events/keystrokes/whatever to a window, i know x has such programs, and windows surely must too, but again, this is all going to be os-specific
<kof673> people who might have to "script" such a thing want a program that does that
<kof673> i think beos let you multihead too.........it has been around a long time :D
<kof673> you should ask in a windows channel though, just saying stuff like that is not impossible
stolen has joined #osdev
hwpplayer1 has quit [Read error: Connection reset by peer]
<kof673> a long time ago i recall a python script that used some kind of opengl x server, to divide the screen into 32 or however many "tiles" for a "wall" of lcd monitors, giant display.....it could do 3d acceleration too i believe, never used it just read about it. there are people who want stuff like that too, a giant wall :D that split things up over the network i believe, to separate machines
\Test_User has quit [Quit: \Test_User]
\Test_User has joined #osdev
housemate has quit [Quit: "I saw it in a tiktok video and thought that it was the most smartest answer ever." ~ AnonOps Radio [some time some place] | I AM THE DERIVATIVE I AM GOING TANGENT TO THE CURVE!]
\Test_User has quit [Ping timeout: 252 seconds]
<kof673> i might be misremembering, x has/had a thing called "xinerama" to join screens IIRC, but before that, you could do multihead, and move cursor across monitors IIRC, but not display a window or move a window across IIRC and you ran 2 separate window managers IIRC, could even run different WMs for each monitor IIRC lol anyways...not sure if that used separate X servers or what, just remember there was a non-xinerema mode that gave l
<kof673> imited multihead
<kof673> it all depends on the windowing system :D
\Test_User has joined #osdev
housemate has joined #osdev
op has quit [Remote host closed the connection]
<kof673> quoth wikipedia, sounds about right, came with xfree86 4.0, so 3.x maybe did not have this :D > When Xinerama is enabled in the X server, multiple X screens can be unified into a single workspace. This unified work area allows windows to be transferred across X screens.
<clever> Xinerama is also an extension that allows an X client to query what coords on the unified screen map to a given monitor
<clever> so something like a window manager knows where to place a window when you maximize
<zid`> I used xinrama on my dual crt setup
<zid`> so that I could have a single :0 spanning two monitors properly, rather than an :0 and a :1
<kof673> wiki even says xinerama might even have "dead spots" so WM needs to be aware :/
<kof673> i guess a funny person would allow custom dead spots like bad ram to purposely not put something on damaged part of a screen :D
<Ermine> vrr says hi
_ngn- has joined #osdev
_ngn has quit [Ping timeout: 260 seconds]
<vin> Hi, does anyone know know how ordering is maintained between RPQ and WPQ in memory controllers? Does every read request first check WPQ and writes check RPQ? Even then there could be ordering problems.
<vin> For example, If there are 20 entries in RPQ and 2 entries in WPQ and in an application we do a load on address A and then a store on address A. In this scenario because write can reach the memory DIMM before read (longer RPQ), wouldn't there be an ordering issue from programs perspective?
bslsk05 is now known as bslsk06
_ngn- is now known as _ngn
<clever> kof673: ive noticed OBS can record the dead spots, and in some configurations, it alternates between 2 corrupt images
<clever> due to just undefined data and over-draw, stuff leaks into the dead spot, but its also double-buffered
<kof673> it was a joke, i just meant in theory, say the bottom part of the screen is screwed up, just ignore those rows, don't use them :D just allow cutting off borders :D
<kof673> "Letter-boxing" "matte" black bars i guess is the term :D
<clever> heh
GeDaMo has quit [Quit: 0wt 0f v0w3ls.]
goliath has quit [Quit: SIGSEGV]
netbsduser` has quit [Ping timeout: 252 seconds]
orccoin has joined #osdev
<SystemPrompt> apparently that old multi-screen mode is retronamed as zaphod mode (zaphod beeblebrox, multiple heads)
obrien has joined #osdev
<kof673> wiki says for multiseat specifically :D but yeah :D
craigo has quit [Quit: Leaving]
Left_Turn has joined #osdev
Arthuria has quit [Ping timeout: 272 seconds]
Turn_Left has joined #osdev
Left_Turn has quit [Ping timeout: 246 seconds]
kneskade has quit [Remote host closed the connection]
<vin> I wrote a program to test my hypothesis but MC dosen't seem to causing any ordering violations: https://ctxt.io/2/AAB4Qo_HEg But I still don't know how it manages to keep it in order though!!
<bslsk06> ​ctxt.io: Context – share whatever you see with others in seconds
troseman has joined #osdev
troseman has quit [Ping timeout: 252 seconds]
troseman has joined #osdev
orccoin has quit [Ping timeout: 252 seconds]
m3a has quit [Remote host closed the connection]
stolen has quit [Quit: Connection closed for inactivity]
Turn_Left has quit [Read error: Connection reset by peer]
xenos1984 has quit [Quit: Leaving.]
xenos1984 has joined #osdev
hwpplayer1 has joined #osdev