They're never used, so just drop them.
Change-Id: Ie137deed3e7258f9d6af7e0cb508e73df0f53ee0
Signed-off-by: Karl Palsson <karlp@tweak.net.au>
Reviewed-on: http://openocd.zylin.com/4131
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Previously it might read an address multiple times if an abstract
command took longer to execute than expected.
The new implementations reads from the target how far it has gotten
along reading memory, and resumes from there if cmderr=busy.
This ended up being a bigger change than I envisioned, but in the end it
deleted more lines than it added, so I'm happy. :-)
The downloaded program now post-increments, and there's no longer an
attempt to read the current address from the target. This made it easier
to fix the problem where at the start of the loop the current address
was already read (in regular entry) or has not yet been read (when the
first round through the loop encountered busy more than once, or busy
was encountered at least once later on).
The interesting new code concerns ignore_prev_addr and
this_is_last_read.
Additionally, I tweaked some debug output, and optimized
riscv_batch_run() when the batch is empty.
The actual implementation of triggers didn't change between those two
versions, so there's no need to duplicate the code.
In the process, I also fixed a minor multicore bug where tselect didn't
always get written on all harts.
This is the only way the spec guarantees that GPRs are accessible, and
depending on the implementation this might be the only way that CSRs are
accessible.
Also changed the debug code that parses out DMI fields to be simpler to
maintain (albeit a little slower).
riscv013_execute_debug_buffer() now automatically clears cmderr if the
command fails. That feels like the right behavior. (It does return the
error to its caller.)
When first connecting to a target, have the debugger disable any
hardware triggers that are set by a previously connected debugger.
The 0.11 code already did this, but 0.13 did not.
To achieve this I decided to share the code to enumerate triggers
between 0.11 and 0.13, which required me to implement get_register() and
set_register() for 0.11, which made the whole change a lot larger than
you might have guessed.
Hopefully this sets us up to in the future share the code to set/remove
triggers as well.
They were implemented, and people want to keep using them.
Also make OpenOCD tolerate cores that have $misa at 0xf10 instead of the
current address of 0x301.
Actually return an error when we fail to read a CSR.
Tweak cache_set32() debug output.
Commit 47b8cf842 changed the fixed type of the value argument to snprint
but didn't change the format string to match for sizes != 64 bit.
Change-Id: I908b06f49ab69d04224282949190a0de883048e0
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/4167
Tested-by: jenkins
Reviewed-by: Stian Skjelstad <stian@nixia.no>
Reviewed-by: Philipp Guehring <pg@futureware.at>
Now that we're supporting non-RTOS multi-hart mode there's some more
assertions that you're running on the right hart. Those assertions
aren't sane very early in examine, so I avoid them.
Rather than having a bunch of "if rtos" stuff, I now just check "if
hart_enabled". This makes some code paths cleaner, all of which were
buggy in the non-RTOS multi-hart mode.
When I disappeared the polls everywhere I forgot to sanitize the hartid
after halting. This is an invariant that GDB expects: when you return
from a halt whatever thread is marked as currently selected is the
thread that the next register accesses reference.
Main change is to make riscv_addr_t be unsigned. The rest is mechanical
fixing of types, print statements, and a few signed/unsigned compares.
Smoketest indicates everything is working more or less as before.
Supported SoCs: AR71xx, AR724x, AR91xx, AR93xx, QCA9558
Extended and revised version of my original patch submitted by Dmytro
here: http://openocd.zylin.com/#/c/3390
This driver is using pure SPI mode, so the flash base address is not
used except some flash commands (e.g. "flash program") need it to
distinguish the banks.
Example config with all 3 chip selects:
flash bank flash0 ath79 0 0 0 0 $_TARGETNAME cs0
flash bank flash1 ath79 0x10000000 0 0 0 $_TARGETNAME cs1
flash bank flash2 ath79 0x20000000 0 0 0 $_TARGETNAME cs2
Example usage:
> flash probe flash0
Found flash device 'win w25q128fv' (ID 0x001840ef)
flash 'ath79' found at 0x00000000
> flash probe flash1
No SPI flash found
> flash probe flash2
No SPI flash found
> flash banks
> flash read_bank flash0 /tmp/test.bin 0x00000000 0x1000
reading 4096 bytes from flash @0x00000000
wrote 4096 bytes to file /tmp/test.bin from flash bank 0 at offset
0x00000000 in 28.688066s (0.139 KiB/s)
Change-Id: I5feb697722c07e83a9c1b361a9db7b06bc699aa8
Signed-off-by: Tobias Diedrich <ranma+openocd@tdiedrich.de>
Reviewed-on: http://openocd.zylin.com/3612
Tested-by: jenkins
Reviewed-by: Dmytro <dioptimizer@hotmail.com>
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
This means I don't know what hart to look at, so I might as well
invalidate the register cache. Without this, you might get stale
registers the first time you ask for them.
I thought OpenOCD did this, but it looks like that doesn't happen when
runningi in RTOS mode. With this I can get to the end of most of the
RTOS tests, but they SIGINT instead of exiting.
There is no need to implement scan code in functions
mips_ejtag_get_idcode/impcode(), use mips_ejtag_drscan_32().
Impcode/idcode saved in ejtag.info.
Reorder the code in the callers of this functions.
Change-Id: Ia829c783a0b24c6a65cade736113fa6f67b0a170
Signed-off-by: Salvador Arroyo <salvador@telecable.es>
Reviewed-on: http://openocd.zylin.com/4003
Tested-by: jenkins
Reviewed-by: Peter Mamonov <pmamonov@gmail.com>
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Breakpoint setting based on length (kind) only.
Added 16bit aligned 32bit software breakpoints
support and same filtering before setting
breakpoint.
Set the required isa bit in hardware breakpoints.
Drop the isa bit in software breakpoints.
Change-Id: I7020f27be16015194b76f385d9b8e5af496d0dfc
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4124
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Read and save configuration registers, up to 4.
Config3 holds the micromips implementation info.
Added isa implementation info to mips32_common.
Added isa filter to avoid common mistakes, but only
if one isa mode is implemented.
When resuming the isa requested is set if more than
one isa mode is implemented.
Change-Id: I1d6526c5525bffac8d75e031b842b2edc6310e28
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4123
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Needed to run in micromips mode. Seems that if an isa
is supported in debug mode it also supported in kernel
mode. The contrary is not true.
Change-Id: I1feb8e2c376f4db97089f05c20bc0cd177208fb3
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4033
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Micromips is 16bit oriented, branch and jumps are
16 bit based. The upper half 16bits of a 32bit instruction
with the major opcode, must go first in the instruction
stream, hence the SWAP16 macro and swap16 array function,
needed if the code is written as 32 bit word in little endian
cores. Endianess info added to ejtag_iinfo. Pointer to
ejtag_info and isa field added to pracc context.
MIPS32 code are renamed to MIPS32_ISA_...
To select the isa, the new code has an additional isa parameter
(1 for micromips, 0 for mips32).
In JR instruction the isa bit must be set to execute
micromips code.
The suffix u is added to the OP codes to avoid signed/unsigned
comparison errors and to make sure the right shift is
performed logically.
The isa in debug mode is updated in the poll function.
Code for miniprograms, in kernel mode, need to be converted.
CFI code only for mips32.
Change-Id: I79a8b637d49b0e2d92b6dd5eb5aa8aa0520bf938
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4032
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Only reencoded MIPS32 instructions. Added some instructions
for crc code. Micromips isa in debug mode is only needed for
pic32mm cores. Pic32mz seems that only works with MIPS32
isa when in debug mode.
Change-Id: I07059e153a7000ea9204f20b6b37edf6a7623455
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4022
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
This option is needed, for example, when exiting
debug mode in bmips targets. The last instruction
is a NOP, not a DERET. When working in async mode
this check is not done, mips32_pracc_queue_exec() pass
the parameter to mips32_pracc_exec() and never use it.
Change-Id: I4c7ed4feb1588b62e2645b955b501b6671113b36
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4021
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
If max_code is reached realloc memory. If fails to realloc
the error is propagated and every call to pracc_add() returns
immediately. The exec function logs the error.
Change-Id: Idd4ed9d9b8b19b7d6842d0bc5ebb05f943726705
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4020
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
pracc_list points to an array with code in the lower half
and addr in the upper half. Change it to a struct with
an instruction field and an address field.
Requiered to make reallocation easier.
As a side effect the code is less quirky.
Change-Id: Ibf904a33a2f35a7f69284d2a2114f4b4ae79219f
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4019
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
MacOSX tool chain defines __unused in "sys/cdefs.h", causing a collision.
Remove the local define to avoid polluting the compilers internal
symbol namespace.
Change-Id: I16370c4518e6aeec482dd689e7db80628f846ee3
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4118
Reviewed-by: Steven Stallion <sstallion@gmail.com>
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Tested-by: jenkins
I'm not sure why this is necessary, but for some reason GDB is asking
for registers before OpenOCD thinks there's been a halt. This is really
just a workaround, but I need to refactor the v0.11 stuff anyway so I
don't want to figure it out.
This is a major rewrite of the RISC-V v0.13 OpenOCD port. This
shouldn't have any meaningful effect on the v0.11 support, but it does
add generic versions of many functions that will allow me to later
refactor the v0.11 support so it's easier to maintain both ports. This
started as an emergency feature branch and went on for a long time, so
it's all been squashed down into one commit so there isn't a big set of
broken commits lying around. The changes are:
* You can pass "-rtos riscv" to the target in OpenOCD's configuration
file, which enables multi-hart mode. This uses OpenOCD's RTOS
support to control all the harts from the debug module using commands
like "info threads" in GDB. This support is still expermental.
* There is support for RV64I, but due to OpenOCD limitations we only
support 32-bit physical addresses. I hope to remedy this by rebasing
onto the latest OpenOCD release, which I've heard should fix this.
* This matches the latest draft version of the RISC-V debug spec, as of
April 26th. This version fixes a number of spec bugs and should be
close to the final debug spec.
Less code and probably cleaner.
Don't check if it is ever ERROR_OK.
Change-Id: I1045b58fd4542ec24430332f49679364ae97b1dc
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4018
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Implement it as a function, the code was already in. Added optimize
option.
Change-Id: Ib9ad3f00d6c4f0b91c4e4960a50ec8d102f4e333
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4017
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
In most of the cases there is no need to request execution,
the check for a new pracc access already does it.
Requesting execution if not needed makes execution slower and
code larger due the additional checks.
Reduce code in fasdata transfer function.
Call for execution when exiting debug.
Change-Id: I3b45f6d1f62da5fad3e3db84f82a9299b16e1bd9
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4010
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Struct mips32_pracc_context no more in use.
In current code cp0 reg/sel do not requires special handling.
In sync mode ctx.store_count not used, drop check.
In fasdata transfer function use mips32_pracc_read_ctrl_addr()
to reduce code.
Change-Id: Ibd4cfa5a44ebc106ed0db042f4e54a2e0b3d43cb
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4007
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Will be used later, allow queuing all needed scans in a pracc
access. This makes faster execution with ftdi based adapters
working in sync with pracc.
Added now because the overall code is shorter.
Change-Id: Ib32b89307b75785f88870db8d7c9255dc5bbd426
Signed-off-by: Salvador Arroyo <sarroyofdez@yahoo.es>
Reviewed-on: http://openocd.zylin.com/4005
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
The halt event was left pending in the CTI, better to clear it immediately
after debug entry.
Change-Id: I6002f862681baf98769e3c73332a7f7f0ef938c1
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4030
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
read_memory was doing autoexec-initialized commands, but wasn't using
ac_busy_delay to add delay or wait between the commands. This isn't the optimal
solution, but the whole read_memory sequence needs to be made more efficient
anyway, so this was the quick and dirty solution.
When ARM64 support was being merged, a comparison ended up being
inverted. This causes NULL pointer access when target attempts to
use core cache.
Change-Id: Ic8873ddd13dbdd8100856a71b4717f44cd336e23
Signed-off-by: Girts Folkmanis <opensource@girts.me>
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/4042
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Replace the constants with the ones from the ISA spec, since those are
not updated as often.
Also delete a bunch of old code.
Change-Id: I9201b1455d64a9d2d203bb362fefaa68cbf35aeb
after handling of an exception in debug state, immediately
restore the original core state.
Change-Id: Ie53b63c9f19815f717f4df4390fbc13f0a204cc2
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3996
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
to speed up debugging, don't load the complete register context
on a halt event, load only those registers that might be
clobbered during debugging.
Change-Id: I0b58e97aad6f28aefce4a52e870af61e1ef1a44f
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3995
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
small changes to correct code formatting and spelling of some
log messages.
Change-Id: I645e675f8f9f4731b0271ddc55f64e8cf56ec1db
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3994
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
This patch contains a major overhaul of the target run control,
mainly for the sake of satisfying gdbs ideas of how a target
should respond to various control requests for the debugger.
The changes allow gdb a slightly better control on how cores
are stepped: a core can be single-stepped while
other cores remain halted or continue normal execution
until the single-stepped core halts again.
Also, on any halting event (user command or breakpoint) the
system is brought into a stable state with all cores halted
before the halt is signaled to the debugger.
This patch also transitions the target code to make use of the
new CTI abstraction instead of accessing CTI registers directly.
Change-Id: I8ddc9abb119e04580d671b57ee12240c3f5070a0
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3993
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
- rename "cortex_a" command group to "aarch64"
- remove default blank check, checksum and algorithm hooks
since they're not going to work in aarch64 mode anyway.
Change-Id: Ieb0046786ed9425baf6774c68f42a8285cc2aefd
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3991
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Make sure all core register caches are invalidated on reset
assert, make sure to re-init debug registers on deassert.
Change-Id: I82350d04cc3eaae5e35245d13d6c1fb0a8d59807
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3990
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
introduce armv8_set_dbgreg_bits() function to make register
bit-field modifications easier to read.
Change-Id: I6b06f66262587fd301d848c9e0645e8327653de7
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3989
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
The aarch32 register cache is only a separate view of the aarch64
registers. Load aarch32 registers through their aarch64 equivalents.
Change-Id: I3e932dfb782f03d73d30d942b24db340a5749e47
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3988
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Mmu faults can not be prevented on aarch64, they need to be taken and
handled accordingly. Remove the remaining stub code.
Change-Id: I6241efa594fe6b963624f9628cdf1c8e46588223
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3987
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Not specific to ARMv8, the Cross Trigger Interface
deserves an independent access wrapper.
Change-Id: I84f8faad15ed3515e0fff7f6cc5d1109ef91a869
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3986
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Replace loop by right-shift.
Inspired by patch from Alamy Liu
Change-Id: I1285f4f54c0695a93fa42e9863ed8ffa4de00f70
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3985
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Suppress some very verbose LOG_DEBUG's that are not really useful
any more.
Change-Id: I67f10ba9510a9e34a027f378f4b62b8901ddc8a4
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3984
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Reduce SLOCs in SCTLR retrieval and modification functions and make them
less complex.
Change-Id: Ida1a99c223743247f171b52eef80dc9886802101
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3982
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Use the correct opcode for Aarch32 state, both for the breakpoint
instruction itself and the cache handling functions.
Change-Id: I975fa67b1e577b54f5c672a01d516419c6a614b2
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Reviewed-on: http://openocd.zylin.com/3981
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
The remote bitbang implementation is really slow when reading back data.
During many operations, like writing a block of memory, we don't need to
see the return data. By communicating this to the lower layers, we get
huge speedups. Downloads to spike now are 295KB/s.
That means the gdbserver.py download test now runs to completion,
unfortunately it fails. Everything else is still passing, though, so I'm
committing this change.
Change-Id: I44cc9db1ade0908c1a12d09b23fc8e529a802d88
Not a great fix. There's still a problem with accessing 64-bit floating
point registers on 32-bit cores.
24 of the gdbserver.py tests pass now.
Change-Id: I69a88ef5fd5581e2c7bf1d78057fd474ae86ff93
The code doesn't do anything intelligent if errors are found.
But MemTestBlock now also passes. I'm not quite sure why.
Change-Id: I8512f0a96db9e34d3db6a4a9bcef6e56f191d4c1
This includes reading GPRs (although I haven't confirmed the values) and
doing some CSR reading/writing to disable triggers that may be left over
from a previous setting.
Change-Id: I2c627bd002d601e302a40f838087541897c025fd
At framework level, the resume hook is not protected. Make sure to
not attempt a resume if the target is not halted.
Change-Id: I4dd1975a95d6c513bd4f4e999e496bc11182a97a
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Basically port a fix that was already done for the cortex_a target.
Change-Id: I4cf4519159bda03ed611bc0b2e340a5dad2d85fe
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Instead of supplying a local, preinitialized "dscr" variable, use the
cached value from arm_dpm, which is kept up-to-date anyway.
Change-Id: I06d548d4dc6db68b9d984c83ed026fa9069d7875
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
arm commands are mostly unusable anyway, remove them. to be replaced
by aarch64 specific commands later
Change-Id: Ie994771bc0e86cff1c26f68f1f51ce8ec352a509
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
"currentel" special register is not accessible in debug state.
Change-Id: I9022b01b423cd9ae8227ed018d6166078ba44832
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Make debug and error messages more informative, fix spelling and
formatting errors
Change-Id: I7245f42c5153bcc95676270814d30e91c113aaed
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
struct arm_reg::value[] must be 8 byte to hold a 64bit register value.
Change-Id: If253e90731d0ee855eafd9d7b63b91f84630cc7c
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Move clearing of DSCR "Sticky Error" condition to the
exception handling function. Clear once on entering debug state.
Change-Id: Iec1d09d6f2d9cdd7e92953da5ea19f3e399ca12c
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
recommended for Corte-A8 cores, not sure if necessary
for ARMv8 based cores as well.
Change-Id: Ibcb36170c5fac6a6b132de17f734c70a56919f9b
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Remove register cache invalidation and target state changes that are
handled appropriately in other functions.
Change-Id: Ic903f41ddc267f4b8765ea022bd4d6da1017e21f
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
When a PE is in Aarch32 state and gdb asks for a target description,
provide a register view compatible with the "org.gnu.gdb.arm.core"
feature. Only current-mode registers are exported, banked registers are
not visible.
Change-Id: I99a85d94831cf597fe8cff6a0a1818ce0a33613b
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
There's no access to TTBR in EL0. Circumvent by moving the PE to EL1
before reading, and switch back to original mode afterwards.
Change-Id: I22891b958d3d7e6fad1cb27183c192d975d63d89
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
When halting a group of PEs through CTI, HDE must be set in EDSCR for
all of them.
Change-Id: Iaa4bc0b0fe31e46a463c709d8274023225affd85
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
When an armv8-a PE causes an exception while halted, e.g. by performing
a prohibited memory or register access, its state is affected in the
same way as if it was running. That means, a number of registers is
overwritten (notably DLR and DSPSR, but also others) and also
potentially the exception level and therefore also the PE state can
change. This state must be restored before resuming normal operation.
This is done by marking the relevant cached registers "dirty" so that
they are written back before resume.
Change-Id: I9b6967a62d7cb23a477a9f7839f8d2b7087eed09
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Use proper T32 opcodes for cache identification when the PE is in
Aarch32 state
Change-Id: I9cd9169409889273a3fd61167f388e68d8dde86d
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
DCPS only allows to enter higher ELs, for lower ELs you need to
use DRPS. Also, of course the encoding differs between A64 and T32.
Both DCPS and DRPS also clobber DLR and DSPSR, which then need to be
restored on resume.
Change-Id: Ifa3dcfa94212702e57170bd59fd0bb25495fb6fd
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
All register access is now performed through common read/write
functions, which delegate the actual register access to the
armv8_common object. armv8_common contains function pointers
to direct read and write requests to the respective low-level
functions for each PE state.
The respective read/write functions are selected on debug state
entry.
At the same time, T32 opcodes are now formatted for ITR in
dpmv8_exec_opcode() and the T32_FMTITR macro is removed from global
visibility.
Change-Id: I9eaef017c7cc9e0c531e693c534901bfdbdb842c
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Read registers based on current EL instead of PE mode.
Change-Id: I05d3219ac1bf8585e9f4f024a7e8599fea0913b6
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
There's no access to system control register in EL0. Circumvent by
moving the PE to EL1 before reading, and switch back to original mode
afterwards.
Change-Id: I309f4eea5597ffc88fc892e9bbb826982e8a44ec
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Translate from cpsr value to "enum arm_mode" by shifting up 4 bits and
filling the lowest nibble with 0xF.
Change-Id: Ic32186104b0c29578c4f6f99e04840ab88a0017b
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Use AT commands to translate virtual to physical addresses based on
current MMU configuration.
Change-Id: I1bbd7d674c435541b617b17022fa9f7f0f01bdab
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Set HDE bit through helper function instead of manual mem_ap access.
Change-Id: I68c157870f3f3c47a875d425ade6e975d8075424
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
The artificial "os_border" doesn't exist in aarch64 state and is wrong
for aarch32 state as well. Remove it.
Change-Id: I7c673a1404b03aa78dbd505e115fa3a93f7ca05f
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Update cache identification to match functionality present in
armv7a_cache.c
Change-Id: I2dc4bee80f5a22b8728334d40331c183d1406f27
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Move all DPM related functions from aarch64.c to armv8_dpm.c.
Change-Id: I43404ff5db414ae898787a523d3219e5bee44889
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Add database for common, equivalent opcodes for Aarch32 and
Aarch64 execution states
Revisit all functions that access Aarch64 specific registers
or use Aarch64 opcodes and rewrite them to act depending on
current state of the core.
Add core register access functions for Aarch32 state
Add function to determine the core execution state without
reading DSPSR.
Change-Id: I345e9f6d682fb4ba454e4b1d16bb5e1b27570691
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Set up CTI so that halt and resume requests get routed to all PEs in the
SMP group.
Change-Id: Ie92cfd3fe54632e5fdc049a6bf5b24b99451a8c9
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
For now only D-Cache flush (Clean&Invalidate) and I-Cache
invalidate are implemented. That's enough for software breakpoints.
Change-Id: I8e96d645a230b51e3490403f4564e59ba6a76cf3
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
On live hardware, interrupts will happen while the core is
held for stepping. The next step will most of the time execute an
interrupt service instead of the next line of code, which is not
what you expect. Disable interrupts through DSCR before resuming
for a step, and re-enable them again after the step happened.
This should be made configurable, like on cortex_a target.
Change-Id: I94d8ffb58cf7579dedb66bc756b7eb6828b6e8e4
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
External debuggers need to use HLT, not BRK. HLT generates a halting
debug event while BRK generates a debug exception for self-hosted
debugging.
Change-Id: I24024b83668107f73a14cc75d951134917269e5c
Signed-off-by: Matthias Welwarsky <matthias.welwarsky@sysgo.com>
Karl Palsson
Liviu Ionescu
Tim Newsome
Palmer Dabbelt
Megan Wachs
Gleb Gagarin
Andreas Fritiofson
Dmitry Ryzhov
Tobias Diedrich
Salvador Arroyo
Salvador Arroyo
Matthias Welwarsky
Marc Schink
Girts Folkmanis