Andrei Tatar [Tue, 21 Nov 2023 20:11:46 +0000 (21:11 +0100)]
vfscore: Move fd handling into posix-fdtab
This change moves the file descriptor table out of vfscore, along with
the fd-specific syscalls close and dup*.
Implementations of vfscore files are also patched to use posix-fdtab.
Checkpatch-Ignore: USE_NEGATIVE_ERRNO
Checkpatch-Ignore: LINE_SPACING Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1168
Andrei Tatar [Tue, 21 Nov 2023 19:34:43 +0000 (20:34 +0100)]
lib/posix-fdtab: Add legacy interface for vfscore
This change adds support for legacy vfscore files to posix-fdtab,
allowing both new uk_ofiles and vfscore files to be used simultaneously.
This includes legacy as well as type-agnostic ("shim") API additions.
Checkpatch-Ignore: UNNECESSARY_ELSE Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1168
Andrei Tatar [Tue, 21 Nov 2023 19:22:23 +0000 (20:22 +0100)]
lib/posix-fdtab: Introduce fdtab library
This change introduces the posix-fdtab library, tasked with managing the
file descriptor table, mapping integers to open file descriptions.
posix-fdtab exposes functionality through Unikraft-internal APIs and is
independent from the file descriptor table implementation in vfscore,
which it aims to replace.
Checkpatch-Ignore: VOLATILE
Checkpatch-Ignore: LINE_SPACING Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1168
Andrei Tatar [Mon, 20 Nov 2023 15:20:25 +0000 (16:20 +0100)]
lib/ukfile: Introduce the ukfile library
This change introduces the ukfile library to provide a new fundamental
abstraction for "files" in Unikraft. This forms the base for a larger
rearchitecting of vfscore.
Checkpatch-Ignore: ENOSYS
Checkpatch-Ignore: VOLATILE Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1165
Andrei Tatar [Fri, 24 Nov 2023 15:46:14 +0000 (16:46 +0100)]
lib/posix-timerfd: Add syscall interface
This change implements the syscall interface to posix-timerfd, adding
the following syscalls to Unikraft:
- timerfd_create
- timerfd_settime
- timerfd_gettime
Andrei Tatar [Mon, 2 Oct 2023 18:58:34 +0000 (20:58 +0200)]
include/arch: Add atomic fetch & AND operation
This change adds a Unikraft macro for an architecture's atomic fetch &
AND operation, similar to the fetch & OR we already have.
It also fixes the order of the fetch & OR operations to be truthful to
the docstring.
Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1125
Andrei Tatar [Mon, 2 Oct 2023 19:07:40 +0000 (21:07 +0200)]
lib/uklock: Add static initializer for uk_rwlock
Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1126
Andrei Tatar [Mon, 2 Oct 2023 19:04:47 +0000 (21:04 +0200)]
lib/uksched: Export static wait queue initializer
This change adds the convenience macro `UK_WAIT_QUEUE_INITIALIZER` to
the public API of uksched, enabling external code to use initializers
for embedded wait queues.
Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1126
Andrei Tatar [Thu, 16 Nov 2023 18:04:30 +0000 (19:04 +0100)]
include: Add strong/weak reference counting
This change adds a utility header implementing a reference counting
scheme that supports regular (strong) as well as weak references.
Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1160
Andrei Tatar [Wed, 16 Aug 2023 22:45:39 +0000 (00:45 +0200)]
lib/ukcpio: Optimize path handling
This change removes redundant string copies of the destination
extraction path, dropping about half of the total string copy ops.
In addition, some minor code style changes included.
Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1181
Andrei Tatar [Wed, 16 Aug 2023 22:00:31 +0000 (00:00 +0200)]
lib/ukcpio: Refactor CPIO extraction
This change reorganizes the CPIO extraction code, making the parsing of
header fields more explicit, as well as splitting functionality into
dedicated handling functions.
Checkpatch-Ignore: CODE_INDENT
Checkpatch-Ignore: LEADING_SPACE
Checkpatch-Ignore: LINE_SPACING
Checkpatch-Ignore: UNNECESSARY_ELSE Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1181
Andrei Tatar [Wed, 16 Aug 2023 17:58:24 +0000 (19:58 +0200)]
lib/ukcpio: Clean up & organize code
This change is a minor code cleanup with no significant logic changes,
in preparation for a larger refactoring; it includes:
- improved macros with more logical ordering
- code deduplication
- cleaner structured `read_section` function
Checkpatch-Ignore: COMPLEX_MACRO Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1181
Andrei Tatar [Thu, 16 Nov 2023 18:22:22 +0000 (19:22 +0100)]
lib/uktimeconv: Add conversion functions
This change adds a header with utility functions related to various time
formats defined by standard C or POSIX.
This header depends on time headers provided by (no)libc.
Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1162
Andrei Tatar [Mon, 2 Oct 2023 19:04:47 +0000 (21:04 +0200)]
lib/uksched: Export static wait queue initializer
This change adds the convenience macro `UK_WAIT_QUEUE_INITIALIZER` to
the public API of uksched, enabling external code to use initializers
for embedded wait queues.
Andrei Tatar [Tue, 7 Nov 2023 13:59:41 +0000 (14:59 +0100)]
include/arch: Make time constants signed
This change removes the unsigned marker on integer literals that define
time-related constants, as there is no logical reason for time to always
be positive. This also alleviates mixed-signedness warnings when doing
arithmetic on time values, as well as assigning to __snsec.
Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1157
Andrei Tatar [Tue, 7 Nov 2023 13:53:45 +0000 (14:53 +0100)]
include/arch: Add raw compare-exchange operation
This change adds a Unikraft macro for the raw compare-exchange
operation, which in contrast to the existing macro may modify its
arguments.
Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1156
Andrei Tatar [Tue, 7 Nov 2023 14:12:30 +0000 (15:12 +0100)]
lib/vfscore: Fix crash in fstatat on NULL args
This change makes fstatat gracefully handle NULL arguments by returning
-EFAULT, preventing a hard crash.
Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1155
Andrei Tatar [Tue, 7 Nov 2023 14:06:48 +0000 (15:06 +0100)]
lib/posix-time: Allow NULL arg in clock_getres
This change fixes the behavior of clock_getres to not return an error
when receiving a NULL as its output argument, in accordance with docs.
Signed-off-by: Andrei Tatar <andrei@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1154
Sergiu Moga [Wed, 1 Nov 2023 11:02:47 +0000 (13:02 +0200)]
plat/kvm/arm: Switch to auxiliary stack on syscall entry
We would usually use the trap stack to execute the issued
system call. However this may be troublesome if the syscall will
generate page faults for example, which would end up reusing the
same stack.
To fix this, switch to the per-thread auxiliary stack on system call
entry and use that instead. The switch shall be done with the help of
the `tpidr_el1` register to help us, as it will be guaranteed to contain
`struct lcpu` that has an up to date pointer to the current thread's
auxiliary stack.
Sergiu Moga [Sat, 4 Nov 2023 14:54:34 +0000 (16:54 +0200)]
plat/common/x86: Save `ss`/`rsp`/`eflags`/`cs`/`rip` on syscall frame
Previously, we would save zeroes on the syscall entry frame instead
of using the actual values for the registers that would normally be
popped off the stack since they would not be used anyway by a syscall.
For completeness's sake, do still save them, to offer those that may
view the structure a clear view of what the syscall will return to.
Furthermore, we no longer need to clear the stack before syscall exit
(the previous `addq $(6 * 8), %rsp`) because the auxiliary stack is
used instead, so do not do this operation anymore.
Sergiu Moga [Wed, 1 Nov 2023 11:02:47 +0000 (13:02 +0200)]
plat/common/x86: Switch to auxiliary stack on syscall entry
We would usually use the application's stack to execute the issued
system call. However this present a couple of problems such as
if the application's stacks are too small (like it is the case
for Go's goroutines) this will end up generating an unhandled
pagefault or, even worse, corrupting other memory areas beyond
the respective stack. Another problem would be that we will
end up overwriting stuff on the Red Zone for those applications
that employ it.
To fix this, switch to the per-thread auxiliary stack on system call
entry and use that instead. The switch shall be done with the help of
the `swapgs` instruction to help us swap the `gs_base` registers, as
our own `KERNEL_GS_BASE` contains a pointer to the current LCPU's
`struct lcpu` that has an up to date pointer to the current thread's
auxiliary stack.
Sergiu Moga [Wed, 1 Nov 2023 07:02:25 +0000 (09:02 +0200)]
lib/uksched: Update `auxsp` of current LCPU when switching stacks
If uksched/ukthread is enabled, make sure that the current LCPU's
auxiliary stack pointer is updated to always point to the currently
executing thread's auxiliary stack pointer by setting it to that
of the thread it is about to switch to during context switching.
Sergiu Moga [Wed, 1 Nov 2023 06:57:34 +0000 (08:57 +0200)]
lib/uksched: Set bootstrap lcpu's `auxsp` during `uk_sched_start`
Make sure to fill in the value of current lcpu's (bootstrap lcpu in
our case) auxiliary stack pointer with that of the current thread's
(main thread) when initializing scheduling.
If shcheduling/multithreading is not enabled, allocate an auxiliary
stack for bootstrap LCPU. Otherwise, the LCPU will have the same
auxiliary stack as the thread that runs on it and this allocation is
not needed.
Sergiu Moga [Tue, 31 Oct 2023 16:49:00 +0000 (18:49 +0200)]
lib/uksched: Introduce the per-thread auxiliary stack
Add a new field to `struct uk_thread` that can represent a secondary
stack that can be used as a backup stack.
This can become very useful in cases such as when wanting to defer
exception handling without creating another thread. For example,
we want to return from an exception into a function inside the
same thread to be able to do deferred I/O outside exception
context without contaminating the original stack that was present
before the trap. We can avoid polluting the original stack by using
this auxiliary stack instead.
Update each thread creation/deletion method's signature and
implementation accordingly.
Sergiu Moga [Wed, 1 Nov 2023 06:44:07 +0000 (08:44 +0200)]
plat/common: Add `auxsp` field to `struct lcpu`
`struct uk_thread` contains a field called `auxsp` which is meant to
represent an auxiliary stack that may be used when switching stacks
during syscall entries or when simply wanting to have a scratch space
to use during a very fragile state of the system, such as when handling
an exception.
Give access to this field through `struct lcpu` as well. This field
shall represent the auxiliary stack pointer of the thread currently
executing on this LCPU.
If uksched/ukthread is disabled, this can come in handy as simply a
pointer to a scratch space that we can contaminate with whatever we
may want during a more fragile execution context.
Sergiu Moga [Thu, 23 Nov 2023 14:00:34 +0000 (16:00 +0200)]
arch/arm64: Implement stack switching on IRQ/trap
Define a per-cpu buffer whose size is twice that of the configured
stacks and whose definition may be represented through the following
diagram:
```
STACK_SIZE STACK_SIZE
<---------------------><--------------------->
|============================================|
| | |
| trap stack | IRQ stack |
| | |
|=============================================
^
SP_EL0
```
The middle address of this buffer shall be assigned to `SP_EL0`,
a great candidate for a register free of use as a `EL1` only Unikernel.
Now, depending on whether an exception is an IRQ or a trap, the early
assembly entry will switch to either the IRQ stack or the trap stack,
by simply making use of the `SP_EL0` system register.
Sergiu Moga [Thu, 23 Nov 2023 13:11:41 +0000 (15:11 +0200)]
plat/common: Move LCPU `id`/`current`/`idx` methods to arch files
Now that we have a per-cpu system register that holds the current
pointer to the `struct lcpu` element of the global `lcpus` array,
make use of that for each architecture independently and move
the implementations of `ukplat_lcpu_[id|current|idx]` from the
architecture agnostic place to the more architecture specific ones.
Notice that a default definition resorting to the value `0` is still
kept for platforms that do not have SMP and do not have a LCPU API
integrated.
Furthermore, obsolete `CONFIG_UKPLAT_LCPU_IDISIDX` since we can now
directly fetch `idx` from `struct lcpu` pointer stored in the
corresponding system register.
NOTE: `lcpu_get_current` of `x86` still makes use of the now
optimized through `%gs` register by calling `ukplat_lcpu_idx`
which calls `ukplat_lcpu_id`. However, in the future, it
will be better if we make use of the `rdgsbase` instruction
and the like.
Co-authored-by: Michalis Pappas <michalis@unikraft.io> Co-authored-by: Marco Schlumpp <marco@unikraft.io> Signed-off-by: Michalis Pappas <michalis@unikraft.io> Signed-off-by: Marco Schlumpp <marco@unikraft.io> Signed-off-by: Sergiu Moga <sergiu@unikraft.io>
Sergiu Moga [Mon, 23 Oct 2023 16:19:05 +0000 (19:19 +0300)]
lib/devfs: Remove dependency on `LIBVFSCORE_AUTOMOUNT_ROOTFS`
Now that we have `fstab`, we no longer depend on something mounting
a rootfs being preconfigured within the kernel. A user may simply
choose to specify rootfs through the `vfs.fstab` command-line
argument.
Signed-off-by: Sergiu Moga <sergiu@unikraft.io> Reviewed-by: Alexander Jung <alex@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1146
Sergiu Moga [Thu, 23 Nov 2023 12:55:19 +0000 (14:55 +0200)]
plat/common/arm: Init `TPIDR_EL1` in `lcpu_arch_init`
Every LCPU shall have their `TPIDR_EL1` system register assigned
the value of the address of their own `struct lcpu` element in the
global `lcpus` array.
Implement basic methods to be able to read/write from/to the
GS_BASE and KERNEL_GS_BASE MSR's, as well as from an offset relative
to the former's value.
Co-authored-by: Marco Schlumpp <marco@unikraft.io> Signed-off-by: Marco Schlumpp <marco@unikraft.io> Signed-off-by: Sergiu Moga <sergiu@unikraft.io>
Sergiu Moga [Wed, 1 Nov 2023 06:32:07 +0000 (08:32 +0200)]
plat/common/x86: Add macro-definition for `X86_MSR_KERNEL_GS_BASE`
Add a macro-definition for the x86 `KERNEL_GS_BASE` MSR that is usually
used in conjunction with the `swapgs` instruction for an efficient swap
between the `gs_base` register placed in `KERNEL_GS_BASE` and `GS_BASE`.
Co-authored-by: Marco Schlumpp <marco@unikraft.io> Signed-off-by: Marco Schlumpp <marco@unikraft.io> Signed-off-by: Sergiu Moga <sergiu@unikraft.io>
NOTE: This is interdependent on the `app-elfloader` commit that
changes `arch_prctl`'s `ARCH_SET_FS` behavior, entitled:
```
Do not change real `fs_base` register in `arch_prctl`
```
The code may store and restore userland TLS pointer properly in the
system call handler and it does check if it was changes by syscalls
such as `arch_prctl`'s `ARCH_SET_FS` command. However, one aspect
was missed: `__UK_SYSCALL_RETADDR_CLEAR`. The macro
`__UK_SYSCALL_RETADDR_CLEAR` accesses what is meant to be Unikraft's
TLS and is invoked in the time window between TLS store and restore.
In the `ARCH_SET_FS` case, `arch_prctl` sets the the `fs` register
regardless and thus we end up having the userland desired TLS pointer
during a time window where keeping our own TLS pointer is mandatory.
Thus, to avoid this, do not set any `fs` register value during
`arch_prctl` (see top NOTE) and, instead, simply modify the
`_uk_syscall_ultlsp` value which will be used instead to restore
the userland TLS as the last operation.
Alexander Jung [Wed, 22 Nov 2023 13:46:33 +0000 (14:46 +0100)]
build: Introduce `UK_IMAGE_NAME_OVERWRITE` make var
This commit introduces a new make variable,
`UK_IMAGE_NAME_OVERWRITE`, which is used to customize the
output binary name of the kernel image. This is a special,
and experimental, variable which can be helpful in certain
use cases, for example where targets with the same architecture
and platform are present. This variable should be used in
conjunction with a different build directory, `O=`, as the
resulting binary will always have the name from this option.
Signed-off-by: Alexander Jung <alex@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1169
Marco Schlumpp [Mon, 27 Feb 2023 14:15:58 +0000 (15:15 +0100)]
lib/uknetdev: Add ipv4_dns0_addr configuration
Previously, we were only able to statically configure the ip address,
netmask, and gateway, but not the DNS server. This commit adds the
configuration option for a DNS server.
Signed-off-by: Marc Rittinghaus <marc.rittinghaus@kit.edu> Signed-off-by: Marco Schlumpp <marco@unikraft.io> Reviewed-by: Simon Kuenzer <simon@unikraft.io> Approved-by: Simon Kuenzer <simon@unikraft.io>
GitHub-Closes: #1167
Andrei Stan [Mon, 20 Nov 2023 21:32:57 +0000 (23:32 +0200)]
plat/xen/x86_64: Add missing segment alignement for `xen`
Make the second segment in a `xen/x86-64` ELF aligned to the page size.
This fixes a memory overlap issue that triggers an assert during
memory coalescing.
Signed-off-by: Andrei Stan <andreistan2003@gmail.com>
cocodery [Mon, 20 Nov 2023 04:33:41 +0000 (12:33 +0800)]
lib/ukrust: Fix compilation fialure when enable libukrust
This commit aims to fix issue#1147 which compile apps with enabled
libukrust
1. ukrust/Makefile.uk: fix command line argument with high-version
bindgen, from 'size_t-is-usize' to 'no-size_t-is-usize'.
2. ukrust/src/allocator.rs: fix previous defined symbol by comment
them, becasue rustc will help to resolve, and add new symbol
'__rust_no_alloc_shim_is_unstable' for locate.
cocodery [Fri, 17 Nov 2023 16:17:03 +0000 (00:17 +0800)]
build: guarantee files generated by awk-file before others rely on them
Fix sub-problem of issue#1147 when enable libukrust, which will throw an error that
`fatal error: 'uk/bits/libid.h' file not found` because of wrong commands order.
this header file is needed by `bindings_helper.h` but command generates
'uk/bits/libid/h' is ahead from the bindgen command.
Eduard Vintilă [Sun, 12 Nov 2023 13:54:34 +0000 (15:54 +0200)]
plat/common/arm: Zero out `x29` and `x30` before stack switch
In order to properly mark the end of a stacktrace and avoid
undefined behaviour during unwinding, both the frame pointer
(`x29`) and the link register (`x30`) should be
zeroed out before switching to a new stack.
Signed-off-by: Eduard Vintilă <eduard.vintila47@gmail.com>
Andrei Tatar [Mon, 2 Oct 2023 18:58:34 +0000 (20:58 +0200)]
include/arch: Add atomic fetch & AND operation
This change adds a Unikraft macro for an architecture's atomic fetch &
AND operation, similar to the fetch & OR we already have.
It also fixes the order of the fetch & OR operations to be truthful to
the docstring.
projects / unikraft / unikraft.git / log