BTW, how to test semihosting via GDB is simple:
Assuming you have a STM32 semihosted application (in .elf) and openocd is
running:
arm-none-eabi-gdb
...
target remote :
monitor arm semihosting enable
monitor arm semihosting_fileio enable
load
continue
...
Now your file IO is done through
Hi Liviu,
I know. Thanks for the discussion in any case.
---
** [tickets:#232] semihosting_fileio read return wrong number of read bytes**
**Status:** new
**Milestone:** 0.9.0
**Created:** Sun Apr 21, 2019 04:18 PM UTC by hogthrob
**Last Updated:** Sun Apr 21, 2019 09:53 PM UTC
**Owner:** nobod
That's for the project maintainers to decide, I'm not one of them.
---
** [tickets:#232] semihosting_fileio read return wrong number of read bytes**
**Status:** new
**Milestone:** 0.9.0
**Created:** Sun Apr 21, 2019 04:18 PM UTC by hogthrob
**Last Updated:** Sun Apr 21, 2019 09:48 PM UTC
**Owne
I will not use Gerrit, as it exposes more personal information than I want it
to. It not even allows me to delete my account, should have checked this
before.
Sorry, but take my patch or leave it. It is just two lines.
---
** [tickets:#232] semihosting_fileio read return wrong number of
Well, we both are right, sort of (you more, I less, the fix: absolutely):
https://github.com/db4ple/openocd/blob/0a87bac689a271c8e6fb97a48ee23e4d019a4438/src/target/semihosting_common.c#L789
You are right, it is not directly refering to the GDB protocol.
In fact my assumptions were wrong, t
Hello,
Thank you for the report and the patch.
On Sun, Apr 21, 2019 at 08:31:05PM -, hogthrob wrote:
> I have no idea how to use Gerrit, is there some kind of instructions
> available.
There certainly is, the file is called HACKING in the root of the
OpenOCD sources.
Fortunately, it doesn'
If you mean the comments in semihosting_common.c, the SEMIHOSTING_SYS_READ
case, they do not refer to the gdb client protocol, but to the arm semihosting
protol (I contributed that file, and the large comment blocks were copied from
the arm semihosting pdf file).
As for the SF behaviour, no com
Sure. In order to understand this patch, you have to read the comments given on
first half of the SYS_READ in the same file somewhere above explaining the GDB
protocol return value for SYS_READ
The GDB remote semihosting protocol implements SYS_READ but returns the
(requested bytes - actually
I confirm that for partial reads the SEMIHOSTING_SYS_READ call should return
the number of remaining bytes, and the regular (not the fileio) case does this.
However I do not have experience with the fileio use case. Your patch seems ok,
but I'd like someone more experienced to confirm this.
Pe
---
** [tickets:#232] semihosting_fileio read return wrong number of read bytes**
**Status:** new
**Milestone:** 0.9.0
**Created:** Sun Apr 21, 2019 04:18 PM UTC by hogthrob
**Last Updated:** Sun Apr 21, 2019 04:18 PM UTC
**Owner:** nobody
The current implementation of read for semihosting_f
10 matches
Mail list logo