]>
Devi Nivas Git - cs3210-lab1.git/log
Robert Morris [Wed, 31 Aug 2011 00:50:19 +0000 (20:50 -0400)]
enterpgdir -> entrypgdir
Frans Kaashoek [Mon, 29 Aug 2011 21:18:40 +0000 (17:18 -0400)]
Revert "Introduce and use sleeplocks instead of BUSY flags"
My changes have a race with re-used bufs and the code doesn't seem to get shorter
Keep the changes that fixed ip->off race
This reverts commit
3a5fa7ed9020eaf8ab843a16d26db7393b2ec072 .
Conflicts:
defs.h
file.c
file.h
Austin Clements [Mon, 29 Aug 2011 20:14:44 +0000 (16:14 -0400)]
Renumber I_VALID since I_BUSY is gone
Austin Clements [Mon, 29 Aug 2011 20:12:17 +0000 (16:12 -0400)]
dir-locals.el to set style in Emacs
Austin Clements [Mon, 29 Aug 2011 20:12:01 +0000 (16:12 -0400)]
Style nits; indentation and tabs
Frans Kaashoek [Fri, 26 Aug 2011 14:10:37 +0000 (10:10 -0400)]
Merge branch 'master' of git+ssh://amsterdam.csail.mit.edu/home/am0/6.828/xv6
Frans Kaashoek [Fri, 26 Aug 2011 14:08:29 +0000 (10:08 -0400)]
Introduce and use sleeplocks instead of BUSY flags
Remove I_BUSY, B_BUSY, and intrans defs and usages
One spinlock per buf to avoid ugly loop in bget
fix race in filewrite (don't update f->off after releasing lock)
Robert Morris [Fri, 26 Aug 2011 10:47:13 +0000 (06:47 -0400)]
segment register -> segment descriptor
Frans Kaashoek [Thu, 25 Aug 2011 18:39:09 +0000 (14:39 -0400)]
Make cross reference with text work
Frans Kaashoek [Wed, 24 Aug 2011 14:24:40 +0000 (10:24 -0400)]
Eliminate USERTOP; no use for it.
Frans Kaashoek [Tue, 23 Aug 2011 00:07:18 +0000 (20:07 -0400)]
A comment
Frans Kaashoek [Tue, 23 Aug 2011 00:05:15 +0000 (20:05 -0400)]
Remove sys_init syscall
Invoke initlog from forkret on first user process
Frans Kaashoek [Mon, 22 Aug 2011 01:14:42 +0000 (21:14 -0400)]
Merge branch 'master' of git+ssh://amsterdam.csail.mit.edu/home/am0/6.828/xv6
Frans Kaashoek [Mon, 22 Aug 2011 01:14:29 +0000 (21:14 -0400)]
delete dead code
Robert Morris [Fri, 19 Aug 2011 17:30:57 +0000 (13:30 -0400)]
make dirlookup and dirlink more similar
Frans Kaashoek [Thu, 18 Aug 2011 00:52:28 +0000 (20:52 -0400)]
bootloader reads into a physical address (not a virtual address)
Frans Kaashoek [Thu, 18 Aug 2011 00:23:36 +0000 (20:23 -0400)]
Make elf proghdr fields match what objdump prints
Frans Kaashoek [Wed, 17 Aug 2011 00:23:17 +0000 (20:23 -0400)]
Switch back to #define for PHYSTOP. Using the E820 to retrieve the memory map is too complicated (must be done in 16-bit real-mode, probably enlarged bootblock beyond 512 bytes, and a #define requires less explanation).
Frans Kaashoek [Tue, 16 Aug 2011 19:47:22 +0000 (15:47 -0400)]
Clean up memlayout.h
Get rid of last instances of linear address and "la"
Get ready for detecting physical memory dynamically
Frans Kaashoek [Tue, 16 Aug 2011 00:32:04 +0000 (20:32 -0400)]
Update make clean
Frans Kaashoek [Tue, 16 Aug 2011 00:28:28 +0000 (20:28 -0400)]
Update list of files to include in printout
Frans Kaashoek [Tue, 16 Aug 2011 00:21:14 +0000 (20:21 -0400)]
Make Austin happier
2011
Frans Kaashoek [Tue, 16 Aug 2011 00:11:13 +0000 (20:11 -0400)]
Avoid "boot" in xv6
Frans Kaashoek [Mon, 15 Aug 2011 21:41:58 +0000 (17:41 -0400)]
Use 4Mbyte pages during boot
Austin Clements [Mon, 15 Aug 2011 18:11:22 +0000 (14:11 -0400)]
Only commit log to disk if something was written in to it.
Without this, every write to the console results in log writes.
Robert Morris [Mon, 15 Aug 2011 16:44:41 +0000 (12:44 -0400)]
Merge branch 'master' of git+ssh://amsterdam.csail.mit.edu/home/am0/6.828/xv6
Robert Morris [Mon, 15 Aug 2011 16:44:20 +0000 (12:44 -0400)]
avoid deadlock by calling begin_trans() before locking any inodes
Frans Kaashoek [Mon, 15 Aug 2011 16:02:59 +0000 (12:02 -0400)]
Separate more clearly bootloader from xv6 by renaming multiboot.S to entry.S etc.
Maybe the string boot shouldn't appear in xv6 code?
Frans Kaashoek [Fri, 12 Aug 2011 16:02:17 +0000 (12:02 -0400)]
Oops
Robert Morris [Fri, 12 Aug 2011 13:26:23 +0000 (09:26 -0400)]
Merge branch 'master' of git+ssh://amsterdam.csail.mit.edu/home/am0/6.828/xv6
Robert Morris [Fri, 12 Aug 2011 13:25:39 +0000 (09:25 -0400)]
log write() data
usertest for big write()s
push begin_trans/commit_trans down into syscalls
Frans Kaashoek [Fri, 12 Aug 2011 11:31:52 +0000 (07:31 -0400)]
Some comments
Frans Kaashoek [Thu, 11 Aug 2011 16:25:10 +0000 (12:25 -0400)]
Make AP processors boot using bootpgdir
Remove device mapping from bootpgdir
Remove unnecessary vmenable
Set CPUS back to 2 in Makefile
Passes all usertests
Frans Kaashoek [Wed, 10 Aug 2011 03:22:48 +0000 (23:22 -0400)]
One more low address
Frans Kaashoek [Wed, 10 Aug 2011 01:56:43 +0000 (21:56 -0400)]
Tiny bit of cleanup
Frans Kaashoek [Wed, 10 Aug 2011 01:51:17 +0000 (21:51 -0400)]
Shorten boot assembly
Frans Kaashoek [Wed, 10 Aug 2011 01:49:13 +0000 (21:49 -0400)]
x
Frans Kaashoek [Wed, 10 Aug 2011 01:47:40 +0000 (21:47 -0400)]
Remove jmpkstack
Frans Kaashoek [Wed, 10 Aug 2011 01:37:35 +0000 (21:37 -0400)]
Use static page table for boot, mapping first 4Mbyte; no more segment trick
Allocate proper kernel page table immediately in main using boot allocator
Remove pginit
Simplify address space layout a tiny bit
More to come (e.g., superpages to simplify static table)
Frans Kaashoek [Wed, 10 Aug 2011 01:33:59 +0000 (21:33 -0400)]
Use kernel virtual addresses for BIOS memory, etc.
Frans Kaashoek [Tue, 9 Aug 2011 20:09:15 +0000 (16:09 -0400)]
Allow gas to use / as divide (instead of comment)
Maybe not necessary, but don't want to waste time figuring this out again
Frans Kaashoek [Mon, 8 Aug 2011 17:30:08 +0000 (13:30 -0400)]
One definition of several macros and constants
Frans Kaashoek [Mon, 8 Aug 2011 13:20:29 +0000 (09:20 -0400)]
Speedup sbrk tst a bit (forking 100Mbyte processes is slow)
Frans Kaashoek [Mon, 8 Aug 2011 03:03:48 +0000 (23:03 -0400)]
Adjust sbrk test for large user address spaces
All tests pass
Frans Kaashoek [Sun, 7 Aug 2011 16:30:34 +0000 (12:30 -0400)]
oops
Frans Kaashoek [Mon, 1 Aug 2011 01:27:02 +0000 (21:27 -0400)]
Don't map IO space in the user part of the address space
Passes all tests now (but need to update usertests to allow for more than 640k)
Frans Kaashoek [Fri, 29 Jul 2011 11:31:27 +0000 (07:31 -0400)]
Map kernel high
Very important to give qemu memory through PHYSTOP :(
Frans Kaashoek [Thu, 28 Jul 2011 00:49:45 +0000 (20:49 -0400)]
Merge commit 'origin/master' into page
Frans Kaashoek [Thu, 28 Jul 2011 00:35:46 +0000 (20:35 -0400)]
Dirt simple logging
Passes usertests and stressfs
Seems to recover correctly in a number of simple cases
Austin Clements [Fri, 24 Jun 2011 18:20:51 +0000 (14:20 -0400)]
Add -fno-omit-frame-pointer to compile flags.
At some point between gcc 4.4 and 4.6, omitting the frame pointer
became default if any -O is specified. getcallerpcs relies on the
frame pointer, so bring it back.
Nickolai Zeldovich [Fri, 24 Jun 2011 09:26:57 +0000 (05:26 -0400)]
do not run past the end of the disk!
[ cherry-picked
de6075ff0978108daea1d658425bcb37663a98a0 ]
Conflicts:
fs.c
Frans Kaashoek [Thu, 28 Apr 2011 12:01:36 +0000 (08:01 -0400)]
Russ's fix for panic remap
Russ Cox [Mon, 28 Feb 2011 14:32:54 +0000 (09:32 -0500)]
rebuild xv6-rev5.tar.gz
does not affect line numbers
Russ Cox [Mon, 28 Feb 2011 14:32:20 +0000 (09:32 -0500)]
fixes from Peter Froehlich
Russ Cox [Sun, 20 Feb 2011 02:21:49 +0000 (21:21 -0500)]
web: rev5
Russ Cox [Sun, 20 Feb 2011 02:19:58 +0000 (21:19 -0500)]
updated book pdfs
Russ Cox [Sun, 20 Feb 2011 02:17:55 +0000 (21:17 -0500)]
xv6: formatting, cleanup, rev5 (take 2)
Russ Cox [Sun, 20 Feb 2011 02:16:54 +0000 (21:16 -0500)]
xv6: formatting, cleanup, rev5
Russ Cox [Sat, 19 Feb 2011 15:36:14 +0000 (10:36 -0500)]
xv6: refresh from 2010 page
Russ Cox [Sat, 19 Feb 2011 15:23:19 +0000 (10:23 -0500)]
xv6: link to xv6-rev4
Russ Cox [Tue, 11 Jan 2011 18:54:23 +0000 (13:54 -0500)]
missing file memide.c
Russ Cox [Tue, 11 Jan 2011 18:51:40 +0000 (13:51 -0500)]
more trivial cleanup
Russ Cox [Tue, 11 Jan 2011 18:27:45 +0000 (13:27 -0500)]
multiboot support and memory-only (no disk) kernel
Russ Cox [Tue, 11 Jan 2011 18:16:28 +0000 (13:16 -0500)]
mp: do not panic on surprising hw
Russ Cox [Tue, 11 Jan 2011 18:01:13 +0000 (13:01 -0500)]
make new code like old code
Variable declarations at top of function,
separate from initialization.
Use == 0 instead of ! for checking pointers.
Consistent spacing around {, *, casts.
Declare 0-parameter functions as (void) not ().
Integer valued functions return -1 on failure, 0 on success.
Russ Cox [Mon, 10 Jan 2011 15:43:40 +0000 (10:43 -0500)]
mkfs: avoid clash with system struct stat
Russ Cox [Sun, 2 Jan 2011 22:59:57 +0000 (17:59 -0500)]
mkfs: avoid out of bounds access to sb in wsect
Robert Morris [Wed, 29 Sep 2010 18:12:26 +0000 (14:12 -0400)]
even more fabulous exec
Robert Morris [Mon, 27 Sep 2010 20:17:57 +0000 (16:17 -0400)]
check exec() arg length
fix double iunlockput
Robert Morris [Mon, 27 Sep 2010 20:14:33 +0000 (16:14 -0400)]
copyout() copies data to a va in a pagetable, for exec() &c
usertest that passes too many arguments, break exec
Robert Morris [Mon, 20 Sep 2010 10:00:22 +0000 (06:00 -0400)]
map kernel instructions r/o
Robert Morris [Sun, 19 Sep 2010 17:48:22 +0000 (13:48 -0400)]
Merge branch 'master' of git+ssh://amsterdam.csail.mit.edu/home/am0/6.828/xv6
Robert Morris [Sun, 19 Sep 2010 17:47:52 +0000 (13:47 -0400)]
my comment is wrong, exec handles BSS fine
Robert Morris [Sun, 19 Sep 2010 11:18:42 +0000 (07:18 -0400)]
exec questions
Austin Clements [Wed, 15 Sep 2010 20:15:07 +0000 (16:15 -0400)]
Simplify boot loader by removing some JOS'isms
Robert Morris [Mon, 13 Sep 2010 19:34:44 +0000 (15:34 -0400)]
change some comments, maybe more informative
delete most comments from bootother.S (since copy of bootasm.S)
ksegment() -> seginit()
move more stuff from main() to mainc()
Austin Clements [Fri, 3 Sep 2010 19:47:28 +0000 (15:47 -0400)]
Rev 4
Austin Clements [Thu, 2 Sep 2010 23:03:48 +0000 (19:03 -0400)]
Update example in TOC footer
Austin Clements [Thu, 2 Sep 2010 23:01:25 +0000 (19:01 -0400)]
Consistent style in defs.h
Austin Clements [Thu, 2 Sep 2010 22:54:50 +0000 (18:54 -0400)]
Get consoleintr back on one page
Austin Clements [Thu, 2 Sep 2010 22:50:49 +0000 (18:50 -0400)]
Get cgaputc and consputc back on the same page
Austin Clements [Thu, 2 Sep 2010 22:36:06 +0000 (18:36 -0400)]
Get rid of the blank page in picirq.c. It misaligs things now.
Austin Clements [Thu, 2 Sep 2010 22:28:36 +0000 (18:28 -0400)]
Simplify allocuvm/deallocuvm to operate in a contiguous memory model. This makes their interface match up better with proc->sz and also simplifies the callers (it even gets the main body of exec on one page).
Austin Clements [Thu, 2 Sep 2010 21:03:19 +0000 (17:03 -0400)]
Nits
Austin Clements [Thu, 2 Sep 2010 20:39:55 +0000 (16:39 -0400)]
Comment inituvm and loaduvm
Austin Clements [Thu, 2 Sep 2010 20:36:38 +0000 (16:36 -0400)]
Fix comment capitalization
Austin Clements [Thu, 2 Sep 2010 20:23:15 +0000 (16:23 -0400)]
Rearrange vm.c so it's in logical order and prints nicely. Shorten a few functions in uninteresting ways to make them fit.
Austin Clements [Thu, 2 Sep 2010 19:37:05 +0000 (15:37 -0400)]
Simplify inituvm and userinit by assuming initcode fits on a page
Austin Clements [Thu, 2 Sep 2010 19:18:19 +0000 (15:18 -0400)]
Simplify freevm using deallocuvm
Austin Clements [Thu, 2 Sep 2010 18:30:06 +0000 (14:30 -0400)]
Oops. Broke the build when I rearranged proc.c
Austin Clements [Thu, 2 Sep 2010 18:08:45 +0000 (14:08 -0400)]
Remove unused nop_pause function.
Austin Clements [Thu, 2 Sep 2010 08:15:17 +0000 (04:15 -0400)]
Rearrange proc.h and proc.c to get our action-packed spreads back (mostly). They also make sense in this order, so it's not just for page layout.
Austin Clements [Thu, 2 Sep 2010 08:06:35 +0000 (04:06 -0400)]
Re-re-arrange bootasm and bootother. We can't get bootasm and bootmain on the same page, but this will make them easier to flip between.
Austin Clements [Thu, 2 Sep 2010 08:03:42 +0000 (04:03 -0400)]
I totally misunderstood the even/odd spec. Add some docs, remove my 'left+' hack, and orthogonalize left/right versus even/odd so you can specify them independently.
Austin Clements [Thu, 2 Sep 2010 07:57:20 +0000 (03:57 -0400)]
Specify in runoff.spec which side sheet 1 is on, since it keeps changing
Austin Clements [Thu, 2 Sep 2010 07:56:36 +0000 (03:56 -0400)]
Pack five columns into cross-ref. This saves us an entire page and fixes the single dangling line.
Austin Clements [Thu, 2 Sep 2010 07:36:39 +0000 (03:36 -0400)]
Oops, missed a comment about the stack guard
Austin Clements [Thu, 2 Sep 2010 07:25:33 +0000 (03:25 -0400)]
Add Nickolai and myself and scrunch two lines out of the text elsewhere
Austin Clements [Wed, 1 Sep 2010 21:14:58 +0000 (17:14 -0400)]
Slight simplification of copyuvm. We could simplify other things now that processes are contiguous, but we'd have to think harder about the error paths.
Austin Clements [Wed, 1 Sep 2010 20:46:37 +0000 (16:46 -0400)]
Remove the stack guard page. Processes are now contiguous from 0 to proc->sz, which means our syscall argument validation is correct. Add a pointer validation test and remove the stack test, which tested for the guard page.