8130e4752b
Map TSS using RMM
2021-05-07 09:13:25 -06:00
77f6887fbc
Debug use of memory map and env data
2021-05-07 08:54:36 -06:00
38f7884e21
Improvements for aarch64 serial driver on real hardware
2021-05-06 21:25:15 -06:00
36b3a4a49e
Add KERNEL_PERCPU_PML4 on AArch64.
2021-05-06 21:22:48 +02:00
558109a9cb
Put the KPCRs in high memory, in their own PML4.
...
This also removes the need to do another semi-expensive remap when
cloning processes, since the KPCRs (for kernel TLS) are no longer stored
in the user PML4.
2021-05-06 21:21:37 +02:00
2e38fab913
Do not hardcode aarch64 uart
2021-05-06 13:16:50 -06:00
9b4ce0d0cc
WIP: Fix userspace ACPI shutdown.
2021-05-06 19:49:58 +02:00
7ac5bdbae0
WIP: Implement userspace-driven shutdown.
2021-05-06 19:49:58 +02:00
bea6747643
Move all AML code to userspace.
...
Currently, there are some things that need to be set up by userspace
that the kernel previously did. These include telling firmware when the
I/O APIC is used, and most importantly, shutting down the system.
The former is not particularly important, but for the latter I think
that we could implement this using a "shutdown pipe". Essentially it
will be a file that triggers an event shutting down, which would be used
to notify to acpid that the kernel is requesting a shutdown.
2021-05-06 19:49:57 +02:00
46a364471c
Re-init aarch64 serial port for interrupts
2021-05-05 20:46:51 -06:00
b9a89f2160
More fixes for building aarch64
2021-05-03 22:07:40 -06:00
8f50785781
Fixes for building aarch64
2021-05-03 21:57:45 -06:00
8d61c79b23
Use RMM TableKind and fix x86_64 compilation
2021-05-03 21:15:46 -06:00
2aa4d8caf5
Merge remote-tracking branch 'origin/aarch64-rebase' into riscv64
2021-05-03 20:52:59 -06:00
17c261553b
Fixes for building x86_64
2021-05-03 20:43:18 -06:00
dfdb562e6b
Use RMM PhysicalAddress
2021-05-03 20:33:31 -06:00
9f0532b1d5
Remove unused flags from old EntryFlags
2021-05-03 17:29:55 -06:00
b9448274fc
Switch to using RMM PageFlags
2021-05-03 17:28:08 -06:00
dd0616cc8f
Use RMM for TLB flushing
2021-05-03 12:42:16 -06:00
29460e0bff
Use RMM for some arch-specific paging functions
2021-05-03 12:14:49 -06:00
ad39568fe9
Fix unwrapping stack when there are kernel addresses (and there always are)
2021-05-02 19:53:31 -06:00
73c77d756d
Print context name in exception
2021-04-28 21:15:21 -06:00
37e6951501
Print CPU and PID when exception occurs
2021-04-28 20:59:52 -06:00
a9bee0bbdc
Require UTF-8 for context name
2021-04-28 20:03:12 -06:00
b26c3e0ae9
Make context name a RwLock
2021-04-28 20:03:05 -06:00
f90033e0e1
Use PHYS_OFFSET instead of KERNEL_OFFSET to refer to the physmap
2021-04-13 20:38:54 -06:00
959a917ba2
Update RMM
2021-04-13 19:34:27 -06:00
bdc925d275
Use GS for TLS!
...
Previously, the kernel used the regular FS segment for Thread-Local
Storage. The problem however, is that userspace code also uses FS for
TLS, meaning that the kernel would have to switch the FS segment between
user and kernel, _upon every syscall_. This is obviously suboptimal for
performance (especially with fast syscalls such as futex, nanosleep, or
yield).
I had to search LLVM for hours, just to find out that the insertion of
the memory load with FS was actually done in the linker, so I added a
flag for that.
I haven't done any proper benchmarking, but the boot process seems to
have gotten much faster!
2021-02-17 14:44:45 +01:00
1988583e23
Forbid lower-half noncanonical addresses too.
2021-02-17 13:00:07 +01:00
a4b3af34a5
Use correct jump instruction.
2021-02-15 20:05:00 +01:00
c19581282d
Use the correct add instruction when popping CS.
2021-02-15 20:02:35 +01:00
8eb58891aa
Simplify sysretq code.
2021-02-15 19:53:49 +01:00
5b2df9f504
Document why usermode() can omit rcx check.
2021-02-15 19:53:49 +01:00
ff33090fd0
Check whether RCX is canonical in sysretq.
2021-02-15 19:53:41 +01:00
a183953ee8
Motivate usage of the IST without SWAPGS involved.
2021-02-15 19:53:37 +01:00
a3583a10ce
Only swapgs when leaving/entering userspace code.
2021-02-15 19:53:37 +01:00
05db0f5977
Temporarily fix sysretq by swapping gs 4 times.
...
In order words, it swaps gs both directly at the start of the syscall
handler, then swaps it back, and the at the end of the syscall handler.
I cannot tell for sure why this is necessary, but probably since some
interrupt handler will execute swapgs in the wrong order or something.
2021-02-15 19:53:37 +01:00
1a8016b985
Give NMI, #DF, and #MC handlers a special stack.
...
This is done by allocating an extra 64 KiB per CPU, and putting it in
the Interrupt Stack Table.
2021-02-15 19:53:37 +01:00
5a638691e0
Treat GS as always pointing to TSS in kernel space.
2021-02-15 19:53:37 +01:00
c913c3be80
Use sysretq in usermode().
2021-02-15 19:53:24 +01:00
a8dc3fcaf1
Begin using sysretq in the system call handler.
2021-02-15 19:53:01 +01:00
238702f7d1
Require UTF-8 for context name
2021-02-13 12:16:47 -07:00
b9f4a915ea
Make context name a RwLock
2021-02-13 11:10:21 -07:00
00723c4ac2
aarch64: Make IRQs use the exception macros
2021-01-26 19:37:23 +00:00
28dfc0f46b
aarch64: Basic exception handlers
2021-01-26 18:18:19 +00:00
4a215c7c2c
aarch64: exception management and clone fixups
2021-01-26 18:17:09 +00:00
3da345867a
aarch64: paging: Derive Debug, PartialEq for VirtualAddressType
...
This makes asserts on VirtualAddressType equality possible.
2021-01-21 11:38:46 +00:00
452196b81f
aarch64: consts: Use the same USER_TLS_SIZE as x86_64
2021-01-21 11:37:32 +00:00
c188a60871
aarch64: Fix clone_ret
...
FIXME: Explain the magic numbers here later.
2021-01-18 21:47:28 +00:00
67d72532a9
aarch64: usermode: Remove tpidr_el0 manipulation
...
Was using the incorrect USER_TLS_OFFSET instead of USER_TCB_OFFSET. In
any case, this is better done in process::clone.
2021-01-17 10:06:04 +00:00
Jeremy Soller