<rqou>
blargh after enough time firefox will just get wedged to the point where all media won't play
digshadow has joined ##openfpga
<azonenberg>
rqou: i have to restart my FF ~weekly
<azonenberg>
and keep it in a VM so it wont keep growing in ram usage until it bogs the whole machine down
<azonenberg>
my fb vm has only 2GB of RAM for example
<azonenberg>
and if i'm away from home for a couple days when i come back i typically find that it got oom-killed
<azonenberg>
well, social media vm
<qu1j0t3>
that's a p. good strategy but sad that it does this
<azonenberg>
not sure if fb, twitter, or other stuff is the biggest offender
<azonenberg>
I also virtualize for damage control if it gets pwned
<qu1j0t3>
twitter seems to be pretty bad
<qu1j0t3>
i avoid leaving that open for long
<azonenberg>
social media vm is a big offender
<azonenberg>
general browsing is next most
<azonenberg>
even pidgin bogs down after 2-4 months running
<azonenberg>
my mail client never gets restarted except for security patches
<azonenberg>
actually i might restart pidgin more like every month
<azonenberg>
now that i think about it
<rqou>
i'm waiting for firefox to finish their e10s
<azonenberg>
but its certainly way less than a week
<rqou>
then you can have one process per tab
<rqou>
and you can probably ulimit each one individually
<azonenberg>
now that would be nice
<azonenberg>
but the tabs will still leak horribly
<azonenberg>
you'll just restart the tab instead of the browser
<azonenberg>
improvement, i guess :p
<rqou>
then it ooms itself
<rqou>
although apparently on linux you can only ulimit virtual address size
<rqou>
not total committed ram
<azonenberg>
thats annoying
<azonenberg>
i dont care how many files you mmap
<azonenberg>
i care how much ram you're blocking other stuff from using
<rqou>
but CoW makes that complicated
<azonenberg>
(fwiw quotas in antikernel will be based on pages)
<rqou>
(and introduces root exploits :P)
<azonenberg>
And the lack of CoW or shared memory in general will make that easy to implement
<azonenberg>
every page has one and only one owner
<rqou>
that's not efficient
<azonenberg>
even free pages (owned by the dram controller)
<azonenberg>
Depends on your use case
<azonenberg>
for desktop computing? no
<azonenberg>
But i'm targeting deeply embedded
<rqou>
ah works fine there
<azonenberg>
Where you're not going to have 30 apps running with their own copy of qt
<rqou>
in general imho linux desktop really needs to start using features the kernel already has
<rqou>
e.g. ulimit, seccomp
<rqou>
namespaces
<rqou>
hell, even selinux/other MAC frameworks
<azonenberg>
yeah that would be nice
carl0s has joined ##openfpga
<carl0s>
pointfree, eric_j , silly question, do you know if it's possible to know which PLD (0 or 1?, left or right?) was used based on the information on the .rpt file? So far i found the # of the macrocell.
<pointfree>
carl0s: In the .route file you can see if pld0 or pld1 are mentioned.
cosmobird has quit [Ping timeout: 256 seconds]
<pointfree>
I don't know where that information is in rpt file, but from the Cypress PSoC Creator Manual it looks like you can force placement to PLD A or B.
<flaviusb>
azonenberg: Is there a single concept of 'a process' in Antikernel? Like, what is your unit of 'thingness'?
<flaviusb>
Because it seems like 'node' or 'address' or whatever is the closest thing to that. Is that the case?
<carl0s>
hehe sorry, just saw it. on the report file is says UDB(0,4):pld0:mc0, on the rpt file i have [UDB(0,4)][LB=0][MC=0] so i guess LB is the number of the pld?
<flaviusb>
(I am assuming that 'single owner' means 'single name/node has ownership', but I would be very interested if you mean something else)
<carl0s>
pointfree, on the rpt file there'a a tag "Final Placement Summary" then another tag inside "Final Placement Details" and then "Component Details"
<pointfree>
carl0s: That would make sense. I suppose it would change when placement is constrained to the other PLD.
<pointfree>
carl0s: interesting.
<pointfree>
carl0s: Can placement be forced to a specific input term of the PLD? It looks like placement can be forced to a specific macrocell.
<carl0s>
i think so, i'm trying to find the name on the .route file
<carl0s>
the XOR gate name is xor_1, so i think the name field on the directives tab must be \xor_1:pld0:mc0\ or simmilar, i'm testing it with a different mc#, but my pc is slow and takes ages to compile ;(
DocScrutinizer05 has quit [Disconnected by services]
DocScrutinizer05 has joined ##openfpga
tecepe has quit [Ping timeout: 245 seconds]
maaku has quit [Quit: No Ping reply in 180 seconds.]
maaku has joined ##openfpga
<carl0s>
pointfree, there's a whole section explaining the Directives, is starts on page 327 of the Creator User Manual, didn't see it b4, and yes it is possible to specify the macrocell within a specific PLD on a specific UDB :). Format is U(x, y, l)i where x is the row, y the colum of the UDB, l is the PLD descriptor (A or B) and i is the macrocell index, and it must be assigned to a output signal from the set of random logic you want to locate.
discrttm is now known as nonlinear
kuldeep has quit [Read error: Connection reset by peer]
kuldeep has joined ##openfpga
carl0s has quit [Quit: Leaving]
<azonenberg>
flaviusb: Nodes are a superset of processes
<azonenberg>
a process running on a CPU is a node
<azonenberg>
as is the management interface on the CPU
<azonenberg>
or a hardware IP core
<azonenberg>
They serve the same function as processes do in a conventional OS... they can own handles, be sources or destinations for IPC, etc
<azonenberg>
flaviusb: also there's a chan specifically for antikernel
<azonenberg>
#antikernel
<azonenberg>
probably a better fit there than here
<azonenberg>
(esp considering antikernel is currently being developed with mostly proprietary FPGA tools, none of the open tools are mature enough to build it, or target devices that are too small)
Bike has quit [Quit: ten hours in the desert]
kuldeep has quit [Ping timeout: 260 seconds]
digshadow has quit [Quit: Leaving.]
eric_j has quit [Remote host closed the connection]
digshadow has joined ##openfpga
_whitelogger has joined ##openfpga
kuldeep has joined ##openfpga
amclain has quit [Quit: Leaving]
<flaviusb>
Cool.
kuldeep has quit [Quit: Leaving]
cosmobird has joined ##openfpga
scrts has quit [Ping timeout: 244 seconds]
cosmobird has quit [Ping timeout: 268 seconds]
scrts has joined ##openfpga
openfpga-bb has quit [Ping timeout: 260 seconds]
azonenberg has quit [Ping timeout: 258 seconds]
azonenberg_work has quit [Ping timeout: 265 seconds]
carl0s has joined ##openfpga
carl0s has quit [Client Quit]
scrts has quit [Ping timeout: 250 seconds]
scrts has joined ##openfpga
kuldeep has joined ##openfpga
azonenberg has joined ##openfpga
openfpga-bb has joined ##openfpga
kuldeep has quit [Read error: Connection reset by peer]
azonenberg_work has joined ##openfpga
cosmobird has joined ##openfpga
mzpx has joined ##openfpga
cosmobird has quit [Ping timeout: 256 seconds]
mzpx has quit [Read error: Connection reset by peer]
cosmobird has joined ##openfpga
cosmobird has quit [Ping timeout: 260 seconds]
cosmobird has joined ##openfpga
carl0s has joined ##openfpga
carl0s has quit [Quit: Leaving]
cosmobird has quit [Read error: Connection reset by peer]