Øyvind Harboe [Tue, 9 Feb 2010 13:26:57 +0000 (14:26 +0100)]
arm11: improve performance using minidriver hook
zy1000 performance for GDB load went from 100kBytes/s
to 300kBytes/s @ 8 MHz by implementing the inner loop
of unack arm11 memory writes directly on top of the hw
fifo.
David Brownell [Mon, 22 Feb 2010 01:19:08 +0000 (17:19 -0800)]
ft2232 table init cleanup
Use labeled initializers in the table of layouts instead of
positional ones. This ls cleaner and less error prone, plus
it simplifies patches which add members to these structure.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Mon, 22 Feb 2010 00:55:17 +0000 (16:55 -0800)]
Developer's Guide: refresh release procedures
Be a closer match to what I've actually done for the past few cycles.
In particular, hold off pushing repository updates until after the
packages are published, as part of opening the merge window, and
mention the utility commands which actually create the archives.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Sun, 21 Feb 2010 22:58:16 +0000 (14:58 -0800)]
ADIv5: relocate memacess_tck cycles
When using an AP to access a memory (or a memory-mapped register),
some extra TCK (assuming JTAG) cycles should be added to ensure
the AP has enugh time to complete that access before trying to
collect the response.
The previous code was adding these cycles *before* trying to
access (read or write) data to that address, not *after*. Fix
by putting the delays in the right location.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Sun, 21 Feb 2010 22:56:56 +0000 (14:56 -0800)]
ADIv5: remove ATOMIC/COMPOSITE interface mode
This removes context-sensitivity from the programming interface and makes
it possible to know what a block of code does without needing to know the
previous history (specifically, the DAP's "trans_mode" setting).
The mode was only set to ATOMIC briefly after DAP initialization, making
this patch be primarily cleanup; almost everything depends on COMPOSITE.
The transactions which shouldn't have been queued were already properly
flushing the queue.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Sun, 21 Feb 2010 22:54:54 +0000 (14:54 -0800)]
ARM: ADIv5, deadcode cleanup
I have no idea what the scan_inout_check() was *expecting* to achieve by
issuing a read of the DP_RDBUFF register. But in any case, that code was
clearly never being called ("invalue" always NULL) ... so remove it, and
the associated comment.
Also rename it as ap_write_check(), facilitating a cleanup of its single
call site by removing constant parameters.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Sun, 21 Feb 2010 22:53:15 +0000 (14:53 -0800)]
ARM: ADIv5 code shrinkage, cleanup
adi_jtag_dp_scan_u32() now wraps adi_jtag_dp_scan(), removing
code duplication. Include doxygen for the former. Comment
some particularly relevant points. Minor fault handling fixes
for both routines: don't register a callback that can't run,
or return ERROR_OK after an error.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Sun, 21 Feb 2010 22:51:19 +0000 (14:51 -0800)]
ADIv5 clean up AP fault handling
Pass up fault codes from various routines, so their callers
can clean up after failures, and remove the FIXME comments
highlighting those previously goofy code paths.
David Brownell [Sun, 21 Feb 2010 22:48:04 +0000 (14:48 -0800)]
ADIv5 clean up AP selection and register caching
Handling of AP (and AP register bank) selection, and cached AP
registers, is pretty loose ... start tightening it:
- It's "AP bank" select support ... there are no DP banks. Rename.
+ dap_dp_bankselect() becomes dap_ap_bankselect()
+ "dp_select_value" struct field becomes "ap_bank_value"
- Remove duplicate AP cache init paths ... only use dap_ap_select(),
and don't make Cortex (A8 or M3) cores roll their own code.
- For dap_ap_bankselect(), pass up any fault code from writing
the SELECT register. (Nothing yet checks those codes.)
- Add various bits of Doxygen
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Sun, 21 Feb 2010 22:34:33 +0000 (14:34 -0800)]
ARM: keep a handle to the PC
Keep a handle to the PC in "struct arm", and use it.
This register is used a fair amount, so this is a net
minor code shrink (other than some line length fixes),
but mostly it's to make things more readable.
For XScale, fix a dodgy sequence while stepping. It
was initializing a variable to a non-NULL value, then
updating it to handle the step-over-active-breakpoint
case, and then later testing for non-NULL to see if
it should reverse that step-over-active logic. It
should have done like ARM7/ARM9 does: init to NULL.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Sun, 21 Feb 2010 22:32:34 +0000 (14:32 -0800)]
ARM DPM: support adding/removing HW breakpoints
Generalize the core of watchpoint setup so that it can handle
breakpoints too. Create breakpoint add/remove routines which
will use that, and hook them up to target types which don't
provide their own breakpoint support (nothing, yet).
This suffices for hardware-only breakpoint support. The ARM11
code will be able to switch over to this without much trouble,
since it doesn't yet handle software breakpoints. Switching
Cortex-A8 will be a bit more involved.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Sun, 21 Feb 2010 22:28:53 +0000 (14:28 -0800)]
ARMv7-M: start using "struct arm"
This sets up a few of the core "struct arm" data structures so they
can be used with ARMv7-M cores. Specifically, it:
- defines new ARM core_modes to match the microcontroller modes
(e.g. HANDLER not IRQ, and two types of thread mode);
- Establishes a new microcontroller "core_type", which can be
used to make sure v7-M (and v6-M) cores are handled right;
- adds "struct arm" to "struct armv7m" and arranges for the
target_to_armv7m() converter to use it;
- sets up the arm.core_cache and arm.cpsr values
- makes the Cortex-M3 code maintain arm.map and arm.core_mode.
This is currently set up as a parallel data structure, primarily to
minimize special cases for the semihosting support with microcontroller
profile cores.
Later patches can rip out the duplicative ARMv7-M support and start
reusing core ARM code.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Sun, 21 Feb 2010 04:47:38 +0000 (20:47 -0800)]
arm920t line length cleanup
The recent patch to fixbreakpoints and dcache handling added
a bunch of overlong lines (80+ chars) ... shrink them, and do
the same to a few lines which were already overlong.
Also add a few FIXME comments to nudge (a) replacement of some
magic numbers with opcode macros, which will be much better at
showing what's actually going on, and (b) correct return codes.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Øyvind Harboe [Fri, 12 Feb 2010 07:39:56 +0000 (08:39 +0100)]
arm720t: virt2phys callback added
This is a copy and paste of arm926ejs. Not tested, but
ready for testing at least. There is a good chance that
it will work if the generic armv4_5 fn's are robust enough...
Øyvind Harboe [Wed, 10 Feb 2010 19:04:00 +0000 (20:04 +0100)]
arm11: fix another infinite loop bug
reset init would get stuck in an infinite loop when
e.g. khz was too high. Added timeout. This is a copy
of paste of a number of such bugfixes in the arm11
code.
Arm11 code reviewed for further such infinite loop bugs
and I couldn't find any more. Xing fingers it's the last
one...
David Brownell [Sun, 7 Feb 2010 03:19:25 +0000 (19:19 -0800)]
Re-title Developer's Guide
The Doxygen output was previously titled "OpenOCD Reference Manual",
which was quite misleading ... the User's Guide is the reference
manual which folk should consult about how to use the software.
Just rename it to match how it's been discussed previously, and to
bring out its intended audience: developers of this software. As a
rule, Doxygen is only for folk who work with the code it documents.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Thu, 4 Feb 2010 22:39:51 +0000 (14:39 -0800)]
ARMv7-M: make DAP commands verify target is an ARMv7-M
Init the ARMv7-M magic number. Define predicate verifying it.
Use it to resolve a lurking bug/FIXME: make sure the ARMv7-M
specific DAP ops reject non-ARMv7-M targets.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Spencer Oliver [Thu, 4 Feb 2010 10:33:33 +0000 (10:33 +0000)]
CMD: duplicate cmd error msg
When registering cmds we report duplicate attempts to register a cmd
as a LOG_ERROR.
Some situations need this, such as when registering dual flash banks.
http://www.mail-archive.com/openocd-development@lists.berlios.de/msg11152.html
Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
David Brownell [Tue, 2 Feb 2010 18:53:13 +0000 (10:53 -0800)]
NOR: User's Guide updates
Remove long-obsolete text about "erase_check" affecting "flash info" output.
Move parts of that text to "protect_check", where it's still relevant; and
update the "flash info" description to mention the issue.
(This is still awkward. It might be best to make "protect_check" mirror
"erase_check" by dumping what it finds, so "flash info" doesn't dump any
potentially-stale cache info.)
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Edgar Grimberg [Tue, 2 Feb 2010 12:17:26 +0000 (13:17 +0100)]
tcl/str7x: Reset init unlocks the flash
For STR7x flash, the device cannot be queried for the protect status.
The solution is to remove the protection on reset init. The driver
also initialises the sector protect field to unprotected.
[dbrownell@users.sourceforge.net: line length shrinkage]
Signed-off-by: Edgar Grimberg <edgar.grimberg@zylin.com> Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Edgar Grimberg [Tue, 2 Feb 2010 09:39:52 +0000 (10:39 +0100)]
flash/str7x: After reset init the flash is unlocked
The default state of the STR7 flash after a reset init is unlocked.
The information in the flash driver now reflects this.
The information about the lock status cannot be read from the
flash chip, so the user is informed that flash info might not
contain accurate information.
[dbrownell@users.sourceforge.net: line length shrinkage]
Signed-off-by: Edgar Grimberg <edgar.grimberg@zylin.com> Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Nicolas Pitre [Tue, 2 Feb 2010 05:05:42 +0000 (00:05 -0500)]
ARM semihosting: fix EOF handling with SYS_READ
The semihosting interface has a strange convention for read/write where
the unused amount of buffer must be returned. We failed to return the
total buffer size when the local read() call returned 0.
Øyvind Harboe [Thu, 21 Jan 2010 15:41:54 +0000 (16:41 +0100)]
telnet: fix strage blank spaces at beginning of telnet lines
Sometimes we saw two strange blank spaces at the beginning
of the telnet lines.
progress
ogress
>
This patch fixes this problem:
progress
progress
>
The code changes are *reasonably* clean, but perhaps it could be
made a bit more elegant, but I didn't want to change things after
I finished diagnosis/testing & submitting the patch.
The problem was that logging can send the text and the newline
separately in two different requests and the telnet code would
incorrectly remove the prompt from the end of a line.
David Brownell [Sun, 31 Jan 2010 08:26:21 +0000 (00:26 -0800)]
Subject: ADIv5: fix more diagnostics
If the MEM-AP cache is invalid, don't display it; just report that
invalidity as an error. (This bug has been observed with "mdw 0 32"
after just a "reset halt". Some code is being wrongly bypassed...)
If it's valid, display that cache at DEBUG level, not ERROR. Also,
don't assume it's an AHB-AP; it could be another flavor of MEM-AP.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Make messages reference "DAP" if they're actually transport-agnostic, or
"JTAG-DP" when they're JTAG-specific. Saying SWJ-DP is often wrong (on
most Cortex-A8 chips) and is confusing even if correct (since we don't
yet support SWD).
Rename a JTAG-specific routine to jtagdp_transaction_endcheck() to highlight
that it's JTAG-specific, and that identify DAP clients undesirably depending
on JTAG. (They will all need to change for SWD support.)
Shrink a few overlong lines of code. Copy a comment from code removed
in a previous patch (for the ARMv7-M "dap baseaddr" command).
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Alex Austin [Fri, 29 Jan 2010 06:41:44 +0000 (00:41 -0600)]
Clang buildfixes
Building with clang took a few very small changes. The change to
helper/log.h is because clang doesn't like an expression where the
result is unused. In helper/system.h, I just defined true and false
since clang doesn't have them builtin.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Thu, 28 Jan 2010 22:03:29 +0000 (14:03 -0800)]
doc clarifications for server flags
The "-f" is a shortcut for "-c" ... and providing any "-c" options
means the "openocd.cfg" file isn't implicitly used. Both the User's
Guide and the manual page were weak on these points, which has led
to some confusion.
Also update the manual page to include highlights of the search path
mechanism, including the facts that it exists and that "-s" adds to it.
Stop saying only the current directory is involved; the OpenOCD
script library is quite significant.
(Missing: complete manpage coverage of the search path, including a
FILES section listing all components and saying where the script
library is found.)
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Thu, 28 Jan 2010 21:58:20 +0000 (13:58 -0800)]
ARM: reference DPM defn from v6/v7 arch spec
The term "DPM" is probably not well known ("Device Power Management"?),
so identify its source in the current ARM architecture specification.
It's relevant to ARMv6, ARMv7-A, and ARMv7-R ... but not "M" profiles.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Spencer Oliver [Wed, 27 Jan 2010 21:20:18 +0000 (21:20 +0000)]
ARM semihosting: fix writing to stdout
SYS_FLEN would be called before a write on a descriptor to check its size.
Currently lseek would fail with -1 when given the stdout/stderr descriptor.
Changing to use fstat seems to be the standard way of handling this.
Signed-off-by: Spencer Oliver <ntfreak@users.sourceforge.net>
David Brownell [Wed, 27 Jan 2010 21:47:48 +0000 (13:47 -0800)]
Cortex-M3: report lockup, and recover
ARMv7-M defines a "lockup" state that's entered in certain double
fault sequences which can't be recovered from without external help.
OpenOCD has previously ignored this.
Issue a diagnostic saying the chip has locked up, and force exit
from this state by halting the core. It's not clear this is the
best way to handle lockup; but there should now be less confusion.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
simon qian [Sun, 17 Jan 2010 20:56:08 +0000 (04:56 +0800)]
SVF: insert space before '(' and after ')'
See http://forum.sparkfun.com/viewtopic.php?p=90983#90983 for discussion;
basically, the SVF parser wrongly expects "TDI (123)" but the space is
optional and it should accept "TDI(123)" too.
In the same way, "TDI(123)TDO(456)" should work too.
Rather than update the command parsing, this just makes sure the expected
spaces are present.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Sat, 23 Jan 2010 06:37:15 +0000 (22:37 -0800)]
ARM11: fix breakpoints with GDB
This fixes a bug whereby GDB's breakpoints weren't activated.
The root cause is a confused interface to resume(). Fix by
almost ignoring the "handle breakpoints" parameter; it only
seems related to the case of skipping breakpoint-at-PC.
Update a few coments to clarify what's happening.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Fri, 22 Jan 2010 00:45:00 +0000 (16:45 -0800)]
Various doc/comment updates
Doxygen: don't be needlessly verbose; alphabetically sort members
TODO: add random bits; clarify which manuals are referenced
ARM disassembler: mention a few opcodes that still aren't handled
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Fri, 22 Jan 2010 00:15:41 +0000 (16:15 -0800)]
User's Guide secton on target hardware setup
Highlight the needs to properly jumper development boards; to
make the OpenOCD configuration match the jumpering; and to have
a usable "reset-init" method when debugging early boot code.
Specific mention of the "ATX Mode" that seems useful on
many i.MX boards, forcing NAND boot.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Øyvind Harboe [Thu, 21 Jan 2010 09:16:42 +0000 (10:16 +0100)]
target: print reason why GDB halts
If GDB halts unexpectedly, print reason: srst assert or power
out detected.
If polling fails, then things are a bit trickier. We do not
want to spam telnet or the log with polling failed messages.
Leave that case be w/a comment in a code for now.
David Brownell [Wed, 20 Jan 2010 18:43:32 +0000 (10:43 -0800)]
gdb_server: correctly report flash sector sizes
Report each region of same-size sectors separately, instead of
incorrectly reporting that every sector has the same size.
This is a longstanding bug on NOR flash chips with non-uniform
sector sizes. It was largely hidden by other bugs in flash
handling. When some of those were recently fixed, this one was
exposed as a regression on str710.
[oyvind.harboe@zylin.com: update the loop to behave on str7 ]
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com> Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
David Brownell [Wed, 20 Jan 2010 08:04:17 +0000 (00:04 -0800)]
User's guide: mention lpc2000 checksum issue
Folk almost certainly want to have OpenOCD compute the checksum
when they modify the vector table. However, that almost guarantees
that "verify_image" will fail.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>