1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-05-22 11:45:08 +00:00
Commit graph

6988 commits

Author SHA1 Message Date
Idan Horowitz
871a53db76 AK: Make Bitmap construction OOM-fallible 2022-02-11 17:49:46 +02:00
Andreas Kling
c797eaa9f8 Kernel/Net: Don't update TCP socket "last sent ACK" field too early
Defer updating this field until after the last fallible operation has
succeeded.
2022-02-11 12:45:38 +01:00
Andreas Kling
06cf01cd1e Kernel/E1000: Bump RX/TX buffer count to 256/256
We were frequently dropping packets when downloading large files.
Then we had to wait for TCP retransmission which slowed things down.

This patch dramatically improves E1000 throughput by increasing the
number of RX/TX buffers from 32/8 to 256/256.

The largest chunk of JavaScript from Discord now downloads in roughly
1 second instead of 7 seconds. :^)
2022-02-11 12:45:38 +01:00
Andreas Kling
2ff9db0245 Kernel: Make contiguous VM objects use "user physical pages" by default
If someone specifically wants contiguous memory in the low-physical-
address-for-DMA range ("super pages"), they can use the
allocate_dma_buffer_pages() helper.
2022-02-11 12:45:38 +01:00
Idan Horowitz
1aad64fbe4 Kernel: Workaround QEMU hypervisor.framework CPUID max leaf bug
This works around issue #10382 until it is fixed on QEMU's side.
Patch from Anonymous.
2022-02-11 02:45:34 +02:00
Liav A
41dae9b3c7 Kernel: Convert i8042 code to use the ErrorOr pattern more broadly
Not only does it makes the code more robust and correct as it allows
error propagation, it allows us to enforce timeouts on waiting loops so
we don't hang forever, by waiting for the i8042 controller to respond to
us.

Therefore, it makes the i8042 more resilient against faulty hardware and
bad behaving chipsets out there.
2022-02-10 15:42:56 +02:00
Liav A
dc41a0b830 Kernel: Check i8042 existence before trying to use it
If we don't do so, we just hang forever because we assume there's i8042
controller in the system, which is not a valid assumption for modern PC
hardware.
2022-02-10 15:42:56 +02:00
Liav A
c6c3e2a7fd Kernel: Instantiate a TextModeConsole early on if there's no framebuffer
If the bootloader that loaded us is providing a framebuffer details from
the Multiboot protocol then we can instantiate a framebuffer console.
Otherwise, we should use a text mode console, assuming that the BIOS and
the bootloader didn't try to modeset the screen resolution so we have is
a VGA 80x25 text mode being displayed on screen.

Since "boot_framebuffer_console" is no longer a good representative as a
global variable name, it's changed to g_boot_console to match the fact
that it can be assigned with a text mode console and not framebuffer
console if needed.
2022-02-09 21:05:48 +00:00
Liav A
278b0aa629 Kernel/Graphics: Don't assert when disabling TextModeConsole
Not sure how it's useful to do so, let's not assert if something tries
to disable it. If we will use TextModeConsole as a boot console, that
console will be disabled after loading an appropriate console to replace
it.
2022-02-09 21:05:48 +00:00
Liav A
90a194377c Kernel/Graphics: Untie Text mode console from VGACompatibleAdapter class
Instead, we can construct this type of object without having to
instantiate a VGACompatibleAdapter object first.
This can help instantiate such console very early on boot to aid debug
issues on bare metal hardware.
2022-02-09 21:05:48 +00:00
Lenny Maiorani
c6acf64558 Kernel: Change static constexpr variables to constexpr where possible
Function-local `static constexpr` variables can be `constexpr`. This
can reduce memory consumption, binary size, and offer additional
compiler optimizations.

These changes result in a stripped x86_64 kernel binary size reduction
of 592 bytes.
2022-02-09 21:04:51 +00:00
Lenny Maiorani
138d54e595 AK+Kernel: Alphabetize debug macros
This is not ASCII-betical because `_` comes after all the uppercase
characters. Treating `_` as a ` ` (space character), these lists are
now alphabetical.
2022-02-09 17:59:19 +00:00
Andrew Kaster
353e72ac9b LibC+Kernel: Remove global variable use from snprintf and fprintf
The global variable use in these functions is super thread-unsafe and
means that any concurrent calls to sprintf or fprintf in a process
could race with each other and end up writing unexpected results.
We can just replace the function + global variable with a lambda that
captures the relevant argument when calling printf_internal instead.
2022-02-09 06:22:33 +00:00
Jamie Mansfield
8964f754f5 Kernel: Fix bug in TCP state handling in SynSent
When receiving SYN while in SynSent, we now reply with SYN|ACK in
addition to the SynSent->SynReceived transition.
2022-02-07 13:18:13 +01:00
Andreas Kling
cda56f8049 Kernel: Robustify and rename Inode bound socket API
Rename the bound socket accessor from socket() to bound_socket().
Also return RefPtr<LocalSocket> instead of a raw pointer, to make it
harder for callers to mess up.
2022-02-07 13:02:34 +01:00
sin-ack
24fd8fb16f Kernel: Ensure socket is suitable for writing in sys$sendmsg
Previously we would return a bytes written value of 0 if the writing end
of the socket was full. Now we either exit with EAGAIN if the socket
description is non-blocking, or block until the description can be
written to.

This is mostly a copy of the conditions in sys$write but with the "total
nwritten" parts removed as sys$sendmsg does not have that.
2022-02-07 12:21:45 +01:00
Andreas Kling
83523cabda Kernel: Fix bugs in TCP state handling in FinWait1 & FinWait2
1. When receiving FIN while in FinWait1, we now reply with ACK
   in addition to the FinWait1->Closing transition.

2. When receiving FIN|ACK while in FinWait1, we now reply with
   ACK and transition from FinWait1->TimeWait.

3. When receiving FIN while in FinWait2, we now reply with ACK.
2022-02-06 22:13:13 +01:00
Andreas Kling
7247f0204d Kernel: Send only FIN when shutting down TCP socket from ESTABLISHED
We were previously sending FIN|ACK for some reason.
2022-02-06 22:13:13 +01:00
Hendiadyoin1
fbb798f98c AK: Move integral log2 and exp to IntegerMath.h 2022-02-06 17:52:33 +00:00
Andreas Kling
04539d4930 Kernel: Propagate sys$profiling_enable() buffer allocation failure
Caught a kernel panic when enabling profiling of all threads when there
was very little memory available.
2022-02-06 01:25:32 +01:00
Andreas Kling
1d411a46c1 Kernel: Put kmalloc heap expansion debug spam behind KMALLOC_DEBUG 2022-02-05 22:50:39 +01:00
Liav A
b9f9cbb12c Kernel/Interrupts: Remove stale MSIHandler class
When we implement MSI support, we can rely on the IRQHandler class for
installing IRQ handlers at the right location.
2022-02-05 18:13:02 +00:00
Tom
4c506f91fe Kernel: Disable BootFramebufferConsole when drivers create a new one
When GraphicsManagement initializes the drivers we can disable the
bootloader framebuffer console. Right now we don't yet fully destroy
the no longer needed console as it may be in use by another CPU.
2022-02-04 21:34:12 +01:00
Tom
24f2f3ba4e Kernel: Set up an initial boot framebuffer console
Instead of seeing a black screen until GraphicsManagement was fully
initialized, this allows us to see the console output much earlier.
So, if the bootloader provided us with a framebuffer, set up a console
as early as possible.
2022-02-04 21:34:12 +01:00
Tom
eb446725d5 Kernel: Separate GenericFramebufferConsole implementation
The GenericFramebufferConsoleImpl class implements the logic without
taking into account any other details such as synchronization. The
GenericFramebufferConsole class then is a simple wrapper around
GenericFramebufferConsoleImpl that takes care of synchronization.

This allows us to re-use this implementation with e.g. different
synchronization schemes.
2022-02-04 21:34:12 +01:00
Idan Horowitz
3dc8bbbc8b Kernel: Remove the infallible make_ref_counted<T> factory function
This function had no users, nor should it ever be used, as all
allocation failures in the Kernel should be explicitly checked.
2022-02-03 23:33:20 +01:00
Idan Horowitz
a65bbbdb71 Kernel: Convert try_make_ref_counted to use ErrorOr
This allows more ergonomic memory allocation failure related error
checking using the TRY macro.
2022-02-03 23:33:20 +01:00
Idan Horowitz
8289727fac Kernel: Stop using the make<T> factory method in the Kernel
As make<T> is infallible, it really should not be used anywhere in the
Kernel. Instead replace with fallible `new (nothrow)` calls, that will
eventually be error-propagated.
2022-02-03 23:33:20 +01:00
Idan Horowitz
e5e7cb822a Kernel: Ignore allocation failures when trying to retransmit packets
We ignore allocation failures above the first 16 guaranteed socket
slots, as we will just retransmit their packets the next time around.
2022-02-03 23:33:20 +01:00
Idan Horowitz
2dc91865a4 Kernel: Stop allocating VirtIO configuration structs on the heap
These are trivially-copyable 12-byte structs, so there's no point in
allocating them on the heap.
2022-02-03 23:33:20 +01:00
Andreas Kling
34f6c88ffd Revert "Kernel: Protect InodeWatcher internals with spinlock instead of mutex"
This reverts commit 0bebf013e3.

This caused a deadlock when handling a crashed process, so let's revert
it until we can figure out what went wrong.
2022-02-03 18:25:55 +01:00
Andreas Kling
e7dc9f71b8 Kernel: Protect Inode flock list with spinlock instead of mutex 2022-02-03 17:28:45 +01:00
Andreas Kling
a81aebfd6e Kernel: Remove unnecessary mutex for ubsan-is-deadly ProcFS node 2022-02-03 16:11:26 +01:00
Andreas Kling
e86ab57078 AK+Kernel+LibSanitizer: Store "ubsan-is-deadly" flag as Atomic<bool> 2022-02-03 16:11:26 +01:00
Andreas Kling
0bebf013e3 Kernel: Protect InodeWatcher internals with spinlock instead of mutex 2022-02-03 16:11:26 +01:00
Andreas Kling
64254b5df8 Kernel: Protect PCI access with spinlock instead of mutex 2022-02-03 16:11:26 +01:00
Andreas Kling
200589ba27 Kernel: Turn VirtIOGPU operation lock from mutex into spinlock 2022-02-03 16:11:26 +01:00
Andreas Kling
ca42621be1 Kernel: Protect FramebufferDevice with spinlock instead of mutex 2022-02-03 16:11:26 +01:00
Andreas Kling
ddde9e7ee5 Kernel: Protect global device map with spinlock instead of mutx 2022-02-03 16:11:26 +01:00
Andreas Kling
e0d9472ced Kernel: Protect Inode's list of watchers with spinlock instead of mutex 2022-02-03 16:11:26 +01:00
Andreas Kling
210689281f Kernel: Protect mounted filesystem list with spinlock instead of mutex 2022-02-03 16:11:26 +01:00
Andreas Kling
3becff9eae Kernel: Protect network adapter list with spinlock instead of mutex 2022-02-03 16:11:26 +01:00
Andreas Kling
0899153170 Kernel: Protect PTYMultiplexer freelist with spinlock instead of mutex 2022-02-03 16:11:26 +01:00
Andreas Kling
6fbb924bbf Kernel: Protect ARP table with spinlock instead of mutex 2022-02-03 16:11:26 +01:00
Andreas Kling
248832f438 Kernel: Convert OpenFileDescriptor from mutex to spinlock
A mutex is useful when we need to be able to block the current thread
until it's available. This is overkill for OpenFileDescriptor.

First off, this patch wraps the main state member variables inside a
SpinlockProtected<State> to enforce synchronized access. This also
avoids "free locking" where figuring out which variables are guarded
by which lock is left as an unamusing exercise for the reader.

Then we remove mutex locking from the functions that simply call through
to the underlying File or Inode, since those fields never change anyway,
and the target objects perform their own synchronization.
2022-02-03 16:11:26 +01:00
Andreas Kling
35e24bc774 Kernel: Move Spinlock lock/unlock functions out of line
I don't see why these have to be inlined everywhere in the kernel.
2022-02-03 16:11:26 +01:00
Pankaj Raghav
d234e6b801 Kernel: Add polling support to NVMe
Add polling support to NVMe so that it does not use interrupt to
complete a IO but instead actively polls for completion. This probably
is not very efficient in terms of CPU usage but it does not use
interrupts to complete a IO which is beneficial at the moment as there
is no MSI(X) support and it can reduce the latency of an IO in a very
fast NVMe device.

The NVMeQueue class has been made the base class for NVMeInterruptQueue
and NVMePollQueue. The factory function `NVMeQueue::try_create` will
return the appropriate queue to the controller based on the polling
boot parameter.

The polling mode can be enabled by adding an extra boot parameter:
`nvme_poll`.
2022-02-02 18:26:59 +01:00
Pankaj Raghav
aa832ee251 Kernel: Add conditional call to disable_irq in IRQHandler constructor
There is no use in calling disable_irq function in the IRQHandler
constructor if irq was not registered before. So add a condition where
we call disable_irq only if the irq was registered before.
2022-02-02 18:26:59 +01:00
Pankaj Raghav
e5a6d12ff8 Kernel: Add nvme_poll command line parameters
As we don't currently support MSI(X) interrupts, it could be an issue
to boot on some newer hardware. NVMe devices support polling mode
where the driver actively polls for completion instead of waiting for
an interrupt.
2022-02-02 18:26:59 +01:00
Andreas Kling
d85f062990 Revert "Kernel: Only update page tables for faulting region"
This reverts commit 1c5ffaae41.

This broke shared memory as used by OutOfProcessWebView. Let's do
a revert until we can figure out what went wrong.
2022-02-02 11:02:54 +01:00