Nothing fundamentally changed, just update to new feature bits and print
them at device detection time.
Try to negotiate the guest feature set as well, though nothing
fundamentally changes at this time.
In the non VM path the existing routine wouldn't subtract from len, so
the function would (properly) return bytes transferred instead of zero.
The wrapping code was written to assume 0 and not bytes transferred,
which seemed like a workaround for broken code. Change the inner routine
to always return bytes transferred and adjust wrapper routines
accordingly.
Use a new directory iterator routine that tracks the offset within the
directory and handles transitions between sectors and clusters.
Redo the entry parsing code to make a single pass across the long file
names to handle crossing sector boundaries.
Move the directory handling code to dir.cpp
Handle walking the entire directory structure
Make file open be a much simpler function that just calls dir routines
TODO:
properly deal with FAT16 and FAT12 root directories
remove runtime mallocs in dir routines
If we were booted at EL2 (e.g. when passing -machine
virt,virtualization=on), we need to use SMC instead of HVC for PSCI
calls. Change psci_call() to do this and add a flag to do-qemuarm to
allow testing this scenario.
Clang will omit the initial comparison in this while loop on the
assumption that two different variables will not have the same address,
which will lead to a crash if the kernel has no constructors.
It will, however, retain the comparison for incomplete arrays
because they may have zero size (and therefore may alias with another
variable). Therefore, change the declarations of the start/end symbols
for the constructor list to incomplete arrays.
Now seems to be capable of mounting a simple fat32 volume and
successfully reading a single file in the root directory that spans a
few 512 byte clusters.
If two mount points had similar names but one was longer, the path
matching logic would be triggered on the shorter name.
ie, /foo and /foo2 were illegal, since /foo2 path matching would match
against /foo. Tighten the logic a bit to match against the matching
element of the passed in path.
Note: heirarchial path matching still works, so /foo/bar/baz will match
against a mount point at /foo/bar. Debatable if heirarchial mount points
should work at the moment with this simple logic, but it's there for
now.
Port to the really neat 68010 based board at https://rosco-m68k.com/
Port Features:
-10Mhz 68010
-1MB ram
-Dual UART + timer implemented as a 68c681 chip
-timer running at 1Khz, UART A for console
-interrupt driven RX support
Some amount of extending of the 68k exceptinon code was needed to
support the autovectored irqs that the 68681 uart uses. Added build
system support for 68010.
It is possible for early initialization functions such as lk_main()
to contain NEON instructions because we don't build the kernel with
-mgeneral-regs-only. As a result we can end up taking an FPU exception
before we are ready to handle it.
We didn't have this problem when starting at a higher exception level
than EL1 because we turned off FPU traps in arm64_elX_to_el1(). But we
neglected to do so when starting at EL1. Fix the problem by moving the
CPACR_EL1 manipulation out of arm64_elX_to_el1() and into arm_reset().
Previously was hard coding the instructions to work around a limitation
of the assembler that did not allow using fpu instructions when the code
was being compiled without support. Move the zeroing routine into a
separate assembly file and override the architure at the top.
