probe_rs/vendor/infineon/sequences/xmc4000.rs
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467
//! Sequences for XMC4000
use crate::architecture::arm::armv7m::{Aircr, Dhcsr, FpCtrl, FpRev1CompX, FpRev2CompX};
use crate::architecture::arm::memory::ArmMemoryInterface;
use crate::architecture::arm::sequences::{ArmDebugSequence, ArmDebugSequenceError};
use crate::architecture::arm::ArmError;
use std::sync::Arc;
use std::sync::Mutex;
use std::thread;
use std::time::{Duration, Instant};
use crate::architecture::arm::communication_interface::DapProbe;
use crate::{probe::DebugProbeError, MemoryMappedRegister};
/// An Infineon XMC4xxx MCU.
#[derive(Debug)]
pub struct XMC4000 {
reset_state: Mutex<Option<ResetState>>,
}
impl XMC4000 {
/// Create the sequencer for an Infineon XMC4000.
pub fn create() -> Arc<dyn ArmDebugSequence> {
Arc::new(Self {
reset_state: Mutex::new(None),
})
}
}
#[derive(Debug)]
enum ResetState {
/// We are commanding a "Warm Reset Halt", per XMC4700/XMC4800 reference manual v1.3 § 28.4.2.
///
/// This variant contains the state we clobbered while setting a breakpoint to halt on the jump
/// to user code.
Warm { fpctrl_enabled: bool, fpcomp0: u32 },
/// We are commanding a "Cold Reset Halt", per XMC4700/XMC4800 reference manual v1.3 § 28.4.2.
///
/// This variant has no state to track.
Cold { tssw_start_at: Instant },
}
impl Default for ResetState {
fn default() -> Self {
ResetState::Warm {
fpctrl_enabled: false,
fpcomp0: 0,
}
}
}
bitfield::bitfield! {
/// SCU->STCON startup configuration register, as documented in XMC4700/XMC4800 reference
/// manual v1.3 § 11-73.
#[derive(Copy,Clone)]
pub struct Stcon(u32);
impl Debug;
/// > HW Configuration
/// > At PORESET the following values are latched:
/// > HWCON.0 = not (TMS)
/// > HWCON.1 = TCK
/// >
/// > 0b00B: Normal mode, JTAG
/// > 0b01B: ASC BSL enabled
/// > 0b10B: BMI customized boot enabled
/// > 0b11B: CAN BSL enabled
pub hwcon, _: 1, 0;
/// > SW Configuration
/// > Bit[9:8] is copy of Bit[1:0] after PORESET
/// > …
/// > Note: Only reset with Power-on Reset
pub swcon, set_swcon: 11, 8;
}
impl Stcon {
const ADDRESS: u64 = 0x50004010;
}
impl ArmDebugSequence for XMC4000 {
// We have a weird halt-after-reset sequence. It's described only in prose, not in a CMSIS pack
// sequence. Per XMC4700/XMC4800 reference manual v1.3 § 28-8:
//
// > A Halt after system reset (Warm Reset) can be achieved by programming a break point at the
// > first instruction of the application code. Before a Warm Reset the CTRL/STAT.CDBGPWRUPREQ
// > and the DHCSR.C_DEBUGEN setting has to be ensured. After a system reset, the HAR situation
// > is not considered, as the reset is not coming from "Power-on Reset".
// >
// > Note: The CTRL/STAT.CDBGPWRUPREQ and DHCSR.C_DEBUGEN does not have to be set after a
// > system reset, if they have already been set before.
// >
// > A tool hot plug condition allows to debug the system starting with a tool registration
// > (setting CTRL/STAT.CDBGPWRUPREQ register) and a debug system enable (setting the
// > DHCSR.C_DEBUGEN register). Afterwards break points can be set or the CPU can directly by
// > HALTED by an enable of C_HALT.
//
// So:
// * ResetCatchSet must determine the first user instruction and set a breakpoint there.
// * ResetCatchClear must restore the clobbered breakpoint, if any.
fn reset_catch_set(
&self,
core: &mut dyn ArmMemoryInterface,
core_type: probe_rs_target::CoreType,
debug_base: Option<u64>,
) -> Result<(), ArmError> {
tracing::trace!("performing XMC4000 ResetCatchSet");
// Did we just come out of a cold reset?
if let Some(tssw_start_at) = self
.reset_state
.lock()
.map(|v| match v.as_ref() {
Some(ResetState::Cold { tssw_start_at }) => Some(*tssw_start_at),
_ => None,
})
.unwrap()
{
// The core will halt itself after SSW execution _if_ we connected fast enough
// Attempt to determine how much time passed since the system software started running
let tssw_elapsed = tssw_start_at.elapsed();
if tssw_elapsed > Duration::from_micros(2_500) {
// We might have missed the window
tracing::warn!(
"{:?} has elapsed since power on reset, exceeding typical tSSW of 2.5ms",
tssw_elapsed
);
}
// See if we halted
match spin_until_core_is_halted(core, Duration::from_millis(3)) {
Err(ArmError::Timeout) => {
// We missed the boat
tracing::info!("Core did not halt after cold boot; performing a warm reset");
// Clear our cold boot status, since we failed
self.reset_state.lock().map(|mut s| s.take()).unwrap();
// Perform a warm reset
self.reset_catch_set(core, core_type, debug_base)?;
self.reset_system(core, core_type, debug_base)?;
}
Err(e) => return Err(e),
Ok(()) => {
// We halted after cold reset, which is what our caller wanted
// Job done
return Ok(());
}
}
}
// We need to find the "first instruction of the application code". The XMC4xxx system
// software ("SSW") has a vector table at the start of boot ROM. XMC4700/XMC4800 reference
// manual v1.3 § 2-31:
//
// > On system reset, the vector table is fixed at address 0x00000000.
//
// That reset vector runs first, but it doesn't qualify as "application code". § 27.2:
//
// > This section describes the startup sequence of the XMC4[78]00 as a process taking place
// > before user application software takes control of the system.
//
// So: where does the application begin? That depends on the boot mode! The boot mode
// depends on the System Control Unit (SCU) Startup Control (STCON) register, which includes
// both a software-settable field and a power-on-resettable field. We want a normal boot,
// because normal boots are sane and reasonable.
//
// Read STCON.
let stcon = Stcon(core.read_word_32(Stcon::ADDRESS)?);
// The software-settable SWCON field is authoritative for system resets, per § 27.2.3:
//
// > HWCON bit field is read only for PORST (Power ON Reset). For every other reset type
// > (available in SCU_RSTSTAT) register, the SWCON field is assessed.
//
// Set it to a normal boot if needed.
if stcon.swcon() != 0 {
let mut stcon = stcon;
stcon.set_swcon(0);
core.write_word_32(Stcon::ADDRESS, stcon.0)?;
}
// § 27.3.1 describes the normal boot mode, which happens after firmware initialization:
//
// > Firmware essentially reprograms the Cortex M4’s SCB.VTOR register with the start
// > address of flash (0C000000H) and passes control to user application by programing
// > register R15 (Program Counter) with reset vector contents. The reset vector contents
// > point to a routine that could be in either the cached or the uncached address space of
// > the flash.
//
// Therefore, "the first instruction of the application code" for normal boot mode is the
// instruction pointed to by the reset vector at the start of flash, i.e. 0x0C000004.
//
// This is also why we have to use a breakpoint instead of trapping the reset vector. The
// application's entrypoint is a normal jump from the firmware, not an exception dispatched
// to the reset vector.
let application_entry = core.read_word_32(0x0C000004)? - 1;
// Read FP state so we can restore it later
let fp_ctrl = FpCtrl(core.read_word_32(FpCtrl::get_mmio_address())?);
let fpcomp0 = core.read_word_32(FpRev1CompX::get_mmio_address())?;
// Indicate we're in the midst of a warm reset
self.reset_state
.lock()
.map(|mut m| {
m.replace(ResetState::Warm {
fpctrl_enabled: fp_ctrl.enable(),
fpcomp0,
})
})
.unwrap();
// Enable FP
let mut fp_ctrl = FpCtrl(0);
fp_ctrl.set_enable(true);
fp_ctrl.set_key(true);
core.write_word_32(FpCtrl::get_mmio_address(), fp_ctrl.into())?;
// Set a breakpoint at application_entry
let val = if fp_ctrl.rev() == 0 {
FpRev1CompX::breakpoint_configuration(application_entry)?.into()
} else if fp_ctrl.rev() == 1 {
FpRev2CompX::breakpoint_configuration(application_entry).into()
} else {
return Err(ArmDebugSequenceError::custom(format!(
"xmc4000: unexpected fp_ctrl.rev = {}",
fp_ctrl.rev()
))
.into());
};
core.write_word_32(FpRev1CompX::get_mmio_address(), val)?;
tracing::debug!("Set a breakpoint at {:08x}", application_entry);
core.flush()?;
Ok(())
}
fn reset_catch_clear(
&self,
core: &mut dyn ArmMemoryInterface,
_core_type: probe_rs_target::CoreType,
_debug_base: Option<u64>,
) -> Result<(), ArmError> {
tracing::trace!("performing XMC4000 ResetCatchClear");
// Grab the prior breakpoint state
let reset_state = self
.reset_state
.lock()
.map(|mut m| m.take().unwrap_or_default())
.unwrap();
match reset_state {
ResetState::Warm {
fpctrl_enabled,
fpcomp0,
} => {
// Put FPCTRL back
let mut fpctrl = FpCtrl::from(0);
fpctrl.set_key(true);
fpctrl.set_enable(fpctrl_enabled);
core.write_word_32(FpCtrl::get_mmio_address(), fpctrl.into())?;
// Put FPCOMP0 back
core.write_word_32(FpRev1CompX::get_mmio_address(), fpcomp0)?;
}
ResetState::Cold { .. } => {
// No op
}
}
Ok(())
}
fn reset_system(
&self,
core: &mut dyn ArmMemoryInterface,
_core_type: probe_rs_target::CoreType,
_debug_base: Option<u64>,
) -> Result<(), ArmError> {
// XMC4700/XMC4800 reference manual v1.3 § 27.2.2.2:
// > Since the Reset Status Information in register SCU.RSTSTAT is the accumulated reset
// > type, it is necessary to clean the bitfield using the SCU register RSTCLR.RSCLR before
// > issuing a System Reset, else SSW will enter the boot mode reflected in STCON.HWCON.
//
// ARM KB https://developer.arm.com/documentation/ka002435:
// > After the user program has handled the the reset sources itself, it needs to clear them
// > by executing:
// >
// > SCU_RESET->RSTCLR = SCU_RESET_RSTCLR_RSCLR_Msk ;
// >
// > before the next reset. This will clear also the PORST bit. And if the next reset is no
// > Power On Reset, the SSW will also not do HAR.
//
// This is normally handled by the runtime, but let's be defensive.
//
// SCU_RESET->RSTCLR is at 0x5000_4408, and RSCLR is the low bit.
core.write_word_32(0x5000_4408, 1)?;
tracing::debug!("Cleared SCU_RESET->RSTSTAT");
let mut aircr = Aircr(0);
aircr.vectkey();
aircr.set_sysresetreq(true);
core.write_word_32(Aircr::get_mmio_address(), aircr.into())?;
tracing::debug!("Resetting via AIRCR.SYSRESETREQ");
// Spin until CoreSight indicates the reset was processed
let start = Instant::now();
loop {
let dhcsr = Dhcsr(core.read_word_32(Dhcsr::get_mmio_address())?);
// Wait until the S_RESET_ST bit is cleared on a read
if !dhcsr.s_reset_st() {
tracing::debug!("Detected reset via S_RESET_ST");
break;
}
if start.elapsed() > Duration::from_millis(500) {
tracing::error!("XMC4000 did not reset as commanded");
return Err(ArmError::Timeout);
}
}
// Now that we commanded a reset, we've lost access to everything but CoreSight.
// XMC4700/XMC4800 reference manual v1.3 § 28-7:
//
// > For security reasons it is required to prevent a debug access to the processor before
// > and while the boot firmware code from ROM (SSW) is being executed. A bit DAPSA, (DAP
// > has system access) in the SCU is implemented, allowing the access from CoreSight™ debug
// > system to the processor core. The default value of this bit is disabled debug access.
// > The register is reset by System Reset. The System Reset disables the debug access each
// > time SSW is being executed. At the end of the SSW the DAPSA is enabled always
// > (independent of any other register setting or signaling value), to allow debug access
// > to the CPU. A tool accessing the SoC during the SSW execution time reads back a zero
// > and a write is going to a virtual, none existing address.
//
spin_until_dapsa_is_clear(core)?;
// If we are intending to halt after reset, spin and wait for that here
if self.reset_state.lock().map(|v| v.is_some()).unwrap() {
tracing::debug!("Waiting for XMC4000 to halt after reset");
// We're doing a halt-after-reset
// Wait for the core to halt
spin_until_core_is_halted(core, Duration::from_millis(1000))?;
} else {
tracing::debug!("not performing a halt-after-reset");
}
Ok(())
}
fn reset_hardware_assert(&self, interface: &mut dyn DapProbe) -> Result<(), ArmError> {
tracing::trace!("performing XMC4000 ResetHardwareAssert");
use crate::architecture::arm::Pins;
// We want to drive nRST, TCK, and TMS
let mut pin_select = Pins(0);
pin_select.set_nreset(true);
pin_select.set_swclk_tck(true);
pin_select.set_swdio_tms(true);
// We want to drive nRST low to command the reset
let mut pin_output = Pins(0);
pin_output.set_nreset(false);
// HWCON is latched at power-on reset to be [TCK, !TMS], and we want HWCON to be zero, so
// we want to drive TCK low and TMS high.
pin_output.set_swclk_tck(false);
pin_output.set_swdio_tms(true);
loop {
match interface.swj_pins(pin_output.0 as u32, pin_select.0 as u32, 0) {
Err(DebugProbeError::CommandNotSupportedByProbe {
command_name: "swj_pins",
}) if pin_select.swdio_tms() => {
// J-Link probes return this error when we try to set pins besides nRST
// Settle for resetting, but warn the user that HWCON is uncontrolled
tracing::debug!("swj_pins(nRST|TCK|TMS) unsupported by probe; falling back to swj_pins(nRST)");
tracing::warn!(
"This probe cannot manipulate HWCON, so the boot mode after power on reset cannot be controlled"
);
pin_select = Pins(0);
pin_select.set_nreset(true);
}
Err(other) => return Err(other.into()),
Ok(_) => break,
}
}
// When we return, the caller will to attach the probe and set DEBUGEN. XMC4000 needs that
// to happen while the SSW is booting, _after_ releasing nRST. This means we need to
// deassert nRST here rather than waiting for ResetHardwareDeassert.
//
// Wait a moment for the reset signal to settle.
thread::sleep(Duration::from_millis(100));
// Indicate to ourselves that we're doing a cold reset, and that the system software began
// executing now
self.reset_state
.lock()
.map(|mut s| {
s.replace(ResetState::Cold {
tssw_start_at: Instant::now(),
})
})
.unwrap();
// Deassert nRST
pin_output.set_nreset(true);
interface.swj_pins(pin_output.0 as u32, pin_select.0 as u32, 0)?;
// Race! :(
Ok(())
}
fn reset_hardware_deassert(&self, memory: &mut dyn ArmMemoryInterface) -> Result<(), ArmError> {
tracing::trace!("performing XMC4000 ResetHardwareDeassert");
// We already deasserted nRST in ResetHardwareAssert, because that's how Cold Reset Halts
// work on this platform.
// We should however wait until the SSW is ready.
spin_until_dapsa_is_clear(memory)?;
Ok(())
}
}
fn spin_until_core_is_halted(
core: &mut dyn ArmMemoryInterface,
timeout: Duration,
) -> Result<(), ArmError> {
let start = Instant::now();
loop {
let dhcsr = Dhcsr(core.read_word_32(Dhcsr::get_mmio_address())?);
if dhcsr.s_halt() {
tracing::debug!("Halted after reset");
return Ok(());
} else if start.elapsed() > timeout {
tracing::error!("XMC4000 did not halt after reset");
return Err(ArmError::Timeout);
}
}
}
fn spin_until_dapsa_is_clear(core: &mut dyn ArmMemoryInterface) -> Result<(), ArmError> {
let start = Instant::now();
loop {
// DAPSA isn't directly accessible because of course it isn't.
//
// Read the SCU module ID register, which is guaranteed to be nonzero. If DAPSA is set,
// we'll read it normally, and we can go on with our lives. If DAPSA is clear, we'll
// read a zero.
let scu_module_id = core.read_word_32(0x5000_4000)?;
if scu_module_id != 0 {
tracing::debug!("DAPSA is set");
break Ok(());
} else {
tracing::trace!("DAPSA is clear");
if start.elapsed() > Duration::from_millis(500) {
tracing::error!("timed out waiting for DAPSA to clear, indicating SSW hang");
break Err(ArmError::Timeout);
}
}
}
}