Inverting the tree turns all of the innermost stack frames into roots,
allowing them to accumulate their total sample counts with other
instances of the same frame being innermost. This is an essential
feature of any cool profiler, and now we have it. :^)
A client that only ever does synchronous IPC calls from its side would
never actually process incoming asynchronous messages since they would
arrive while waiting for a synchronous response and then end up sitting
forever in the "unhandled messages" queue.
We now always handle unhandled messages using a deferred invocation.
This fixes the bug where Audio.MenuApplet didn't learn that the muted
state changed in response to its own request to change it. :^)
This patch introduces the second MenuApplet: Audio. To make this work,
menu applet windows now also receive mouse events.
There's still some problem with mute/unmute via clicking not actually
working, but the call goes from the applet program over IPC to the
AudioServer, where something goes wrong with the state change message.
Need to look at that separately.
Anyways, it's pretty cool to have more applets running in their own
separate processes. :^)
Instead of implementing menu applets as their own thing, they are now
WSWindows of WSWindowType::MenuApplet.
This makes it much easier to work with them on the client side, since
you can just create a GWindow with the right type and you're in the
menubar doing applet stuff :^)
To enforce this, we create two separate mappings of the same underlying
physical page. A writable mapping for the kernel, and a read-only one
for userspace (the one returned by sys$get_kernel_info_page.)
This patch adds a single "kernel info page" that is mappable read-only
by any process and contains the current time of day.
This is then used to implement a version of gettimeofday() that doesn't
have to make a syscall.
To protect against race condition issues, the info page also has a
serial number which is incremented whenever the kernel updates the
contents of the page. Make sure to verify that the serial number is the
same before and after reading the information you want from the page.
Every process keeps its own ELF executable mapped in memory in case we
need to do symbol lookup (for backtraces, etc.)
Until now, it was mapped in a way that made it accessible to the
program, despite the program not having mapped it itself.
I don't really see a need for userspace to have access to this right
now, so let's lock things down a little bit.
This patch makes it inaccessible to userspace and exposes that fact
through /proc/PID/vm (per-region "user_accessible" flag.)
Currently only Ext2FS and TmpFS supports InodeWatchers. We now fail
with ENOTSUPP if watch_file() is called on e.g ProcFS.
This fixes an issue with FileManager chewing up all the CPU when /proc
was opened. Watchers don't keep the watched Inode open, and when they
close, the watcher FD will EOF.
Since nothing else kept /proc open in FileManager, the watchers created
for it would EOF immediately, causing a refresh over and over.
Fixes#879.
I had to change the layout of RegisterDump a little bit to make the new
IRQ entry points work. This broke get_register_dump_from_stack() which
was expecting the RegisterDump to be badly aligned due to a goofy extra
16 bits which are no longer there.
This logic is all taken care of by GAbstractColumnView now, so we can
simply delete GTreeView::context_menu_event(). :^)
Fixes an issue mentioned in #826
Instead of having a common entry point and looking at the PIC ISR to
figure out which IRQ we're servicing, just make a separate entryway
for each IRQ that pushes the IRQ number and jumps to a common routine.
This fixes a weird issue where incoming network packets would sometimes
cause the mouse to stop working. I didn't track it down further than
realizing we were sometimes EOI'ing the wrong IRQ.
Instead of fetching these from JSON in every paint event, we now have a
separate "SampleData" vector that can be iterated.
This optimization was made possible by profiling ProfileViewer and then
analyzing the profile with ProfileViewer! :^)
You can now select the time range you want on the profile timeline.
The tree view will update automatically as you alter the range.
Unfortunately this causes the treeview to collapse all of its nodes.
It would be nice to solve this somehow in the future so that nodes
can stay open.
The majority of the time in NetworkTask was being spent in allocating
and deallocating KBuffers for each incoming packet.
We'll now keep up to 100 buffers around and reuse them for new packets
if the next incoming packet fits in an old buffer. This is pretty
naively implemented but definitely cuts down on time spent here.
4KB gets pretty mmap/munmap heavy when downloading larger files,
so bump this a bit to reduce time spent in memory allocation.
This can be improved in various ways, but I'm not sure what the
best way forward is at the moment.
Since stream sockets don't actually need to deliver packets-at-a-time
data in recvfrom(), they can just buffer the payload bytes instead.
This avoids keeping one KBuffer per incoming packet in the receive
queue, which was a big performance issue in ProtocolServer.
This code is definitely not perfect and is something we should keep
improving over time.
