Mercurial > nand-ecoscentric
annotate packages/hal/sparclite/arch/current/ChangeLog @ 6:d376b777e2ce ecos-sw-1999-05-14
Merge from eCos master repository on 1999-05-14-19:27:43-BST
| author | jlarmour |
|---|---|
| date | Fri, 14 May 1999 12:20:00 +0000 |
| parents | 1d7f19c9e4d1 |
| children | ece80412419a |
| rev | line source |
|---|---|
|
6
d376b777e2ce
Merge from eCos master repository on 1999-05-14-19:27:43-BST
jlarmour
parents:
4
diff
changeset
|
1 1999-05-13 Hugo Tyson <hmt@cygnus.co.uk> |
|
d376b777e2ce
Merge from eCos master repository on 1999-05-14-19:27:43-BST
jlarmour
parents:
4
diff
changeset
|
2 |
|
d376b777e2ce
Merge from eCos master repository on 1999-05-14-19:27:43-BST
jlarmour
parents:
4
diff
changeset
|
3 * include/hal_arch.h (HAL_THREAD_INIT_CONTEXT): Align stack |
|
d376b777e2ce
Merge from eCos master repository on 1999-05-14-19:27:43-BST
jlarmour
parents:
4
diff
changeset
|
4 (rather conservatively) before use. |
|
d376b777e2ce
Merge from eCos master repository on 1999-05-14-19:27:43-BST
jlarmour
parents:
4
diff
changeset
|
5 |
|
4
1d7f19c9e4d1
Merge from eCos master repository on 1999-05-11-21:11:10-BST
jlarmour
parents:
2
diff
changeset
|
6 1999-04-28 Bart Veer <bartv@cygnus.co.uk> |
|
1d7f19c9e4d1
Merge from eCos master repository on 1999-05-11-21:11:10-BST
jlarmour
parents:
2
diff
changeset
|
7 |
|
1d7f19c9e4d1
Merge from eCos master repository on 1999-05-11-21:11:10-BST
jlarmour
parents:
2
diff
changeset
|
8 * src/PKGconf.mak: |
|
1d7f19c9e4d1
Merge from eCos master repository on 1999-05-11-21:11:10-BST
jlarmour
parents:
2
diff
changeset
|
9 Remove the -n argument to tail, it does not appear to be required |
|
1d7f19c9e4d1
Merge from eCos master repository on 1999-05-11-21:11:10-BST
jlarmour
parents:
2
diff
changeset
|
10 on any supported host and causes problems with some |
|
1d7f19c9e4d1
Merge from eCos master repository on 1999-05-11-21:11:10-BST
jlarmour
parents:
2
diff
changeset
|
11 implementations of tail. |
|
1d7f19c9e4d1
Merge from eCos master repository on 1999-05-11-21:11:10-BST
jlarmour
parents:
2
diff
changeset
|
12 |
| 2 | 13 1999-04-15 Jonathan Larmour <jlarmour@cygnus.co.uk> |
| 14 | |
| 15 * src/sparclite.ld: Define __bss_end at the end of the BSS | |
| 16 * src/hal_boot.c (hal_zero_bss): Stop at __bss_end rather than _end | |
| 17 when clearing BSS | |
| 18 | |
| 19 These fix PR 19750 | |
| 20 | |
| 21 1999-04-14 Jonathan Larmour <jlarmour@cygnus.co.uk> | |
| 22 | |
| 23 * src/PKGconf.mak (EXTRAS): Don't generate extras.o here any more | |
| 24 But do define EXTRAS every time for the linker script. | |
| 25 | |
| 26 1999-04-09 Hugo Tyson <hmt@cygnus.co.uk> | |
| 27 | |
| 28 * include/hal_intr.h (HAL_RESTORE_INTERRUPTS et al): Put in 3 nops | |
| 29 after any writes to the psr; TBH I don't think they're strictly | |
| 30 necessary when not altering the window state, but for complete | |
| 31 safety and adherence to the book of words.... | |
| 32 | |
| 33 1999-04-08 John Dallaway <jld@cygnus.co.uk> | |
| 34 | |
| 35 * src/*.ld: Revised SECTION_* macro arguments to | |
| 36 avoid padded output sections (PR 19787) | |
| 37 | |
| 38 1999-04-07 Hugo Tyson <hmt@cygnus.co.uk> | |
| 39 | |
| 40 * include/hal_io.h: If not the simulator, get hal_hwio.h too - | |
| 41 which describes board-and-CPUvariant-specific IO needs. | |
| 42 | |
| 43 1999-04-07 Hugo Tyson <hmt@cygnus.co.uk> | |
| 44 | |
| 45 * include/hal_intr.h (HAL_VSR_SET_TO_ECOS_HANDLER): Define this | |
| 46 macro for use in tests that are interested in exception diddling. | |
| 47 Redefine CYGNUM_HAL_ISR_MIN to describe the number of interrupt | |
| 48 sources; while there are XSRs too, this definition is better for | |
| 49 the error checking in kernel interrupt stuff. Changed the symbols | |
| 50 used for defining tables sizes to match, since there are still 27 | |
| 51 VSRs and ISRs/XSRs and their data. | |
| 52 | |
| 53 * src/hal_intr.c: Use the correct symbol for the size of the | |
| 54 isr/vsr/their-data tables. | |
| 55 | |
| 56 1999-03-26 Hugo Tyson <hmt@cygnus.co.uk> | |
| 57 | |
| 58 * tests/sparc_ex.c (cyg_[user_]start): | |
| 59 If there is no kernel, cyg_user_start() is never called, main gets | |
| 60 in and steals the CPU. So we hang. So if there is no kernel, | |
| 61 define cyg_start() instead to wrest control. | |
| 62 | |
| 63 1999-03-22 Hugo Tyson <hmt@cygnus.co.uk> | |
| 64 | |
| 65 * include/hal_arch.h (CYGNUM_HAL_STACK_SIZE_TYPICAL): | |
| 66 And this time make the edit correctly ;-/ | |
| 67 | |
| 68 1999-03-22 Hugo Tyson <hmt@cygnus.co.uk> | |
| 69 | |
| 70 * include/hal_arch.h: | |
| 71 Use CYGNUM_HAL_STACK_SIZE_TYPICAL for the stack size instead of | |
| 72 CYGNUM_HAL_MINIMUM_STACK_SIZE. | |
| 73 | |
| 74 1999-03-17 Hugo Tyson <hmt@cygnus.co.uk> | |
| 75 | |
| 76 * src/vectors.S (rom_vectors_end, rom_vectors): Add two new | |
| 77 symbols for copying trampoline code into RAM in ROM startup. This | |
| 78 keeps ROM startup as similar as possible to RAM, for simplicity; | |
| 79 no re-ordering of vectors.S's contents. | |
| 80 | |
| 81 * src/hal_boot.c (hal_copy_data): Copy in long sized units rather | |
| 82 than long longs, better to cope with ROM data alignment. | |
| 83 Also add some LED output (if SLEB hardware) for startup to help | |
| 84 with debugging ROM start. | |
| 85 | |
| 86 1999-03-17 John Dallaway <jld@cygnus.co.uk> | |
| 87 | |
| 88 * src/PKGconf.mak: Remove dependence on echo '-e' switch. | |
| 89 | |
| 90 1999-03-16 Hugo Tyson <hmt@cygnus.co.uk> | |
| 91 | |
| 92 * src/vectors.S (genuine_reset): Add a little extra debug to SLEB | |
| 93 LEDs, and use the genuine top-of-memory stack when running on | |
| 94 hardware; only define cyg_bootup_stack in the simulator case. | |
| 95 | |
| 96 * src/sparclite.ld: Define rom_data_start, regardless of start | |
| 97 type, no harm is done. | |
| 98 | |
| 99 1999-03-12 Gary Thomas <gthomas@cygnus.co.uk> | |
| 100 | |
| 101 * include/pkgconf/hal_sparclite.h: | |
| 102 * include/hal_arch.h: Moved definition of 'CYGNUM_HAL_MINIMUM_STACK_SIZE' | |
| 103 | |
| 104 1999-03-11 Hugo Tyson <hmt@cygnus.co.uk> | |
| 105 | |
| 106 * src/icontext.c (cyg_hal_sparc_[gs]et_gdb_regs): | |
| 107 Two new routines to communicate our nice, optimized stack save | |
| 108 states to a reg list for GDB. My maximal register saves are not | |
| 109 really suitable for passing to GDB in any case, they're for humans | |
| 110 to debug the interrupt code more than anything else, and my | |
| 111 minimal save sets naturally contain pointers and ways of finding | |
| 112 the rest of the reg set, but it depends on whether it's an | |
| 113 interrupt save or a synchronous one, and given the SPARClite's | |
| 114 stack layout, it's a bit hairy. | |
| 115 | |
| 116 * src/vec_ivsr.S (hal_default_interrupt_vsr): | |
| 117 Do save %g0 even in a minimal context, so that the code for | |
| 118 communicating register state with GDB can tell it's a minimum | |
| 119 context. Costs one store cycle per interrupt. | |
| 120 | |
| 121 * include/hal_arch.h (HAL_[GS]ET_GDB_REGISTERS): | |
| 122 Define and call routines cyg_hal_sparc_[gs]et_gdb_regs for | |
| 123 communicating register state of a thread to GDB. | |
| 124 | |
| 125 * src/vectors.S (real_vector): Reinstate the anullment of the | |
| 126 instructions that were blamed for register corruption earlier; | |
| 127 they are correct and safe. | |
| 128 | |
| 129 1999-03-10 Jesper Skov <jskov@cygnus.co.uk> | |
| 130 | |
| 131 * include/hal_intr.h (HAL_INTERRUPT_IN_USE): Added. | |
| 132 | |
| 133 1999-03-09 Gary Thomas <gthomas@cygnus.co.uk> | |
| 134 | |
| 135 * src/hal_boot.c: Added missing <cyg/infra/cyg_type.h> include file. | |
| 136 | |
| 137 1999-03-05 Gary Thomas <gthomas@cygnus.co.uk> | |
| 138 | |
| 139 * src/sparclite.ld: | |
| 140 * src/PKGconf.mak: Clean up I/O package changes. | |
| 141 | |
| 142 1999-03-04 Jonathan Larmour <jlarmour@cygnus.co.uk> | |
| 143 | |
| 144 * src/sparclite.ld: | |
| 145 Unconditionalize use of libextras.a and add to GROUP() | |
| 146 | |
| 147 1999-03-04 Gary Thomas <gthomas@cygnus.co.uk> | |
| 148 | |
| 149 * src/PKGconf.mak: | |
| 150 * src/sparclite.ld: New I/O package support. | |
| 151 | |
| 152 * src/vectors.S: Fix register corruption problem - stems from | |
| 153 use of annulled instructions in window overflow/underflow code. | |
| 154 | |
| 155 1999-02-25 Hugo Tyson <hmt@cygnus.co.uk> | |
| 156 | |
| 157 These log entries from 1999-02-22 to 1999-02-24 were executed in | |
| 158 a branch; this change is a merge of that branch into the trunk. | |
| 159 | |
| 160 (1999-02-24 Hugo Tyson <hmt@cygnus.co.uk>) | |
| 161 | |
| 162 * tests/sparc_ex.c (sparc_ex_main): | |
| 163 For all traps except the USER_TRAP, install the default VSR | |
| 164 hal_default_exception_vsr(); this is so that this test catches | |
| 165 exception ie. is able to work, even when we are playing nice with | |
| 166 CygMon and GDB. The USER_TRAP is left so that breakpoints work, | |
| 167 for the test infrastructure. | |
| 168 | |
| 169 (1999-02-22 Hugo Tyson <hmt@cygnus.co.uk>) | |
| 170 | |
| 171 * src/vectors.S (real_vector): Remove old testing code. Add delay | |
| 172 NOPs after writes to %wim and %psr for safety. Leave installing | |
| 173 the trap vector to platform-specific code, along with RAM control | |
| 174 and cache initialization. | |
| 175 | |
| 176 * src/vec_ivsr.S (hal_default_interrupt_vsr): Add delay NOPs after | |
| 177 writing the %wim before doing a restore; that was a bug that was. | |
| 178 | |
| 179 * src/vec_xvsr.S (hal_default_exception_vsr): same change. | |
| 180 | |
| 181 * src/hal_intr.c: Remove platform-specific clock period variable; it | |
| 182 happens that both platforms sim and sleb have one, but it's not a | |
| 183 requirement of the target architecture. | |
| 184 | |
| 185 * src/hal_boot.c (cyg_hal_start): Move diddling the sim-specific | |
| 186 watchdog into platform-specific startup code viz. new routines | |
| 187 hal_board_{pre/post}start(). | |
| 188 | |
| 189 * src/context.S: a couple of extra NOPs to ensure that %wim and | |
| 190 %psr changes work on real hardware. | |
| 191 | |
| 192 (1999-02-22 Hugo Tyson <hmt@cygnus.co.uk>) | |
| 193 | |
| 194 * include/hal_arch.h (HAL_SET_GDB_REGISTERS): | |
| 195 Remove warning about unused var 'cos of the incompleteness of | |
| 196 these macros; gdb interworking is not yet supported. | |
| 197 | |
| 198 1999-02-25 Nick Garnett <nickg@cygnus.co.uk> | |
| 199 | |
| 200 * include/vectors.h: | |
| 201 Changed label used to access scheduler lock to one that is not | |
| 202 mangled by C++. This is intended to make support for interrupt | |
| 203 handling in non-kernel configurations easier. | |
| 204 | |
| 205 1999-02-20 Jonathan Larmour <jlarmour@cygnus.co.uk> | |
| 206 | |
| 207 * include/hal_arch.h: | |
| 208 Rename deliver_exception() -> cyg_hal_deliver_exception() | |
| 209 QA improvements | |
| 210 | |
| 211 * include/hal_intr.h: | |
| 212 Reorganise vector/interrupt/exception names according to purpose | |
| 213 Rename exception_handler() -> cyg_hal_exception_handler() | |
| 214 QA improvements | |
| 215 | |
| 216 * src/hal_intr.c: | |
| 217 Rename exception_handler() -> cyg_hal_exception_handler() | |
| 218 Rename deliver_exception() -> cyg_hal_deliver_exception() | |
| 219 Rename CYG_ISR_COUNT -> CYGNUM_HAL_ISR_COUNT | |
| 220 QA improvements | |
| 221 | |
| 222 * src/vec_ivsr.S: | |
| 223 * src/vec_xvsr.S: | |
| 224 Rename CYG_ISR_COUNT -> CYGNUM_HAL_ISR_COUNT | |
| 225 QA improvements | |
| 226 | |
| 227 * tests/sparc_ex.c: | |
| 228 Rename CYG_EXCEPTION_* -> CYGNUM_HAL_EXCEPTION_* | |
| 229 Rename CYG_VECTOR_UNALIGNED -> CYGNUM_HAL_VECTOR_UNALIGNED | |
| 230 | |
| 231 1999-02-12 Hugo Tyson <hmt@cygnus.co.uk> | |
| 232 | |
| 233 * src/context.S (hal_thread_load_context, hal_setjmp): | |
| 234 Disable interrupts (but not traps) while doing the flush of all my | |
| 235 callers and their callers' callers register sets to the stack. | |
| 236 There is evidence that the inconsistency in the stack during this | |
| 237 is a bad thing, though I have to admit that I don't see exactly | |
| 238 how it goes wrong. Partly, it's impossible to debug, so if the | |
| 239 problem I saw (a SEGV in instrumentation of interrupt_end()) is | |
| 240 reproduced with this change, I'll have more to go on. This is a | |
| 241 conservative change in that sense. | |
| 242 | |
| 243 * include/hal_arch.h (HAL_FLUSH_REGISTERS_TO_STACK): | |
| 244 Disable interrupts but not traps while doing this. Ditto. | |
| 245 | |
| 246 1999-02-12 Hugo Tyson <hmt@cygnus.co.uk> | |
| 247 | |
| 248 * include/pkgconf/hal_sparclite.h (CYGNUM_HAL_MINIMUM_STACK_SIZE): | |
| 249 Define a minimal sensible stack size for apps to use. Note that | |
| 250 this file is included by hal_arch.h and so is available to all HAL | |
| 251 users, and in fact almost everyone by default. | |
| 252 | |
| 253 1999-02-12 Hugo Tyson <hmt@cygnus.co.uk> | |
| 254 | |
| 255 * src/hal_intr.c (exception_handler): | |
| 256 Include <cyg/infra/cyg_ass.h> so that the CYG_FAIL() in some | |
| 257 configurations actually compiles (actually: links). | |
| 258 combo{22,18,10,6} were thus afflicted. | |
| 259 | |
| 260 1999-02-11 Hugo Tyson <hmt@cygnus.co.uk> | |
| 261 | |
| 262 * tests/sparc_ex.c (do_test): Make the test run for much less time | |
| 263 when CYG_USE_TRACING and in any case bleat periodically about how | |
| 264 successful it is to prevent testing system timeouts. Also move | |
| 265 the test routines (those which provoke the exception) to the end | |
| 266 so that (ever helpful) -O3 cannot inline them and then "optimize" | |
| 267 the result and thus spoil the whole damned test too. | |
| 268 | |
| 269 1999-02-08 John Dallaway <jld@cygnus.co.uk> | |
| 270 | |
| 271 * include/pkgconf/hal_sparclite.h: Correct capitalization of | |
| 272 'SPARClite'. | |
| 273 | |
| 274 1999-02-08 John Dallaway <jld@cygnus.co.uk> | |
| 275 | |
| 276 * src/sparclite.ld: New MLT-compatible linker script | |
| 277 * src/PKGconf.mak: Add build rules for sparclite.ld | |
| 278 | |
| 279 1999-02-04 Hugo Tyson <hmt@cygnus.co.uk> | |
| 280 | |
| 281 * src/vec_ivsr.S (hal_default_interrupt_vsr): Preserve global | |
| 282 registers over interrupts, and implement support for | |
| 283 CYGDBG_HAL_COMMON_INTERRUPTS_SAVE_MINIMUM_CONTEXT ie. save a | |
| 284 minimal register set rather than a maximal. Create a saved | |
| 285 register set whose address is passed to the interrupt handler | |
| 286 (though it may well be mostly empty). | |
| 287 [also tidied up trailing tabs and long blank lines] | |
| 288 | |
| 289 * src/vec_xvsr.S (hal_default_exception_vsr): Preserve all | |
| 290 registers over interrupts; create a saved register set whose | |
| 291 address is passed to the exception handler so it can modify it. | |
| 292 [also tidied up trailing tabs and long blank lines] | |
| 293 | |
| 294 * src/context.S( hal_setjmp, hal_longjmp): | |
| 295 (hal_thread_switch_context,hal_thread_load_context): Implement | |
| 296 CYGDBG_HAL_COMMON_CONTEXT_SAVE_MINIMUM support; ie. save a minimal | |
| 297 register set rather than a maximal. | |
| 298 | |
| 299 * include/hal_arch.h (HAL_FLUSH_REGISTERS_TO_STACK): New macro | |
| 300 useful for writing exception handlers, for example. | |
| 301 (HAL_GET_GDB_REGISTERS, HAL_SET_GDB_REGISTERS): provided initial | |
| 302 dummy versions of these. | |
| 303 | |
| 304 * tests/sparc_ex.c: A test of the exception mechanism: | |
| 305 deliberately cause unaligned accesses of various sizes and check | |
| 306 we get exactly one trap and correctly step over it by modifying | |
| 307 the saved register set whose address the exception ISR is handed. | |
| 308 This test should be regarded as documentation for the exception | |
| 309 handler API/ABI. | |
| 310 | |
| 311 * tests/PKGconf.mak (TESTS): Build sparc_ex. | |
| 312 | |
| 313 1999-02-01 Hugo Tyson <hmt@cygnus.co.uk> | |
| 314 | |
| 315 * src/vectors.S (genuine_reset): | |
| 316 Include <cyg/hal/halboot.si> to set up RAM controllers, cache and | |
| 317 the like - this is rather platform dependant. | |
| 318 | |
| 319 1999-02-01 Hugo Tyson <hmt@cygnus.co.uk> | |
| 320 | |
| 321 * include/hal_intr.h (HAL_TRANSLATE_VECTOR): | |
| 322 Include a definition of this important macro. Not. | |
| 323 It's used when interrupt chaining is selected, BUT we the HAL make | |
| 324 no distinction, it's up to the kernel to look after things. So | |
| 325 this macro is an assignment. | |
| 326 | |
| 327 1999-01-28 Hugo Tyson <hmt@cygnus.co.uk> | |
| 328 | |
| 329 * include/hal_intr.h: | |
| 330 Get interrupt controller definitions from a board-specific file: | |
| 331 #include <cyg/hal/hal_xpic.h> | |
| 332 Get clock control definitions from a board-specific file: | |
| 333 #include <cyg/hal/hal_clock.h> | |
| 334 Delete the original definitions from this central file. | |
| 335 | |
| 336 * include/hal_cache.h: REMOVED to the simulator tree in the first | |
| 337 instance; more generally to various board-specific files. | |
| 338 | |
| 339 | |
| 340 1999-01-26 Hugo Tyson <hmt@cygnus.co.uk> | |
| 341 | |
| 342 * include/hal_cache.h (HAL_DCACHE_ENABLE) (and its friends): | |
| 343 Define initial empty placeholder versions of these so that cache | |
| 344 tests build. Well it's a start. | |
| 345 | |
| 346 1999-01-21 Jonathan Larmour <jlarmour@cygnus.co.uk> | |
| 347 | |
| 348 * src/hal_boot.c (cyg_hal_invoke_constructors): | |
| 349 Add code to deal with CYGSEM_HAL_STOP_CONSTRUCTORS_ON_FLAG | |
| 350 Tidy up and update description header | |
| 351 Shorten needlessly long lines | |
| 352 Remove all traces of non-CYG_KERNEL_USE_INIT_PRIORITY code | |
| 353 | |
| 354 1999-01-21 Hugo Tyson <hmt@cygnus.co.uk> | |
| 355 | |
| 356 * src/vectors.S (genuine_reset): Set the RAM size to 4MB in the | |
| 357 erc32 simulator. | |
| 358 | |
| 359 * include/hal_io.h: New file; macros to access IO space, really a | |
| 360 place holder at present. | |
| 361 | |
| 362 * include/hal_intr.h: Do not define WATCHDOG() macro. | |
| 363 | |
| 364 * src/hal_intr.c (hal_default_isr): Do not mess with the watchdog | |
| 365 in the the default interrupt handler. | |
| 366 | |
| 367 * src/hal_boot.c (hal_zero_bss): use an efficient 8-byte zero. | |
| 368 (hal_copy_data): use an efficient 8-byte copy. | |
| 369 (cyg_hal_start): turn off the ERC32 watchdog timer before we start | |
| 370 anything very time consuming. | |
| 371 | |
| 372 NB: the 8-byte copy changes require alignment in the linker | |
| 373 script: see the Changelog in hal/sparclite/sim/... for details. | |
| 374 | |
| 375 1999-01-19 Hugo Tyson <hmt@cygnus.co.uk> | |
| 376 | |
| 377 Modified files in arch: | |
| 378 * include/hal_arch.h | |
| 379 * include/hal_intr.h | |
| 380 * src/PKGconf.mak | |
| 381 * src/context.S | |
| 382 * src/hal_intr.c | |
| 383 * src/vec_ivsr.S | |
| 384 * src/vec_xvsr.S | |
| 385 * src/vectors.S | |
| 386 | |
| 387 New files in arch: | |
| 388 * src/hal_boot.c (bss initialization, constructors &c) | |
| 389 * include/hal_cache.h (empty but required header) | |
| 390 | |
| 391 Modified files in sim: | |
| 392 * include/hal_diag.h | |
| 393 | |
| 394 Lots more progress; essentially, everything now works. | |
| 395 | |
| 396 Simulator's treatment of stdio is poor though, it doesn't work | |
| 397 under pkgtest. This has been slowing me up. | |
| 398 | |
| 399 1999-01-14 Hugo Tyson <hmt@cygnus.co.uk> | |
| 400 | |
| 401 * ecc/ecc/hal/sparclite/...: | |
| 402 | |
| 403 Initial checkin of SPARClite HAL; it doesn't work yet, except the | |
| 404 most trivial of tests - timer and interrupt manipulation stuff is | |
| 405 entirely absent. | |
| 406 | |
| 407 Here's some documentation of the initial checkin tree; this will | |
| 408 change, as platform dependent stuff moves correctly into the sim | |
| 409 or <board> directories. | |
| 410 | |
| 411 arch/current/include: | |
| 412 basetype.h nothing much here. | |
| 413 vectors.h definitions of symbols shared between the | |
| 414 various vector code assembler files, can be | |
| 415 included in both C and assembler. | |
| 416 hal_intr.h various macros for clock and interrupt | |
| 417 control plus the eCos vector numbering scheme; we | |
| 418 map thus: | |
| 419 what SPARC trap type eCos vector numbers | |
| 420 ----------------------------------------------- | |
| 421 (not used) --- 0 | |
| 422 interrupts (17-31) 1-15 | |
| 423 user traps (128-255) 16 | |
| 424 exceptions (1-9,36==4) 17-25 | |
| 425 others (10,?) 26 | |
| 426 ----------------------------------------------- | |
| 427 Trap types 5 and 6 which would be eCos vectors 21 | |
| 428 and 22 are handled in the trampoline code in | |
| 429 file src/vectors.S | |
| 430 hal_arch.h definition of a thread context, plus | |
| 431 macros for thread switching and the like. Some | |
| 432 ancillary bit twiddling macros and cache barriers | |
| 433 too. | |
| 434 | |
| 435 arch/current/include/pkgconf: | |
| 436 hal_sparclite.h pkgconf config/header file. | |
| 437 | |
| 438 arch/current/src: | |
| 439 vectors.S boot-up stuff, trampoline code that hangs | |
| 440 on the interrupt/trap vector, handlers for window | |
| 441 under/overflow. Saves some state in registers | |
| 442 then jumps to the interrupt or exception VSR | |
| 443 respectively. | |
| 444 Also instantiates the VSR table. | |
| 445 vec_ivsr.S the default interrupt VSR; it establishes | |
| 446 a C calling environment (ie. anticipates a window | |
| 447 overflow) then locks the scheduler, calls the ISR | |
| 448 with appropriate arguments, and then interrupt_end() | |
| 449 before undoing that environment, anticipating | |
| 450 window underflow and returning to the interrupted | |
| 451 instruction. | |
| 452 vec_xvsr.S the default exception AKA trap VSR; does | |
| 453 the same as the interrupt one but without the | |
| 454 scheduler lock or interrupt end. | |
| 455 hal_intr.c instantiates the ISR table used by the | |
| 456 code above. | |
| 457 context.S context switch code, saves and loads up a | |
| 458 whole register state, coroutine or longjump-like. | |
| 459 icontext.c initializes a context to "jump" to in | |
| 460 context.S at the birth of a new thread. | |
| 461 | |
| 462 sim/current/include: | |
| 463 hal_diag.h macros to deal with debugging output via a | |
| 464 fake serial device of some kind; initially empty, | |
| 465 but definitely board/sim dependent. | |
| 466 | |
| 467 sim/current/include/pkgconf: | |
| 468 hal_sparclite_sim.h pkgconf config/header file. | |
| 469 | |
| 470 sim/current/src: | |
| 471 sim.ld linker script bits for "sim" target. | |
| 472 | |
| 473 //=========================================================================== | |
| 474 //####COPYRIGHTBEGIN#### | |
| 475 // | |
| 476 // ------------------------------------------- | |
| 477 // The contents of this file are subject to the Cygnus eCos Public License | |
| 478 // Version 1.0 (the "License"); you may not use this file except in | |
| 479 // compliance with the License. You may obtain a copy of the License at | |
| 480 // http://sourceware.cygnus.com/ecos | |
| 481 // | |
| 482 // Software distributed under the License is distributed on an "AS IS" | |
| 483 // basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See the | |
| 484 // License for the specific language governing rights and limitations under | |
| 485 // the License. | |
| 486 // | |
| 487 // The Original Code is eCos - Embedded Cygnus Operating System, released | |
| 488 // September 30, 1998. | |
| 489 // | |
| 490 // The Initial Developer of the Original Code is Cygnus. Portions created | |
| 491 // by Cygnus are Copyright (C) 1998,1999 Cygnus Solutions. All Rights Reserved. | |
| 492 // ------------------------------------------- | |
| 493 // | |
| 494 //####COPYRIGHTEND#### | |
| 495 //=========================================================================== | |
| 496 | |
| 497 //EOF ChangeLog |