Make the J-Link driver handle everything needed for FPGA programming,
this includes arbitrary long scans and STABLECLOCKS command.
Also, bump to the latest upstream libjaylink to properly support this.
This code is heavily inspired by Andreas Fritiofson's ftdi.c.
Change-Id: Ic5fd87aa88b58ff1138dc2e0a197bb52321b1541
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/2946
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
BeagleBone debian 7 builds produce:
jlink.c: In function 'jlink_speed':
jlink.c:218:11: error: declaration of 'div' shadows a global declaration [-Werror=shadow]
jlink.c: In function 'check_trace_freq':
jlink.c:1065:54: error: declaration of 'div' shadows a global declaration [-Werror=shadow]
jlink.c: In function 'config_trace':
jlink.c:1101:11: error: declaration of 'div' shadows a global declaration [-Werror=shadow]
Fix this by changing the local variable to 'divider'.
Change-Id: I96a0cc0f7d4d4af5a56aa1e918e5416d3c61cbfe
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Reviewed-on: http://openocd.zylin.com/3185
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
BeagleBone debian 7 builds produce:
adi_v5_jtag.c: In function 'jtag_ap_q_bankselect':
adi_v5_jtag.c:336:11: error: declaration of 'select' shadows a global declaration [-Werror=shadow]
Fix this by changing the local variable to 'sel'.
Change-Id: I8e29662ac12bc77d38d5064046d59b7364853cd9
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Reviewed-on: http://openocd.zylin.com/3184
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
The existing arm_adi_v5.c code decodes CoreSight peripherals based
on the part number field. However, these are specific to a
particular manufacturer (often ARM). The same part number from
two different manufacturers (distinct designer ids) should not
decode as the same CoreSight peripheral.
The Analog Devices ADSP-SC58x and ADSP-BF70x have peripherals that
overlap with existing OpenOCD decoding. The part number is the
same as existing OpenOCD decoding, but have a different JEP106 code.
Most, if not all, of the existing part number entries in
arm_adi_v5.c are probably specific to ARM. Change all entries
suspected to be designed by ARM to match only ARM's designer ID.
However, to preserve legacy behavior, existing non-ARM entries are
encoded with a wildcard so that they will behave in the same way as
the existing legacy code. It is desirable, however, to start
encoding the data with designer codes to avoid such ambiguity.
Revising the code to check both the part number and designer id
seemed to a warrant a const array lookup table instead of a
multi-tiered switch statement.
Also try to sync part identification IDs with relevant ARM docs.
Change-Id: Iac1374e4cfc6f04cebb479c0e3fa9bde527cc4a3
Signed-off-by: Peter Lawrence <majbthrd@gmail.com>
[andreas.fritiofson@gmail.com: change JEP106 to designer ID, cleanup]
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3128
Tested-by: jenkins
The HLA target shares an examine handler with cortex_m but since it
lacks direct access to DAP, some operations need to be omitted.
Change-Id: Ifdd9d3da4a3a3c2e1c9721284b21d041b3ccaa7a
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/3183
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
work around issues with software breakpoints when the text segment
is mapped read-only by the OS. Set DACR to "all-manager" to bypass
TLB permission checks on memory access.
Change-Id: I79fd9b32b04a4d538d489896470ee30b26b72b30
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3107
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Change tests when reading from 'value' in sysfs from =='0' to !='1'.
This guards against broken sysfs GPIO implementations that return
non-zero for high rather than just '1' while still being clean and
correct code. Note that sysfs will never output a leading zero even
in a very broken implementation as that is covered in gpiolib.c, not
the offending driver.
Tested against broken Freescale kernel 3.14.38 on i.MX6SL.
Change-Id: Id05567bb8504b1babef33d6ee5172bceefeca8b8
Signed-off-by: Matthew Campbell <mcampbell@izotope.com>
Reviewed-on: http://openocd.zylin.com/3121
Tested-by: jenkins
Reviewed-by: Antonio Borneo <borneo.antonio@gmail.com>
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
On JTAG, all reads are pipelined. If you read a register, the result is not
delivered inside the request that issued the read, it is delivered in the
following request. The current code therefore issues a scan of the RDBUFF
register after each read. This adds a superfluous transaction after each
read.
This patch follows a strategy similar to what SWD already implements.
It also leverages that all JTAG reads are pipelined, i.e. the result
will be clocked out in the next JTAG data phase, no matter if it's
READ or WRITE. Therefore it's never necessary to explicitly read RDBUFF
other than for the very last READ before a dap_run().
Change-Id: Ie40b1fef3203f0cdcb503f40dcbd2a68b0f9776c
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3167
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Debug initialization blindly selects AP#0 as default, which is the AHB-AP
in many cases. This sets the default for target_read/write functions.
However, AHB-AP is the wrong choice, because it bypasses caches on read
and write and also makes some peripherals inaccessible (e.g. l2 outer
caches). This patch explicitely selects the APB-AP (debug_ap) as the
default.
Change-Id: I13f9e0750186d35dcfc135c8d67d437c5884d9c4
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3113
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Remove entirely the concept of a "selected" AP that has to be maintained
between calls. All the information the DAP ops need are now provided to
each call through the AP/DAP pointer.
Consolidate the cache of the SELECT fields into one single field caching
the entire register.
Change-Id: I2e1c93ac5ee8ac38a7d680ca2c660c30093a6b87
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3165
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Move the mandatory dap_ap_select() call into the dap_queue_ap_read/write
wrapper.
This avoids the need for dap_ap_select() and the notion of a "current" AP
within target code.
Change-Id: I5cde8f3eef2c662f7458be6f3b3dd44ea693bd74
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3164
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Make dap apsel without arguments show current state instead of changing
to AP 0.
Change-Id: I75ea10e3e1b8a067f2dc417ec6691dc7ceec1af6
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3163
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
All AP operations should select the AP to use before calling it so
there's no point in restoring the previous value afterwards.
The explicit call to dap_ap_select() before all AP operations should be
moved into dap_queue_ap_read/write() which then would have to take the
AP as an argument instead of the DAP.
Change-Id: Icacb0c76ef2a5ac36b4d2f26b52ec01a8850286e
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3156
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
It's currently set during target creation but the AP that will be used
for the target is not even known.
Change-Id: I4502e7eb1fa8d90f746445b8cf8a4c21cb7d519e
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3155
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The AP for which the TAR/CSW is printed may not be the one that caused
the failure. Remove the flawed output entirely. The correct info is
printed in mem_ap_read/write anyway.
Change-Id: I97580a0662dcf02e80646e45445cdbfc251122d8
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3154
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
All mem_ap_* functions now make sure the SELECT register is updated with
the AP number that it's operating on. This shouldn't have to be handled
explicitly.
Change-Id: Ib193d8930fabb6a25715064355f98258c9580b5d
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3153
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
This function does two separate things, powering up the DP and setting
up a MEM-AP. But the DP needs to be powered before even searching for a
MEM-AP to initialize and targets may have multiple MEM-APs that need
initializing.
Split the function into dap_dp_init() and mem_ap_init() and change all
call sites to use the appropriate one.
Change-Id: I92f55e09754a93f3f01dd8e5aa1ffdf60c856126
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3151
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reduce use of magic numbers and add AXI type MEM-AP detection. Don't try
to call dap_rom_display on a non-existent AP.
AP identification is unique per designer, so make sure the JEDEC code
matches ARM when interpreting the AP type.
Change-Id: I8e86b7de61811382afe99bf15094ab71b43f5fdf
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3150
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Change the debug_ap and memory_ap fields of the cortex_a target and
the debug_ap field of the cortex_m target to be pointers to the
struct adiv5_ap instead of AP numbers in some known DAP.
This reduces the dependency on the DAP struct in the targets and
enables MEM-AP accesses to take the relevant AP as parameter.
Change-Id: I39d7b134d78000564b7eec5bff464adf0ef89147
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3147
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
This required fixing the AP ID parsing in dap_find_ap() to
match IHI0031C. The AXI type was added too.
Change-Id: I44577a7848df37586e650dce0fb57ac26f5f858c
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3146
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
This will make it possible to reference directly the AP used for debug
in the target instance and remove the DAP reference. This will in turn
enable getting rid of the need to select an "active" AP in the DAP (using
dap apsel).
Change-Id: I265846a427c714204f4fd3df3cdb75843686c2d0
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3144
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Not clear what it was supposed to be used for. It probably shouldn't.
Change-Id: Ife1d833e59ba80f93876447d752a0ca7e7b57b0f
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3143
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
The Cortex-A and Cortex-M keeps an arm_jtag struct around just to be
able to pass a pointer to it to one common JTAG function which anyway
only uses the TAP field.
Refactor the function to take a TAP directly, remove the legacy struct
from cortex instances and store the TAP pointer only in the DAP.
Cortex-M makes a call to arm_jtag_setup_connection() with the struct
but the function does nothing useful for a Cortex-M target so remove
the call.
Change-Id: I3b33709ef55372ef14522ed4337e9f2e817ae3ab
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3142
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Making the SWD driver aware of the DAP that controls it is a layering
violation.
The only usage for the DAP pointer is to store the number of idle cycles
the AP may need to avoid WAITs. Replace the DAP pointer with a cycle
count hint instead to avoid future misuse.
Change-Id: I3e64e11a43ba2396bd646a4cf8f9bc331805d802
Signed-off-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Reviewed-on: http://openocd.zylin.com/3141
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Separate out the values from adiv5_dap that are associated with a specific AP into a new struct, so we can properly support multiple APs. Remove the DAP struct from the armv7* structs, because we can have multiple CPUs per DAP, and we shouldn't have multiple DAP structs. Tidy up a few places where ap_current is used incorrectly.
Change-Id: I0c6ef4b49cc86b140366347aaf9b76c07cbab0a8
Signed-off-by: Patrick Stewart <patstew@gmail.com>
Reviewed-on: http://openocd.zylin.com/2984
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Prepare to support multiple cortex-m cores on one DAP. Uses mem_ap_sel_*
functions and removes mem_ap_* functions. Adds a new debug_ap
parameter to the cortex_m (currently set to zero as in existing code).
Change-Id: I6926029d1e7bf44a42d453d1aff349bda824ba72
Signed-off-by: Patrick Stewart <patstew@gmail.com>
Reviewed-on: http://openocd.zylin.com/2983
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
ERROR_WAIT is better than ERROR_FAIL in timeout condition.
Change-Id: Iefe837f276a9091ce6c18db5947212c449f49d89
Signed-off-by: Alamy Liu <alamy.liu@gmail.com>
Reviewed-on: http://openocd.zylin.com/2934
Tested-by: jenkins
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
This is a TODO in the src/target/arm_adi_v5.h for MEM-AP registers.
Some new registers are introduced in ADIv5.2 specification.
MEM_AP_REG_MGT (0x20) // Memory Barrier Transfer register
MEM_AP_REG_TAR64 (0x08) // Bits[63:32] of Transfer Address
MEM_AP_REG_BASE64 (0xF0) // Bits[63:32] of Debug Base Address
Refer to
7.5 MEM-AP register summary in
IHI0031C: ARM Debug Interface Architecture Specification ADIv5.0 to ADIv5.2
Change-Id: I3bc4296a04c35f5c64f851e5865d3099922613fa
Signed-off-by: Alamy Liu <alamy.liu@gmail.com>
Reviewed-on: http://openocd.zylin.com/2904
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Problem
No matter what target->coreid is, it always shows
Detected core 0 dbgbase: ...
In dap_lookup_cs_component(), it decreases the core index value to zero
in order to find the desired core.
The reference to coreidx is necessary considering "a device which has nested
ROM tables, with each core described in its own table." (by Paul Fertser).
Change-Id: I9b56d45d6edf6639e748a625ab27787f8e5a5776
Signed-off-by: Alamy Liu <alamy.liu@gmail.com>
Reviewed-on: http://openocd.zylin.com/2902
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
They are only used to initialize the flash bank sectors and never modified.
Explicitly specify the array length while at it.
Cleanup before adding XMC4800 support.
Change-Id: I2985b9a9946b67798dbfd47d8b219d93a7ffc3da
Signed-off-by: Andreas Färber <afaerber@suse.de>
Reviewed-on: http://openocd.zylin.com/3131
Tested-by: jenkins
Reviewed-by: Jeff Ciesielski <jeffciesielski@gmail.com>
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
cortex_a_read_apb_ab_memory_fast() uses the wrong order of ITR and DSCR
writes when setting up the transfer. ARM DDI0406C says in C8.2 regarding
"Fast mode" operation to first switch to fast mode and then latch the
instruction in ITR. Current implementation first wrote ITR, causing
the instruction to be executed immediately, then switched to fast mode
without an instruction latched. Repeated reading of DTRTX didn't
execute LDC and thus replicated its current content into the whole buffer.
This patch uses the following, revised algorithm:
1) switch to non-blocking mode and issue the LDC for the first word
2) if more than one word is to be read:
- switch to fast mode
- latch the LDC instruction into ITR (it is _not_ executed)
- issue (count-1) reads of DTRTX register, each read returns the current
content of DTRTX and re-issues the latched instruction
-> now the second-to-last word is in the buffer and the LDC for the last
word has been issued.
3) wait for the last instruction to complete
4) switch back to non-blocking mode
5) Read DTRTX for the last (or: only) word and put it into the buffer
Change-Id: I44f5c585962ffa5af257c3d5a2a802c122b6b1e4
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3122
Tested-by: jenkins
Reviewed-by: Christopher Head <chead@zaber.com>
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
When accessing memory through the ARM core, privilege levels and mmu
access permissions observed. Thus it depends on the current mode of the
ARM core whether an access is possible or not. the ARM in USR mode can
not access memory mapped to a higher privilege level. This means, if the
ARM core is halted while executing at PL0, the debugger would be
prevented from setting a breakpoint at an address with a higher privilege
level, e.g. in the OS kernel. This is not desirable.
cortex_a_check_address() tried to work around this by predicting if an
access would fail and switched the ARM core to SVC mode. However, the
prediction was based on hardcoded address ranges and only worked for
Linux and a 3G/1G user/kernel space split.
This patch changes the policy to always switch to SVC mode for memory
accesses. It introduces two functions cortex_a_prep_memaccess() and
cortex_a_post_memaccess() which bracket memory reads and writes. These
function encapsulate all actions necessary for preparation and cleanup.
Change-Id: I4ccdb5fd17eadeb2b66ae28caaf0ccd2d014eaa9
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3119
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Tested-by: jenkins
when disabling the mmu to access physical addresses, normally the d-cache
must be disabled as well. Disabling the d-cache also requires a full
clean&invalidate. However, since all memory writes are treated as write-
through no-allocate and memory reads do not allocate cache lines,
effectively the d-cache state does not change at all. We can therefore
save the the d-cache disabling and flushing.
This patch also simplifies the function a bit.
Change-Id: Ia17c56a28f432156429cd4596107e3652b788e63
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3114
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
for minimal impact on the hardware state, force all memory accesses to
bypass the caches and tlbs. This may actually be the default, but ARM
recommends in DDI0406C to set proper default values on debug init.
Change-Id: If5ac097b6ee725c047b1e86c2f90eabe16b98c7b
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3079
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Tested-by: jenkins
Other cache functions use an updated pattern for the address range loop.
Bring dcache clean and flush functions in line.
Change-Id: Iccb4a05c49054471033a3403363110cb08245d5b
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3035
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Call armv7a_l1_d_cache_flush_virt() before writing the breakpoint,
to make sure the d-cache is clean and invalid at the breakpoint
location down to PoC.
Call armv7a_l1_d_cache_inval_virt() after writing the breakpoint
again, so that d-cache will pick up the modified code.
Call armv7a_l1_i_cache_inval_virt() after writing the breakpoint
to memory to make the change visible to the CPU.
Change-Id: I24fc27058d99cb00d7f6002ccb623cab66b0d234
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3033
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Tested-by: jenkins
D-Cache invalidate is a dangerous operation. It will only work correctly
if full cache lines are invalidated. When partial cache lines are
invalidated, i.e. the target address range does not start and end
at a cache line boundary, cpu data writes outside of the target range
will be dropped. This patch adds special treatment for partial cache
lines by doing a clean & invalidate on the partial lines before
invalidating the rest of the range.
Change-Id: I64099ddb058638e990a7eb0ee911b9cc8f6f8901
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3034
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
The following changes are implemented:
- Clean&Invalidate the VA range to PoC *before* the write takes place
- Remove SMP handling since DCCIMVA instruction already maintains SMP
coherence.
- Remove separate Invalidate step
Change-Id: I19fd3cc226d8ecf2937276fc63258b6a26e369a7
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3027
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Tested-by: jenkins
This patch adds a function for cleaning & invalidating a virtual
address range from the architecture caches down to the point of
coherence.
Change-Id: I4061ab023a3797fabc967f3a34498034841d52c6
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3026
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Tested-by: jenkins
There's only one function left that handles cache info display,
no need any more for a function pointer and runtime initialization.
Change-Id: I90b09577f81607917b11f0ab5600a0e2dce223e2
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3025
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Tested-by: jenkins
ARMv7 architecture allows up to 7 cache levels that are architecturally
visible, as opposed to "system caches", which are outside of the domain
defined by ARMv7 and require separate management. This patch enables
detection and identification of caches at all levels. It also implements
a new "flush-all" function that cleans & invalidates all cache levels to
the "Point of Coherence".
Change-Id: Ib77115d6044d39845907941c6f031e208f6e0aa5
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3024
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Tested-by: jenkins
This patch is on the path to unified handlers for both inner and
outer caches. It removes the special overrides installed when
an outer cache is configured.
Change-Id: I747f2762c6c8c76c700341cbf6cf500ff2a51476
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3022
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
The outer cache is not necessarily at L2 in a system. Rename functions
to make that clear.
Change-Id: Ia636a4844f50634f2bdf5cdce285febc1a47c11f
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3020
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
This patch introduces, new command set and handlers for l1 and l2x caches.
Patch set 10 folded the following changes into this one:
Ib1a2a1fc1b929dc49532ac13a78e8eb796ab4415
If8d87a03281d0f4ad402909998e7834eb4837e79
I0749f129fa74e04f4e9c20d143a744f09ef750d8
Change-Id: I849f4d1f20610087885eeddefa81d976f77cf199
Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/2800
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
This was needed for ahb access
Change-Id: I638f45a276a593c08140b5d9d7480617aa85f096
Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Reviewed-on: http://openocd.zylin.com/2796
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Tested-by: jenkins
Problem
dap->ap_current is register value, not field value.
it restores invalid ap when it calls dap_ap_select(dap, ap_old) later.
* assume the current ap is 1, dap->ap_current value would be (1 << 24).
ap_old = dap->ap_current; <-- ap_old = 1<<24 = 0x1000000.
...
dap_ap_select(dap, ap_old); <-- select 0x1000000, not 1.
* All AP registers accessing fail afterwards.
One of the reproducible case(s): CORE residents in AP >= 1
dap_lookup_cs_component() being used to find PE(*).
In most cases, PE would be found in AP==0, hence the problem is hidden.
When AP number is 1, dap->ap_current would have the value of 1<<24.
Anyone get the AP value with dap->ap_current and resotre it later would
select the wrong AP and all accessing later would fail.
The ARM Versatile and/or FPGA would have better chance to provide this
kind of environment that PE residents in AP>=1. As they have an 'umbrella'
system at AP0, and main system at AP>=1.
* PE: Processing Element. AKA Core. See ARM Glossary at
http://infocenter.arm.com/help/topic/com.arm.doc.aeg0014g/ABCDEFGH.html
Fix
Use dap_ap_get_select() to get ap value.
a. Retrieve current ap value by calling dap_ap_get_select();
src/flash/nor/kinetis.c
src/target/arm_adi_v5.c
b. The code is correct (dap->ap_current >> 24), but it's better to use
dap_ap_get_select() so everything could be synchronized.
src/flash/nor/sim3x.c
Change-Id: I97b5a13a3fc5506cf287e299c6c35699374de74f
Signed-off-by: Alamy Liu <alamy.liu@gmail.com>
Reviewed-on: http://openocd.zylin.com/2935
Reviewed-by: Andreas Färber <afaerber@suse.de>
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
If supported, the maximum transport speed is now retrieved from the
device.
Change-Id: I614f405ec91cf199c851781785fd26cbd10c37a6
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/2955
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
This patch uses libjaylink which is a library to access J-Link
devices. As other tools which are not in the scope of OpenOCD also
need to access J-Link devices a library is used. A firmware upgrade
tool and an advanced configuration tool for J-Link devices are under
development.
Further versions of libjaylink will support additional features
OpenOCD could benefit from. This includes TCP/IP as additional
possibility to connect to J-Link devices as well as power tracing and
device internal communication. The latter is used to access
peripherals on some development boards (e.g EFM32 STK and DVK).
Integration of libjaylink is realized with a git submodule like
jimtcl. As libjaylink depends on libusb-1.0 only, no additional
dependency is introduced for OpenOCD.
All low-level JTAG and SWD implementations of the current driver are
left untouched and therefore no incompabilities are to be expected.
Improvements of this patch:
* Support for more USB Product IDs, including those with the new
scheme (0x10xx). The corresponding udev rules are also updated.
* Device selection with serial number and USB address.
* Adaptive clocking is now correctly implemented and only usable for
devices with the corresponding capability.
* The target power supply can now be switched without the need for
changing configuration and power cycling the device.
* Device configuration is more restrictive and only allowed if the
required capabilities are available.
* Device configuration now shows the changes between the current
configuration of the device and the values that will be applied.
* Device configuration is verified after it is written to the device
exactly as the vendor software does.
* Connection registration is now handled properly and checks if the
maximum number of connections on a device is reached. This is also
necessary for devices which are attached via USB to OpenOCD as
some device models also support connections on TCP/IP.
* Serial Wire Output (SWO) can now be captured. This feature is not
documented by SEGGER however it is completely supported by
libjaylink.
This patch and libjaylink were tested on Ubuntu 14.04 (i386),
Debian 7 (amd64), FreeBSD 10.0 (amd64) and Windows XP SP3 (32-bit)
with the following device and target configurations:
* JTAG: J-Link v8.0, v9.0 and v9.3 with AT91SAM7S256
* SWD: SiLabs EFM32 STK 3700 (EFM32GG990F1024)
* SWD: J-Link v8.0, v9.0 and v9.3 with EFM32GG990F1024
* SWD: XMC 2Go (XMC1100)
* SWD: XMC1100 Boot Kit (XMC1100)
* SWD: IAR Systems / Olimex Eval Board (LPC1343F)
* SWD: Nordic Semiconductor nRF51 Dongle (nRF51422)
* SWD: SiLabs EZR32 WSTK 6220A (EZR32WG330FG60G)
Except for Windows XP all builds are tested with Clang in addition to
GCC. This patch and libjaylink are not tested on OSX yet.
Change-Id: I8476c57d37c6091c4b892b183da682c548ca1786
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/2598
Tested-by: jenkins
Reviewed-by: Andreas Färber <afaerber@suse.de>
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
http://support.code-red-tech.com/CodeRedWiki/DebugAccessChip
> Note that once you have recovered debug access to your MCU, then in most cases you should then modify your Debug Configuration to turn vector catch off again. If this is not done, then this can cause problems in some circumstances with some versions of the Code Red IDE. For example with NXP LPC13xx parts, connecting more than once to the MCU with vector catch enabled can lead to the part ID being incorrectly read - which can again cause debug connections to fail
This patch adds an alternative part ID for LPC1343. With this patch "program" command works fine for flashing.
Change-Id: I8632e898a4c33102455925d25715b4f4edfa1d97
Signed-off-by: Jakub Kubiak <jakub@kubiak.es>
Reviewed-on: http://openocd.zylin.com/2782
Tested-by: jenkins
Reviewed-by: Andreas Färber <afaerber@suse.de>
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
This adds docs, example config, flash driver.
Driver is only supports K1921VK01T model for now.
Change-Id: I135259bb055dd2df1a17de99f066e2b24eae1b0f
Signed-off-by: Bogdan Kolbov <kolbov@niiet.ru>
Reviewed-on: http://openocd.zylin.com/3011
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
The probe and info methods had duplicate sections decoding family names
to generate a human friendly part name. Extract this to a common
helper.
Change-Id: I4c6309d83c601e154b7c14ad9c15c53854ee1e98
Signed-off-by: Karl Palsson <karlp@tweak.net.au>
Reviewed-on: http://openocd.zylin.com/2932
Tested-by: jenkins
Reviewed-by: Andreas Färber <afaerber@suse.de>
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Basic basic support to get running, magic numbers taken from revision
0.90 of the reference manual.
Signed-off-by: Karl Palsson <karlp@tweak.net.au>
Change-Id: Iff6ab94d30698f056ef09f7a856b7285fed8f441
Reviewed-on: http://openocd.zylin.com/2931
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
By saving a pointer to the tail of the list,
we don't need to traverse the entire command
queue before we're able to append an item to
it.
With this patch, I see a 10% improvement when
using the embedded XDS100v2 on AM437x IDK board
to load a 4MiB binary (linux zImage) to DDR
with load_image.
IOW, we went from ~305kB/sec to ~336kb/sec.
Change-Id: Idb55d49f0d0106043374ab520b2f3b6b32f2c50f
Signed-off-by: Felipe Balbi <balbi@ti.com>
Reviewed-on: http://openocd.zylin.com/2709
Tested-by: jenkins
Reviewed-by: Stian Skjelstad <stian@nixia.no>
Reviewed-by: Daniele Emancipato <daniele12457@hotmail.com>
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
The svf_progress_enabled variable is global, hence its lifetime is not
limited and it retains the value from the previous run. Fix this by
explicit assignment.
Change-Id: Id6f4fa88f39521606342a37f6876a0948ac5406e
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/3111
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
This makes SVF error output match actual line numbers in the file
processed.
Change-Id: I1fa4b9d0891e4358b7beada516945d5331ebe182
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/2945
Tested-by: jenkins
Kinetis driver checks MDM STAT register to detect secured state of MCU.
An unsecured clean device typically triggered a huge fat alarm message.
Now when driver detects secured state it tries to halt MCU and then
re-reads status register.
Command "mdm mass_erase" used to deassert reset when finished
and MCU started looping in hard fault - WDOG reset cycle.
Now "reset halt" is issued. Clean flash is not run after mass_erase.
Change-Id: I23f393509fbd8751d44ffc744ff2d67f1074f74e
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/3010
Tested-by: jenkins
Reviewed-by: Thomas Schmid <thomas@rfranging.com>
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
SAMD driver suffered from following problems:
1) Flash was erased as an integral part of flash write.
It was not documented so with usual workflow it resulted
in erasing flash twice (and reducing flash lifespan)
and in almost double flashing time.
2) Sector erase was silently skipped if "is_erased" flag was set.
"is_erased" logic was not reliable, e.g. when a row write
was aborted after successful write of some pages, sector was
still considered as erased. "is_erased" flag could not
cope with flash writes from a user program.
3) Writing of a block with start address unaligned to a flash page
resulted in failed assert and OpenOCD abort.
4) Disabling cache in bit 18 of 16-bit halfword never worked.
MCU implements cache invalidate in hardware so there is no need
to take care about. This bug was reported by Tony DiCola.
New code does not erase flash in write operation.
Instead it traditionally relies on erasing flash beforehand.
"is_erased" logic and cache disabling is completely removed.
It simplifies write procedure a lot and flash write is now faster.
The change partly solves ticket #109 SAMD/SAM4L driver doubles flash erase.
Change-Id: I582b497d01a351575533a1f8c9810a4413be0216
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/3045
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
This adds support for the Atmel SAML21 variant B parts.
There is minimal change between the two variants, but in
variant B the automatic page write which the at91samd flash
driver relies on to be enabled is disabled by default.
With this patch the write row function will now issue a page write
command after each of the four pages in the row if the MANW (manual
write) bit is set. This also fixes flash write for the SAMC20/SAMC21
devices which have the MANW bit set by default as well.
I have also moved the device ID (DID) register bitfield extraction
from the find_part into helper macros. These can be used in the future
if there are more workarounds for specific devices.
Tested (programming) on:
ATSAML21-XPRO
ATSAML21-XPRO-B
SAMC21 Xplained Pro
SAMD21 Xplained Pro
SAMD20 Xplained Pro
Change-Id: I401a8aa1efd64730840c0d62cf49a1e880ea5900
Signed-off-by: Andreas Loehre <alohre@gmail.com>
Reviewed-on: http://openocd.zylin.com/2903
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
cmsis-dap protocol has both DAP_Connect and DAP_Disconnect commands.
Logically if cmsis_dap_swd_switch_seq() calls DAP_Connect in connected
state it should call DAP_Disconnect first.
Doing so surprisingly solves problems on Atmel EDBG with target SAMD/R/L/C.
Without this change SWD communication breaks after reset run/halt
in config "reset_config srst_only" and reconnect trials repeatedly
fail with "SWD ack not OK: 0 JUNK"
Change-Id: Ie951098d5e0c83f388e2df414608aaabec2135c9
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/3098
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
Atmel introduced a "Device Service Unit" (DSU) that holds the CPU
in reset if TCK is low when srst (RESET_N) is deasserted.
Function is similar to SMAP in ATSAM4L, see http://openocd.zylin.com/2604
Atmel's EDBG adapter handles DSU reset correctly without this change.
An ordinary SWD adapter leaves TCK in its default state, low.
So without this change any use of sysresetreq or srst
locks the chip in reset state until power is cycled.
A new function dsu_reset_deassert is called as reset-deassert-post event handler.
It optionally prepares reset vector catch and DSU reset is released then.
Additionally SWD clock comment is fixed in at91samdXX.cfg and clock is
lowered a bit to ensure a margin for RC oscillator frequency deviation.
adapter_nsrst_delay 100 is commented out because is no more necessary after
http://openocd.zylin.com/2601
Change-Id: I42e99b1b245f766616c0a0d939f60612c29bd16c
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/2778
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
According to Infineon, XMC4500 EES AA13 with date codes before GE212 -
as seen on an XMC4500 General App Kit - had a zero SCU_IDCHIP register.
Handle this by extending our checks to not error out on zero SCU_IDCHIP
and by printing a useful info string in that case.
Change-Id: Ic2d641a314627dd5a1ff775a0113999191b95e3d
Signed-off-by: Andreas Färber <afaerber@suse.de>
Reviewed-on: http://openocd.zylin.com/2751
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-by: Jeff Ciesielski <jeffciesielski@gmail.com>
This is a complete flash driver for the Infineon XMC4xxx family of
microcontrollers, based on the TMS570 driver by Andrey Yurovsky.
The driver attempts to discover the particular variant of MCU via a
combination of the SCU register (to determine if this is indeed an
XMC4xxx part) and the FLASH0_ID register (to determine the variant).
If this fails, the driver will not load.
The driver has been added to the README and documentation.
Tests:
* Hardware: XMC4500 (XMC4500_relax), XMC4200 (XMC4200 enterprise)
* SWD + JTAG
* Binary: 144k, 1M
Note:
* Flash protect only partly tested. These parts only allow the flash
protection registers (UCB) to be written 4 times total, and my devkits
have run out of uses (more on the way)
Future Work:
* User 1/2(permalock) locking support via custom command
* In-memory flash loader bootstrap (flashing is rather slow...)
Change-Id: I1d3345d5255d8de8dc4175cf987eb4a037a8cf7f
Signed-off-by: Jeff Ciesielski <jeffciesielski@gmail.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
Reviewed-on: http://openocd.zylin.com/2488
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
This is a driver for the Atmel Cortex-M7 SAMV, SAMS, and SAME.
I started with the at91sam4.c driver and then restructured it
significantly to try to simplify it and limit the functionality
to just a flash driver, as well as to comply with the style guide.
Change-Id: I5340bf61f067265b8ebabd3adad45be45324b707
Signed-off-by: Morgan Quigley <morgan@osrfoundation.org>
Reviewed-on: http://openocd.zylin.com/2952
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Commit 68101e67ac introduced a
regression which resulted for ever-growing registers list (as output
by "reg" command), its contents were doubled every reset (actually,
every examination).
Change-Id: Ie3409c795160a2fc840a5e8a892928df0bcc0c57
Reported-by: Daniele Emancipato <daniele12457@hotmail.com>
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-on: http://openocd.zylin.com/3100
Tested-by: jenkins
Reviewed-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
This patch brings the calculation of the address ranges handled by
ttbr0 and ttbr1 registers in line with ARM DDI 0406C, Table B3-1
Change-Id: Ib807c4b1cb328a6f661e1a0898e744e60d3eccac
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3006
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
If ttbcr is changed after the debugger has examined a target for the
first time, address translations may fail. This problem does not show up
with Linux because it doesn't use ttbr1, but it shows with other OS that
use this feature. If the debugger connects to the target while it's in
u-boot, all address translations will fail after the OS has booted and
the target can not be debugged.
This patch reads the ttbcr in armv7a_mmu_translate_va() and compares it
a cached value. If a difference is detected, armv7a_read_ttbcr() is called
to re-parse the ttb configuration and update the cache.
Change-Id: I1c3adf53ea9d748a0e1e3091d9581e5c43ed64e8
Signed-off-by: Matthias Welwarsky <matthias@welwarsky.de>
Reviewed-on: http://openocd.zylin.com/3005
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
kinetis_write() with byte count not divisible by prog_section_chunk_bytes
computed wrong wc and therefore paded section chunk by some
random data instead of 0xff
Change-Id: Ic7c66d8a3ceacda9e611e98b9fbf943b8001774b
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/2994
Tested-by: jenkins
Reviewed-by: Thomas Schmid <thomas@rfranging.com>
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Kx devices with FTFA flash need the watchdog disabled when programming.
I tried to keep overhead as small as possible and re-use registers that
were already inquired (e.g. sim_sdid).
Change-Id: Ibc29a26ec34102d78a6c3920dd16f63e134a8f6f
Signed-off-by: Thomas Schmid <thomas@rfranging.com>
Reviewed-on: http://openocd.zylin.com/2986
Tested-by: jenkins
Reviewed-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
This makes it easier to relocate the install tree of OpenOCD from where
it was originally built (for example, if put onto a different machine),
without having to change scripts or add something to the command line
every time.
Change-Id: Ia5edf0eba166f7a999f267bd6a92402dab9b399e
Signed-off-by: Jonathan Larmour <jifl@eCosCentric.com>
Reviewed-on: http://openocd.zylin.com/3004
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
The memory leak occurs when opening a file fails. It can be
reproduced by using the "flash verify_bank" command with a filename
that does not exist.
Change-Id: I60b7b545c18793d750ff75d08124fde3f0aa6f64
Signed-off-by: Marc Schink <openocd-dev@marcschink.de>
Reviewed-on: http://openocd.zylin.com/2998
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
If we work on smp system, the output of step command will depend
on Id of default target.
This patch adds additional information to help find what on which
core is happening.
Example of LOG after this patch.
imx6.cpu.1: target state: halted
^^^^^^^^^^
target halted in ARM state due to breakpoint, current mode: Supervisor
cpsr: 0x60000093 pc: 0x80076c0c
MMU: enabled, D-Cache: enabled, I-Cache: enabled
imx6.cpu.0: target state: halted
^^^^^^^^^^
target halted in ARM state due to debug-request, current mode: Supervisor
cpsr: 0x20000193 pc: 0x802ccb6c
MMU: enabled, D-Cache: enabled, I-Cache: enabled
Change-Id: I536a2cce33b5ab10af9de2a43b9960320c17729f
Signed-off-by: Oleksij Rempel <external.Oleksij.Rempel@de.bosch.com>
Reviewed-on: http://openocd.zylin.com/2691
Tested-by: jenkins
Reviewed-by: Spencer Oliver <spen@spen-soft.co.uk>
A segmentation fault in cortex_m_endreset_event() is sometimes raised
with very broken target like Kinetis Kx with erased flash and active WDOG.
Debugging revealed that cortex_m->dwt_num_comp is 4 and
dwt_list is NULL at cortex_m:290
Change-Id: I229c59d6da13d816df513d1dbb19968e4b5951e2
Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
Reviewed-on: http://openocd.zylin.com/2989
Reviewed-by: Thomas Schmid <thomas@rfranging.com>
Tested-by: jenkins
Reviewed-by: Paul Fertser <fercerpav@gmail.com>
Memory for the symbol table was allocated by malloc but not initialized other
than with the symbol name. Therefore `address` and `optional` members were
having arbitrary values leading to every symbol being optional most of the
time which messes up RTOS auto-detection. Memory will now be zero-initialized
as in other RTOS implementations.
Change-Id: I6c6e31ec1ef7e043061adf8c695b2139620e005d
Signed-off-by: Daniel Krebs <github@daniel-krebs.net>
Reviewed-on: http://openocd.zylin.com/3017
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
According to "man 2 poll" the correct header to include is poll.h, not
sys/poll.h. Reported by a build against musl.
Change-Id: I5298b49dc947d1a368e423104c0c0c7b9bdd1a10
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
Reviewed-on: http://openocd.zylin.com/2947
Tested-by: jenkins
In the case that the STKALIGN bit is set on Cortex M processors, on
entry to an exception - the processor can store an additional 4 bytes
of padding before regular stacking to achieve 8-byte alignment on
exception entry. In the case that this padding is present, the
processor will set bit (1 << 9) in the stacked xPSR register. Use the
new calculate_process_stack callback to take into account the xPSR
register and use it on the standard Cortex_M3 stacking.
Note: Change #2301 had some misinformation regarding the padding. On
Cortex-M the padding is stored BEFORE stacking so xPSR is always
available at a fixed offset.
Tested on a Cortex-M0+ (Atmel SAMR21) board which has STKALIGN fixed
to a '1' such that this alignment always occurs on non-aligned stacks.
Behavior of xPSR verified via the (bad-sorry) assembly program below by
setting a breakpoint on the SVC_Handler symbol. The first time
SVC_Handler is triggered the stack was 0x20000ff8, the second time
SVC_Handler is triggered the stack was 0x20000ffc. Note that in both
cases the interrupt handler gets 0x20000fd8 for a stack pointer.
GDB exerpt:
Breakpoint 1, 0x000040b6 in Reset_Handler ()
(gdb) hbreak SVC_Handler
Hardware assisted breakpoint 2 at 0x40f8
(gdb) cont
Continuing.
Breakpoint 2, 0x000040f8 in SVC_Handler ()
(gdb) print $msp
$3 = (void *) 0x20000fd8
(gdb) x/9w $msp
0x20000fd8: 0x1 0x2 0x3 0x4
0x20000fe8: 0x88160082 0xa53 0x40ce 0x21000000
0x20000ff8: 0x0
(gdb) cont
Continuing.
Breakpoint 2, 0x000040f8 in SVC_Handler ()
(gdb) print $msp
$4 = (void *) 0x20000fd8
(gdb) x/9w $msp
0x20000fd8: 0x1 0x2 0x3 0x4
0x20000fe8: 0x88160082 0xa53 0x40e8 0x21000200
0x20000ff8: 0x0
Assembly program:
.cpu cortex-m0plus
.fpu softvfp
.thumb
.syntax unified
.section .vectors
@ pvStack:
.word 0x20001000
@ pfnReset_Handler:
.word Reset_Handler + 1
@ pfnNMI_Handler:
.word 0
@ pfnHardFault_Handler:
.word 0
@ pfnReservedM12:
.word 0
@ pfnReservedM11:
.word 0
@ pfnReservedM10:
.word 0
@ pfnReservedM9:
.word 0
@ pfnReservedM8:
.word 0
@ pfnReservedM7:
.word 0
@ pfnReservedM6:
.word 0
@ pfnSVC_Handler:
.word SVC_Handler + 1
.section .text
.global Reset_Handler
Reset_Handler:
cpsie i
ldr r0, .stack_start
ldr r2, .stack_last
eors r1, r1
.loop_clear:
str r1, [r0]
adds r0, r0, #4
cmp r0, r2
bne .loop_clear
subs r2, r2, #4
mov sp, r2
movs r0, #1
movs r1, #2
movs r2, #3
movs r3, #4
svc #0
ldr r0, .stack_start
ldr r2, .stack_last
eors r1, r1
.loop_clear2:
str r1, [r0]
adds r0, r0, #4
cmp r0, r2
bne .loop_clear2
mov sp, r2
movs r0, #1
movs r1, #2
movs r2, #3
movs r3, #4
svc #0
.loop:
b .loop
.align 4
.stack_start:
.word 0x20000f00
.stack_last:
.word 0x20000ffc
@ first call - 0x2000fff8 -- should already be aligned
@ second call - 0x2000fffc -- should hit the alignment code
.global SVC_Handler
SVC_Handler:
bx lr
Change-Id: Id0940e6bbd6a59adee1378c0e86fe86830f0c8fc
Signed-off-by: Andrew Ruder <andrew.ruder@elecsyscorp.com>
Cc: Paul Fertser <fercerpav@gmail.com>
Cc: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Cc: Evan Hunter <evanhunter920@gmail.com>
Cc: Jon Burgess <jburgess777@gmail.com>
Reviewed-on: http://openocd.zylin.com/3003
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
Some targets (Cortex M) require more complicated calculations for
turning the stored stack pointer back into a process stack pointer.
For example, the Cortex M stores a bit in the auto-stacked xPSR
indicating that alignment had to be performed and an additional 4
byte padding is present before the exception stacking. This change
only sets up the framework for Cortex-M unstacking and does not
add Cortex-M support.
Note: this also fixes the alignment calculation nearly addressed by
change #2301 entitled rtos/rtos.c: fix stack alignment calculation.
Updated calculation is in rtos_generic_stack_align.
Change-Id: I0f662cad0df81cbe5866219ad0fef980dcb3e44f
Signed-off-by: Andrew Ruder <andrew.ruder@elecsyscorp.com>
Cc: Paul Fertser <fercerpav@gmail.com>
Cc: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Cc: Evan Hunter <evanhunter920@gmail.com>
Cc: Jon Burgess <jburgess777@gmail.com>
Reviewed-on: http://openocd.zylin.com/3002
Reviewed-by: Andreas Fritiofson <andreas.fritiofson@gmail.com>
Tested-by: jenkins
