This allows disassembly of binaries with SSE2 instructions in them.
SSE2 also extends all MMX instructions without affecting the mnemonic,
therefore these are just directed to the same function for now.
The UserspaceEmulator does not know this as of
this commit.
This almost fully implements the SSE extension, similar to the x87 and
MMX extensions, using a separate class "SoftVPU".
Currently missing are all shadow and exception checks, as well as the
denormals-are-zero and flush-to-zero flags.
Also missing are some integer-SIMD functions.
The obsolete ttyname and ptsname syscalls are removed.
LibC doesn't rely on these anymore, and it helps simplifying the Kernel
in many places, so it's an overall an improvement.
In addition to that, /proc/PID/tty node is removed too as it is not
needed anymore by userspace to get the attached TTY of a process, as
/dev/tty (which is already a character device) represents that as well.
...and implement SoftCPU::read_memory<T> with it.
This allows the MMU to read a typed object (using 1-byte reads), which
is significantly nicer to use than reading the struct fields manually.
Instead of making it hold the shadow data as another `T`, make it hold
the data as a byte array, and allow it to read the byte array as `T`.
This makes it much easier to make a "read_typed" function in the MMU.
While trying out `ue --profile` today, I received an invalid json
profile. After poking around at the file it looks like we never close
the `events: [..` array that we generate, and thus end up with an
invalid document.
The fix is straight forward, always emit the closing brace.
Apologies for the enormous commit, but I don't see a way to split this
up nicely. In the vast majority of cases it's a simple change. A few
extra places can use TRY instead of manual error checking though. :^)
Previously, Emulator::virt$execve would not report ENOENT and EACCES
when the binary to be executed was nonexistent or not executable. This
broke the execp family of functions, which rely on ENOENT being reported
in order to know that they should continue searching $PATH.
Emulator::virt$execve would construct command lines such as
`/bin/UserspaceEmulator echo -- hello` instead of
`/bin/UserspaceEmulator -- echo hello`, which naturally caused problems.
This commit moves the "--" to the correct place.
This function is an extended version of `chmod(2)` that lets one control
whether to dereference symlinks, and specify a file descriptor to a
directory that will be used as the base for relative paths.
This commit fixes an issue where zero-alignment would lead to the
requested range being allocated outside of the total available range,
hitting an assert in RangeAllocator::allocate_anywhere().
If we do not mark these ranges as reserved, RangeAllocator might later
give us addresses that overlap these, which then causes an assertion
failure in the SoftMMU. This behavior led to recurring CI failures, and
sometimes made programs as simple as `/bin/true` fail.
Fixes "Crash 1" reported in #9104
Most other syscalls pass address arguments as `void*` instead of
`uintptr_t`, so let's do that here too. Besides improving consistency,
this commit makes `strace` correctly pretty-print these arguments in
hex.
This feature was introduced in version 4.17 of the Linux kernel, and
while it's not specified by POSIX, I think it will be a nice addition to
our system.
MAP_FIXED_NOREPLACE provides a less error-prone alternative to
MAP_FIXED: while regular fixed mappings would cause any intersecting
ranges to be unmapped, MAP_FIXED_NOREPLACE returns EEXIST instead. This
ensures that we don't corrupt our process's address space if something
is already at the requested address.
Note that the more portable way to do this is to use regular
MAP_ANONYMOUS, and check afterwards whether the returned address matches
what we wanted. This, however, has a large performance impact on
programs like Wine which try to reserve large portions of the address
space at once, as the non-matching addresses have to be unmapped
separately.
