 tarfeef101
|
c6987a2215
trying to add vm into the memory init so we never do physical allocation. not working rn though, can't run more than a couple tests sequentially without a tlb miss (Accessing NULL)
|
6 rokov pred |
|
tarfeef101
|
a4fb607edb
there were a few cases of getppages getting called directly, which caused issues after vm_init is true, as that sets first and last to 0 in ram.c, which is used by stealmem. so those would fail after init. hence the issue. i simply changed those to use alloc_kpages, so now there should be no exceptions. ran one set of tests, but those pass now
|
6 rokov pred |
|
tarfeef101
|
ee5567e1fd
got memory management working (no paging). tested. need to do paging, then try cool shit like VM in ram.c for never using physical allocation. i'm probably forgetting something, but that's all i can remember rn
|
6 rokov pred |
|
tarfeef101
|
e22dd1c049
tlb now randomly kickes entries, code segments are read only
|
6 rokov pred |
|
tarfeef101
|
ca95959290
new branch for a3
|
6 rokov pred |
 Tareef Dedhar
|
9b4c9e283d
Merge branch 'a2b' of tarfeef101/cs350 into master
|
6 rokov pred |
 tarfeef101
|
43b65fa2e9
Merge branch 'a2b' of ssh://git.tareef.tech:10022/tarfeef101/cs350 into a2b
|
6 rokov pred |
|
tarfeef101
|
bd91fe35d2
modified buildscripts to be working directory agnostic, also fixed a small memory leak
|
6 rokov pred |
|
Tareef Dedhar
|
a546dc9ed7
Merge branch 'a2b' of tarfeef101/cs350 into master
|
6 rokov pred |
|
tarfeef101
|
7376c59940
testing with cloudflare dns crap going on
|
6 rokov pred |
|
tarfeef101
|
23af2b9c19
leslie is a god
|
6 rokov pred |
|
tarfeef101
|
2c7d850820
debugging of my weird error. details are on piazza, and can be inferred through comments and printfs when running uw-testbin/argtesttest
|
6 rokov pred |
|
tarfeef101
|
74156bfff5
confirmed, bug with argtesttest is with a2b, not a2a (assuming a2a is actually fixed, of course)
|
6 rokov pred |
|
tarfeef101
|
55542e5358
fixed a2a crap
|
6 rokov pred |
|
tarfeef101
|
4a78e046da
fuck piazza
|
6 rokov pred |
|
tarfeef101
|
11fc2beeac
got execv mostly working... arg2 and argc are kinda off with argtestest
|
6 rokov pred |
|
tarfeef101
|
8b37d7e1e7
tested uw-testbin/argtest with argument passing (no execv), and it works flawlessly. so issue must be with execv specific code as discussed before, but now we have also validated the functionality of the argument copying
|
6 rokov pred |
|
tarfeef101
|
03416ba12a
seems that arg passing may be working after all, but the bad memory reference i get on hogparty is simply due to the execv base functionality not working. hogparty isn't supposed to use argv. perhaps i am not setting up the userspace pointers properly...
|
6 rokov pred |
|
tarfeef101
|
b9192e35bf
believe argument passing in runprogram working, have base execv without agrs working. with args is still not great
|
6 rokov pred |
|
tarfeef101
|
4c3717c727
have a building kernel, can run old tests, but cannot run new ones. execv clearly needs work
|
6 rokov pred |
|
tarfeef101
|
8f2d17547d
started execv, not sure how caret.proj will do stuff, guess i'll find out when i repoen it.
|
6 rokov pred |
|
tarfeef101
|
ab2a8632a3
new branch for a2b
|
6 rokov pred |
|
Tareef Dedhar
|
9b74ba733d
Merge branch 'a2a' of tarfeef101/cs350 into master
|
6 rokov pred |
|
tarfeef101
|
a76a6813f0
rebuilt user level programs as per guidelines pre-submission
|
6 rokov pred |
|
tarfeef101
|
a680c9cb45
removing old code, missed 1 element of the processes array in init
|
6 rokov pred |
|
tarfeef101
|
41d0d447bd
added a script for non-complete building. also, fixed sys exit. issue was not deleting children that are zombies when parents die. i do that now
|
6 rokov pred |
|
tarfeef101
|
2c8fab64f4
using the proper palin test, i am passing. think my sleep is the issue, or according to piazza the sleep is exposing issues with address spaces which could be a fork issue.
|
6 rokov pred |
|
tarfeef101
|
42307f5601
using proclock more
|
6 rokov pred |
|
tarfeef101
|
a982f748c5
actually update retval in getpid
|
6 rokov pred |
|
tarfeef101
|
0e648f4fd6
added another condition to thread to allow for threads to die if their process has exited but exists because it wants to be usable by family|
|
6 rokov pred |
