On 12/5/24 09:29, Cédric Le Goater wrote:
On 12/5/24 18:09, Pierrick Bouvier wrote:
On 12/5/24 00:18, Cédric Le Goater wrote:
Hello Pierrick,
Thanks for improving the documentation !
On 12/4/24 21:39, Pierrick Bouvier wrote:
Signed-off-by: Pierrick Bouvier <pierrick.bouv...@linaro.org>
---
docs/system/arm/fby35.rst | 3 +++
1 file changed, 3 insertions(+)
diff --git a/docs/system/arm/fby35.rst b/docs/system/arm/fby35.rst
index bf6da6baa2a..ed9faef363c 100644
--- a/docs/system/arm/fby35.rst
+++ b/docs/system/arm/fby35.rst
@@ -45,3 +45,6 @@ process starts.
$ screen /dev/tty0 # In a separate TMUX pane, terminal window, etc.
$ screen /dev/tty1
$ (qemu) c # Start the boot process once screen is setup.
+
+This machine model supports emulation of the boot from the CEO flash device by
+setting option ``execute-in-place``.
Could we add a link to the "Boot options" section of the
"Aspeed family boards" document instead ?
The problem is that the aspeed board has other options (fmc-model, spi-model),
so it would be confusing for the reader.
ok. Just to clarify, the Aspeed machines always boot from CE0 (if you
exclude rainier).
Ok, I was not aware of that.
I'm not familiar with those machine models, and just have been a
documentation updater here, so my knowledge is limited.
With the "execute-in-place" option, the CPU fetches instructions to
execute by reading CE0 and not from a preloaded ROM initialized at
machine init time. I think it is good to mention the difference and
the impact on performance : It is more accurate modeling but slower.
I'll add this.
This machine model supports emulation of the boot from the CE0 flash
device by setting option ``execute-in-place``. When using this option,
the CPU fetches instructions to execute by reading CE0 and not from a
preloaded ROM initialized at machine init time. As a result, execution
will be slower.
Is that ok for you?
It's worth repeating what the option means IMHO.
sure. np.
Thanks,
C.
