|
Re: RFC: BSD Socket (like) API
Hello Luiz Augusto,
By
Paul Sokolovsky
·
#340
·
|
|
Re: Reminder: advanced compiler novelties are upon us
Hi,
Szymon Janc <szymon.janc@...> writes:
Adding the warnings, I think, is the best way forward.
Keeping the checks after the NULL pointer is accessed gains us little,
we are already in
Hi,
Szymon Janc <szymon.janc@...> writes:
Adding the warnings, I think, is the best way forward.
Keeping the checks after the NULL pointer is accessed gains us little,
we are already in
|
By
Vinicius Costa Gomes
·
#339
·
|
|
Re: Reminder: advanced compiler novelties are upon us
Thanks Paul for your post. There is an issue in the Atmel SAM GMAC Ethernet driver I was working on which made me scratch my head for a while. I was sure it's a gcc bug but in fact I was
Thanks Paul for your post. There is an issue in the Atmel SAM GMAC Ethernet driver I was working on which made me scratch my head for a while. I was sure it's a gcc bug but in fact I was
|
By
Piotr Mienkowski
·
#338
·
|
|
Re: Reminder: advanced compiler novelties are upon us
Indeed.
Even on systems where you do have FLASH at NULL it is often merely a bootloader that the application shouldn't be touching anyway so it doesn't prevent the application running with an MPU or
Indeed.
Even on systems where you do have FLASH at NULL it is often merely a bootloader that the application shouldn't be touching anyway so it doesn't prevent the application running with an MPU or
|
By
Daniel Thompson <daniel.thompson@...>
·
#337
·
|
|
Re: Reminder: advanced compiler novelties are upon us
Hi,
(resending to mailing list)
-Wnull-dereference has effect only if -fdelete-null-pointer-checks is set
So we could either make GCC warn when it is eliminating code under NULL check
or tell it
Hi,
(resending to mailing list)
-Wnull-dereference has effect only if -fdelete-null-pointer-checks is set
So we could either make GCC warn when it is eliminating code under NULL check
or tell it
|
By
Szymon Janc
·
#336
·
|
|
Re: Reminder: advanced compiler novelties are upon us
Hi Marcus,
I rather take slightly more code size than risking to have weird bugs
because some null checks are removed by compiler.
Cheers,
Jukka
Hi Marcus,
I rather take slightly more code size than risking to have weird bugs
because some null checks are removed by compiler.
Cheers,
Jukka
|
By
Jukka Rissanen
·
#335
·
|
|
Re: Reminder: advanced compiler novelties are upon us
<jukka.rissanen@...> wrote:
We should understand the code size implication of that decision before
changing the out of box behaviour.
Another option might be to have a test compile run
<jukka.rissanen@...> wrote:
We should understand the code size implication of that decision before
changing the out of box behaviour.
Another option might be to have a test compile run
|
By
Marcus Shawcroft <marcus.shawcroft@...>
·
#334
·
|
|
Re: Reminder: advanced compiler novelties are upon us
Hi Paul,
Thanks for this info. I have hit by this actually, but did not know the
root cause at that time. When running the failed program, it really
looked there was a compiler bug and actually IMHO
Hi Paul,
Thanks for this info. I have hit by this actually, but did not know the
root cause at that time. When running the failed program, it really
looked there was a compiler bug and actually IMHO
|
By
Jukka Rissanen
·
#333
·
|
|
Re: Minimal Zephyr build
Indeed, I missed that CONFIG_PRINTK=y was set in the .config without me setting it. I can drop 9124 bytes out of my mcuboot build by just turning off printk. I could probably drop more by removing the
Indeed, I missed that CONFIG_PRINTK=y was set in the .config without me setting it. I can drop 9124 bytes out of my mcuboot build by just turning off printk. I could probably drop more by removing the
|
By
David Brown
·
#332
·
|
|
Re: Reminder: advanced compiler novelties are upon us
Hi,
-fno-delete-null-pointer-checks can be set to disable this optimisation.
Other option would be to enable -Wnull-dereference which would warn about
dereferencing NULL pointer (this has effect
Hi,
-fno-delete-null-pointer-checks can be set to disable this optimisation.
Other option would be to enable -Wnull-dereference which would warn about
dereferencing NULL pointer (this has effect
|
By
Szymon Janc
·
#331
·
|
|
Re: Minimal Zephyr build
Printk() doesn't implement either of these. You can see its implementation in misc/printk.c
The problem really isn't the code. Most of the size comes from all the format strings all over the codebase.
Printk() doesn't implement either of these. You can see its implementation in misc/printk.c
The problem really isn't the code. Most of the size comes from all the format strings all over the codebase.
|
By
Boie, Andrew P
·
#330
·
|
|
Re: Minimal Zephyr build
There are ways to reduce printf/printk code foot print by reducing functionality:
* don't need octal printing (unless application requires it)
* don't need float/double printing (unless application
There are ways to reduce printf/printk code foot print by reducing functionality:
* don't need octal printing (unless application requires it)
* don't need float/double printing (unless application
|
By
Chuck Jordan <Chuck.Jordan@...>
·
#329
·
|
|
Re: Minimal Zephyr build
Same on stm32: UART driver uses HAL for driver initialization.
I made an update of stm32cube today and Low Level API is now supported on stm32f4/f3/l4 based boards.
We can save around 1K moving from
Same on stm32: UART driver uses HAL for driver initialization.
I made an update of stm32cube today and Low Level API is now supported on stm32f4/f3/l4 based boards.
We can save around 1K moving from
|
By
Erwan Gouriou
·
#328
·
|
|
Re: Minimal Zephyr build
By
Maureen Helm
·
#327
·
|
|
Re: Minimal Zephyr build
[Boie, Andrew P]
I did some footprint testing with MCUBoot and Arduino 101. I disable CONFIG_MULTITHREADING and a few other unnecessary things.
With the UART and CONFIG_PRINTK turned off, the whole
[Boie, Andrew P]
I did some footprint testing with MCUBoot and Arduino 101. I disable CONFIG_MULTITHREADING and a few other unnecessary things.
With the UART and CONFIG_PRINTK turned off, the whole
|
By
Boie, Andrew P
·
#326
·
|
|
Re: Minimal Zephyr build
If you disable CONFIG_MULTITHREADING, the scheduler goes away. The
kernel operates in coop-mode only with a single thread, which is
basically "never preempt me, except by interrupts and when
If you disable CONFIG_MULTITHREADING, the scheduler goes away. The
kernel operates in coop-mode only with a single thread, which is
basically "never preempt me, except by interrupts and when
|
By
Benjamin Walsh <benjamin.walsh@...>
·
#325
·
|
|
Re: BLE attribute handle
Hi Anila,
I guess you are not familiar with GATT/ATT, the attribute handle comes
from the discovered (bt_gatt_discover) attributes, in zephyr we
represent them as bt_gatt_attr which happens to
Hi Anila,
I guess you are not familiar with GATT/ATT, the attribute handle comes
from the discovered (bt_gatt_discover) attributes, in zephyr we
represent them as bt_gatt_attr which happens to
|
By
Luiz Augusto von Dentz
·
#324
·
|
|
Re: undefined reference to `_legacy_sleep'
Thanks all, it works
I just flash again and again.
Thanks all, it works
I just flash again and again.
|
By
kk <pinganddu90@...>
·
#323
·
|
|
Re: undefined reference to `_legacy_sleep'
OK, thanks.
By
kk <pinganddu90@...>
·
#322
·
|
|
Re: undefined reference to `_legacy_sleep'
Hi Buriez, thanks first!
1. Yes I have set the 115200 for minicom.
2. I have export the "ZEPHYR_FLASH_OVER_DFU=y" and on zephyr git master
but nothing on minicom
My connect picture:
Why I can't
Hi Buriez, thanks first!
1. Yes I have set the 115200 for minicom.
2. I have export the "ZEPHYR_FLASH_OVER_DFU=y" and on zephyr git master
but nothing on minicom
My connect picture:
Why I can't
|
By
kk <pinganddu90@...>
·
#321
·
|