summaryrefslogtreecommitdiff
path: root/vm.c
AgeCommit message (Collapse)AuthorFilesLines
2018-10-09Change mycpu() to use %gs.Frans Kaashoek1-2/+5
2018-10-04Make names line up with textFrans Kaashoek1-9/+9
2018-10-03Names of text are better.Frans Kaashoek1-5/+5
2018-10-02Remove unused typedefsFrans Kaashoek1-5/+5
2018-10-02Simplify by freeing user part of addres pace in one page increments. ThisFrans Kaashoek1-2/+2
undoes commit ffe444 and 052e18, which skipped page directories, but was tailored to two-level page table. Undoing doesn't seem to affect boottime for xv6 visibly.
2018-10-02Simplify freeing page tableFrans Kaashoek1-27/+23
2018-10-02Avoid repition in walkpgdirFrans Kaashoek1-45/+13
2018-09-23Checkpoint port of xv6 to x86-64. Passed usertests on 2 processors a few times.Frans Kaashoek1-63/+150
The x86-64 doesn't just add two levels to page tables to support 64 bit addresses, but is a different processor. For example, calling conventions, system calls, and segmentation are different from 32-bit x86. Segmentation is basically gone, but gs/fs in combination with MSRs can be used to hold a per-core pointer. In general, x86-64 is more straightforward than 32-bit x86. The port uses code from sv6 and the xv6 "rsc-amd64" branch. A summary of the changes is as follows: - Booting: switch to grub instead of xv6's bootloader (pass -kernel to qemu), because xv6's boot loader doesn't understand 64bit ELF files. And, we don't care anymore about booting. - Makefile: use -m64 instead of -m32 flag for gcc, delete boot loader, xv6.img, bochs, and memfs. For now dont' use -O2, since usertests with -O2 is bigger than MAXFILE! - Update gdb.tmpl to be for i386 or x86-64 - Console/printf: use stdarg.h and treat 64-bit addresses different from ints (32-bit) - Update elfhdr to be 64 bit - entry.S/entryother.S: add code to switch to 64-bit mode: build a simple page table in 32-bit mode before switching to 64-bit mode, share code for entering boot processor and APs, and tweak boot gdt. The boot gdt is the gdt that the kernel proper also uses. (In 64-bit mode, the gdt/segmentation and task state mostly disappear.) - exec.c: fix passing argv (64-bit now instead of 32-bit). - initcode.c: use syscall instead of int. - kernel.ld: load kernel very high, in top terabyte. 64 bits is a lot of address space! - proc.c: initial return is through new syscall path instead of trapret. - proc.h: update struct cpu to have some scratch space since syscall saves less state than int, update struct context to reflect x86-64 calling conventions. - swtch: simplify for x86-64 calling conventions. - syscall: add fetcharg to handle x86-64 calling convetions (6 arguments are passed through registers), and fetchaddr to read a 64-bit value from user space. - sysfile: update to handle pointers from user space (e.g., sys_exec), which are 64 bits. - trap.c: no special trap vector for sys calls, because x86-64 has a different plan for system calls. - trapasm: one plan for syscalls and one plan for traps (interrupt and exceptions). On x86-64, the kernel is responsible for switching user/kernel stacks. To do, xv6 keeps some scratch space in the cpu structure, and uses MSR GS_KERN_BASE to point to the core's cpu structure (using swapgs). - types.h: add uint64, and change pde_t to uint64 - usertests: exit() when fork fails, which helped in tracking down one of the bugs in the switch from 32-bit to 64-bit - vectors: update to make them 64 bits - vm.c: use bootgdt in kernel too, program MSRs for syscalls and core-local state (for swapgs), walk 4 levels in walkpgdir, add DEVSPACETOP, use task segment to set kernel stack for interrupts (but simpler than in 32-bit mode), add an extra argument to freevm (size of user part of address space) to avoid checking all entries till KERNBASE (there are MANY TB before the top 1TB). - x86: update trapframe to have 64-bit entries, which is what the processor pushes on syscalls and traps. simplify lgdt and lidt, using struct desctr, which needs the gcc directives packed and aligned. TODO: - use int32 instead of int? - simplify curproc(). xv6 has per-cpu state again, but this time it must have it. - avoid repetition in walkpgdir - fix validateint() in usertests.c - fix bugs (e.g., observed one a case of entering kernel with invalid gs or proc
2018-08-30thanks matanessFrans Kaashoek1-1/+3
2017-08-29fix runoff complaints about pagination and long linesRobert Morris1-1/+2
2017-08-07Merge branch 'nogs'Frans Kaashoek1-15/+6
2017-02-05Free pgdir in setupkvm in case of mappages failureSaarett1-1/+3
There is a potential memory leak when mappages() fails inside setupkvm(). A call to freevm() is added in this case so as to reclaim the lost mapping pages.
2017-02-01Eliminate code for gs trick to track per-cpu state. We rely on lapiccpunum()Frans Kaashoek1-6/+1
to find a per-cpu id with which we locate a cpu's cpu struct.
2017-01-31Read curproc from cpu structure, but be careful because after a schedule eventFrans Kaashoek1-3/+1
myproc() points to a different thread. myproc(); sched(); myproc(); // this proc maybe different than the one before sched Thus, in a function that operates on one thread better to retrieve the current process once at the start of the function.
2017-01-31Start of an experiment to remove the use of gs for cpu local variables.Frans Kaashoek1-12/+10
2017-01-30Fix long-standing switchuvm() inconsistency.Peter Froehlich1-3/+8
switchuvm() is supposed to switch the TSS and page table to the process p it is passed. Alas, instead of using p to access the kstack field, it used the global proc. This worked fine because (a) most uses of switchuvm() pass proc anyway and (b) because in the schedule, where we call switchuvm with the newly scheduled process, we actually set the global proc before the call. But I think it's still a bug, even if it never broke a test case. :-)
2016-09-26fix possible memory leak in deallocuvmCody Cutler1-1/+1
when a zero PDE is encountered while searching for present PTEs to free, resume searching at first entry of the next page table instead of the current entry of the next page table.
2016-09-01curproc -> proc in a commentRobert Morris1-1/+1
2016-08-26better commentFrans Kaashoek1-1/+3
2016-08-26set iomb to forbid i/o instructions from user spaceFrans Kaashoek1-0/+1
add to test that they indeed trap in user space thanks to pablogventura@gmail.com and nwolovick@gmail.com
2016-08-25Remove trailing white space with:Frans Kaashoek1-7/+7
for f in *.{h,c}; do sed -i .sed 's/[[:blank:]]*$//' $f; done (Thanks to Nicolás Wolovick)
2016-08-24p2v -> P2VRobert Morris1-13/+13
2016-08-11nothing muchRobert Morris1-1/+1
2016-08-10tiny clarifications to some gdt code.Robert Morris1-2/+1
2016-08-08shell nitsRobert Morris1-1/+6
fix Regehr complaint
2014-08-29Nice layout for a printout of the codeFrans Kaashoek1-0/+8
2013-03-04Correct a security bug in copyuvm()Stephen Tu1-2/+3
copyuvm() should not allow new copied pages to inherit more permissions than the original pages.
2012-08-28nitsRobert Morris1-4/+4
2012-08-22Runoff still complaints these lines are too long, but it is now ok on printoutFrans Kaashoek1-4/+4
2012-08-22Remove unused argument to setupkvm (thanks to Peter Froehlich)Frans Kaashoek1-1/+1
2011-09-13eliminate enter_alloc -- use kalloc for everythingRobert Morris1-13/+12
2011-09-13more regular kmap[] and descriptionRobert Morris1-19/+22
2011-09-02Nuke data.S, since we do have a custom linker script.Austin Clements1-1/+1
2011-09-02Oops, dropped a parenAustin Clements1-1/+1
2011-09-02clear_pte_u -> clearpteuAustin Clements1-2/+2
2011-09-02Move copy_pte_u; fits vm.c in one less column and aligns proc.hAustin Clements1-13/+13
2011-09-02Cleanup comments and fit setupkvm on same page as kmap, which aligns lots of ↵Austin Clements1-10/+6
other things
2011-09-02StyleAustin Clements1-7/+7
2011-09-02Make mappages fit on the pageAustin Clements1-2/+1
2011-09-01Skip missing page directories in deallocuvmAustin Clements1-1/+3
Previously, deallocuvm scanned from 0 to KERNBASE in one page increments, which had a noticable effect on boot time. Now it skips over missing page directories.
2011-09-01inaccessible page under the user stack page, to help exec deal w/ too-large argsRobert Morris1-0/+13
2011-09-01Merge branch 'master' of git+ssh://amsterdam.csail.mit.edu/home/am0/6.828/xv6Robert Morris1-15/+17
Conflicts: vm.c
2011-09-01fix usertests to correctly test what happens when you callRobert Morris1-2/+3
exec() with arguments that don't fit on a single page.
2011-09-01Fix layout issues for printed versionFrans Kaashoek1-16/+19
2011-08-29Style nits; indentation and tabsAustin Clements1-1/+1
2011-08-24Eliminate USERTOP; no use for it.Frans Kaashoek1-3/+3
2011-08-16Switch back to #define for PHYSTOP. Using the E820 to retrieve the memory ↵Frans Kaashoek1-11/+5
map is too complicated (must be done in 16-bit real-mode, probably enlarged bootblock beyond 512 bytes, and a #define requires less explanation).
2011-08-16Clean up memlayout.hFrans Kaashoek1-9/+17
Get rid of last instances of linear address and "la" Get ready for detecting physical memory dynamically
2011-08-15Make Austin happierFrans Kaashoek1-12/+12
2011
2011-08-15Avoid "boot" in xv6Frans Kaashoek1-2/+2