Now the kernel supports 2 ECAM access methods.
MMIOAccess was renamed to WindowedMMIOAccess and is what we had until
now - each device that is detected on boot is assigned to a
memory-mapped window, so IO operations on multiple devices can occur
simultaneously due to creating multiple virtual mappings, hence the name
is a memory-mapped window.
This commit adds a new class called MMIOAccess (not to be confused with
the old MMIOAccess class). This class creates one memory-mapped window.
On each IO operation on a configuration space of a device, it maps the
requested PCI bus region to that window. Therefore it holds a SpinLock
during the operation to ensure that no other PCI bus region was mapped
during the call.
A user can choose to either use PCI ECAM with memory-mapped window
for each device, or for an entire bus. By default, the kernel prefers to
map the entire PCI bus region.
Apparently we don't enable PCI ECAM (MMIO access to the PCI
configuration space) even if we can. This is a regression, as it was
enabled in the past and in unknown time it was regressed.
The CommandLine::is_mmio_enabled method was renamed to
CommandLine::is_pci_ecam_enabled to better represent the meaning
of this method and what it determines.
Also, an UNMAP_AFTER_INIT macro was removed from a method
in the MMIOAccess class as it halted the system when the kernel
tried to access devices after the boot process.
Object introspection in the Browser's JS console is still not great, but
this makes it a lot easier to find out the exact type of an object by
checking its 'constructor' property.
It also fixes all the things that rely on these properties being set, of
course :^)
This was super confusing as we would check if the exception's value is a
JS::Error and not log it otherwise, even with m_should_log_exceptions
set.
As a result, things like DOM exceptions were invisible to us with
execution just silently stopping, for example.
Not sure if this regressed at some point or just never worked, it
definitely wasn't tested at all. We would always return undefined when
returning from a try statement block, handler, or finalizer.
By using regex::AllFlags::SkipTrimEmptyMatches we get a null string for
unmatched capture groups, which we then turn into an undefined entry in
the result array instead of putting all matches first and appending
undefined for the remaining number of capture groups - e.g. for
/foo(ba((r)|(z)))/.exec("foobaz")
we now return
["foobaz", "baz", "z", undefined, "z"]
and not [
["foobaz", "baz", "z", "z", undefined]
Fixes part of #6042.
Also happens to fix selecting an element by ID using jQuery's $("#foo").
A FrameHostElement is an HTML element (<frame> or <iframe>) that may
have a content frame that participates in the frame tree.
This basically just moves code from <iframe> to a separate base class
so we can share it with <frame> once we implement <frame>.
The end goal of this commit is to allow to boot on bare metal with no
PS/2 device connected to the system. It turned out that the original
code relied on the existence of the PS/2 keyboard, so VirtualConsole
called it even though ACPI indicated the there's no i8042 controller on
my real machine because I didn't plug any PS/2 device.
The code is much more flexible, so adding HID support for other type of
hardware (e.g. USB HID) could be much simpler.
Briefly describing the change, we have a new singleton called
HIDManagement, which is responsible to initialize the i8042 controller
if exists, and to enumerate its devices. I also abstracted a bit
things, so now every Human interface device is represented with the
HIDDevice class. Then, there are 2 types of it - the MouseDevice and
KeyboardDevice classes; both are responsible to handle the interface in
the DevFS.
PS2KeyboardDevice, PS2MouseDevice and VMWareMouseDevice classes are
responsible for handling the hardware-specific interface they are
assigned to. Therefore, they are inheriting from the IRQHandler class.
We do support AHCI now, but the implementation could be incomplete for
some chipsets.
Also, we should write the acronym "Non-volatile Memory Express" as
NVMe. not NVME.
This reverts commit 36a82188a8.
This register is write-only for the firmware (BIOS), and read-only for
us so we shouldn't set the PCI IRQ line never.
The firmware figured out the IRQ routing to the PIC for us, so changing
it won't affect anything. I was mistaken when I thought that changing
the value of this register will allow us to change its interrupt line,
like when changing a PCI BAR to relocate device resources as desired
with the requirements of the OS.
Also handle native and compatibility channel modes together, so if only
one IDE channel was set to work on PCI native mode, we need to handle it
separately, so the other channel continue to operate with the legacy IO
ports and interrupt line.
Update the painting of background images for both <body> nodes and other
non-initial nodes. Currently, only the following values are supported:
repeat, repeat-x, repeat-y, no-repeat
This also doesn't support the two-value syntax which allows for setting
horizontal and vertical repetition separately.
This page tests the following values for background-repeat:
repeat, repeat-x, repeat-y, no-repeat
The test is duplicated for the <body> node and for child <div> nodes,
because the code that paints these nodes are in separate locations.
If we don't limit the sizes of the intermediate results, they will grow
indefinitely, causing each iteration to take longer and longer (in both
memcpy time, and algorithm runtime).
While calculating the trimmed length is fairly expensive, it's a small
cost to pay for uniform iteration times.
The user may now request specific cipher suites, the use of SNI, and
whether we should validate certificates (not that we're doing a good job
of that).
While space distribution along the primary axis of a BoxLayout is
pretty sophisticated, the secondary axis is very simple: we simply
center the widget.
However, this doesn't always look very nice if we don't take margins
into account, so make sure we subtract them from the rect we do all
the centering within.
If you don't need/want to use Fuse+ex2 then half of the existing
install command is unnecessary, and it's hard to pick out which you
do and don't need to, for example, build Lagom. This makes it clear
which commands you can skip if you don't need ex2 support.
This fixes an issue where `undefined.foo = "bar"` would throw a
ReferenceError instead of a TypeError as undefined was also used for
truly unresolvable references (e.g. `foo() = "bar"`). I also made the
various error messages here a bit nicer, just "primitive value" is not
very helpful.
We should be able to get the 'typeof' string for any value directly, so
this is now a standalone Value::typeof() method instead of being part of
UnaryExpression::execute().
This is required for the block formatting context to know the height of
the <br> element while computing the height of its parent. Specifically,
this comes into play when the <br> element is the first or last child of
its parent. In that case, it previously would not have any fragments, so
BlockFormattingContext::compute_auto_height_for_block_level_element
would infer its top and bottom positions to be 0.
This adds a double-click speed slider control to the Mouse Settings
panel, and value labels for both the movement speed and double-click
speed sliders.
To allow for updating and persisting the configured double-click
speed through the WindowServer, two IPC calls - `SetDoubleClickSpeed`
and `GetDoubleClickSpeed` - have been added.
The old approach was more complex and also had a very bad edge case
with lots of collisions. This approach eliminates that possiblility.
It also makes both reading and writing lookups a little bit faster.
We now leverage the VM's promise rejection tracker callbacks and print a
warning in either of these cases:
- A promise was rejected without any handlers
- A handler was added to an already rejected promise
