Greetings All!

I have completed testing, it appears that the smapi error only happens 1 after retrieving mail through a poll, 2 on 64bit arm.

I have tried 32 bit arm works just fine
I have tried 64 bit arm of the following distros, manjaro,debian and ubuntu they all produce the smapi error after first poll and receipt of pkts.

The only difference that happened in testing was the debian and ubuntu trial I had to set perl=0 in order to compile hpt, the manjaro 64 bit perl=1 and built fine.

Not being a programmer, in fact very little knowledge of the subject but 11 installs in 24 hours tells me that perhaps its a problem with 64bit arm?

----
Regards,


Rick Smith (Nitro)

... YKYHW: You consider BBSing better than ice cream.
--- GoldED+/LNX 1.1.5-b20180707
* Origin: ----> Abacus Sysop Point --->>>>bbs.abon.us:2323 (1:105/10.1)

Hello Rick,

Sunday April 11 2021, Rick Smith wrote to All:

I have completed testing, it appears that the smapi error only happens
1 after retrieving mail through a poll, 2 on 64bit arm.

I have tried 32 bit arm works just fine
I have tried 64 bit arm of the following distros, manjaro,debian and ubuntu they all produce the smapi error after first poll and receipt
of pkts.

The only difference that happened in testing was the debian and ubuntu trial I had to set perl=0 in order to compile hpt, the manjaro 64 bit perl=1 and built fine.

Not being a programmer, in fact very little knowledge of the subject
but 11 installs in 24 hours tells me that perhaps its a problem with
64bit arm?

Try to append

-mstructure-size-boundary=64

to OPTCFLAGS in your huskymak.cfg.

You may also try
-mstrict-align

Michael

... node (at) f1042 (dot) ru
--- GoldED+/LNX 1.1.5-b20180707
* Origin: Moscow, Russia (2:5020/1042)

Greetings Michael!

Monday April 12 2021 11:59, you wrote to me about an urgent matter!:

* Forwarded from area 'fidosoft.husky'
Hello Rick,

Sunday April 11 2021, Rick Smith wrote to All:

Not being a programmer, in fact very little knowledge of the
subject but 11 installs in 24 hours tells me that perhaps its a
problem with 64bit arm?

-mstructure-size-boundary=64

to OPTCFLAGS in your huskymak.cfg.

I added this

# C-compiler: optimization
OPTCFLAGS=-c -s -O3 -fomit-frame-pointer -fstrength-reduce -fPIC -mstructure-size-boundary=64

got this during build

gcc: error: unrecognized command-line option ‘-mstructure-size-boundary=64’ make: *** [Makefile:43: adcase.o] Error 1

You may also try
-mstrict-align

I added this

# C-compiler: optimization
OPTCFLAGS=-c -s -O3 -fomit-frame-pointer -fstrength-reduce -fPIC -mstrict-align

Packages built, smapi error returned.


----
Regards,


Rick Smith (Nitro)

... Automate your BBSing! Have your cake and eat it too!
--- GoldED+/LNX 1.1.5-b20180707
* Origin: ----> Abacus Sysop Point --->>>>bbs.abon.us:2323 (1:105/10.1)

Hello Rick,

Monday April 12 2021, Rick Smith wrote to Michael Dukelsky:

I added this

# C-compiler: optimization
OPTCFLAGS=-c -s -O3 -fomit-frame-pointer -fstrength-reduce -fPIC -mstructure-size-boundary=64

got this during build

gcc: error: unrecognized command-line option ‘-mstructure-size-boundary=64’
make: *** [Makefile:43: adcase.o] Error 1

Do you build on your rpi or on a PC? What version of gcc do you use?

Michael

... node (at) f1042 (dot) ru
--- GoldED+/LNX 1.1.5-b20180707
* Origin: Moscow, Russia (2:5020/1042)

Greetings Michael!

Monday April 12 2021 20:20, you wrote to me about an urgent matter!:

* Forwarded from area 'fidosoft.husky'
Hello Rick,

Monday April 12 2021, Rick Smith wrote to Michael Dukelsky:

I added this

# C-compiler: optimization
OPTCFLAGS=-c -s -O3 -fomit-frame-pointer -fstrength-reduce -fPIC
-mstructure-size-boundary=64

Do you build on your rpi or on a PC? What version of gcc do you use?

Yes I build it right on the pi itself, and I have version gcc version 10.2.0 (GCC) installed.



----
Regards,


Rick Smith (Nitro)

... Help save BBSing - Ask your host for INTBBS_WK today!
--- GoldED+/LNX 1.1.5-b20180707
* Origin: ----> Abacus Sysop Point --->>>>bbs.abon.us:2323 (1:105/10.1)

Hello Rick,

Monday April 12 2021, Rick Smith wrote to Michael Dukelsky:

Do you build on your rpi or on a PC? What version of gcc do you
use?

Yes I build it right on the pi itself, and I have version gcc version 10.2.0 (GCC) installed.

I see. There is a possibility that some bug in the code manifests itself only on a 64-bit ARM device. If some owner of such a device finds the bug, it can be fixed. But I am not inclined to buy a Raspberry Pi just to see whether there is such a bug or not, sorry.

In the meantime, you may use a 32-bit OS.

Michael

... node (at) f1042 (dot) ru
--- GoldED+/LNX 1.1.5-b20180707
* Origin: Moscow, Russia (2:5020/1042)

Hi Michael.

13 Apr 21 13:24, you wrote to Rick Smith:

I see. There is a possibility that some bug in the code manifests itself only
on a 64-bit ARM device. If some owner of such a device finds the bug, it can be
fixed. But I am not inclined to buy a Raspberry Pi just to see whether there is
such a bug or not, sorry.

In the meantime, you may use a 32-bit OS.

32bit RPI 2 here, husky defaults with perl, no issues.

gcc version 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04)

Linux 5.4.0-1032-raspi #35-Ubuntu SMP PREEMPT Fri Mar 19 20:56:57 UTC 2021 armv7l GNU/Linux

Distributor ID: Ubuntu
Description: Ubuntu 20.04.2 LTS
Release: 20.04
Codename: focal

'Tommi

--- GoldED+/LNX 1.1.5-b20210402
* Origin: rpi.rbb.bbs.fi (2:221/1)

Hello Rick!

13 Apr 21, Rick Smith wrote to Michael Dukelsky:

device finds the bug, it can be fixed. But I am not inclined to
buy a Raspberry Pi just to see whether there is such a bug or
not, sorry.

Bummer... I would donate a Pi if someone were inclined to research it,
I do not possess such abilities but like hpt enough to contribute

Both of you could think about if a remote ssh session would be an option.

Regards

Kai

--- GoldED+/LNX 1.1.4.7
* Origin: Monobox (2:240/77)

Greetings Kai!

Wednesday April 14 2021 13:06, you wrote to me about an urgent matter!:

Hello Rick!

13 Apr 21, Rick Smith wrote to Michael Dukelsky:

device finds the bug, it can be fixed. But I am not inclined to
buy a Raspberry Pi just to see whether there is such a bug or
not, sorry.

Bummer... I would donate a Pi if someone were inclined to
research it, I do not possess such abilities but like hpt enough
to contribute

Both of you could think about if a remote ssh session would be an
option.

Also an option if someone was so inclined.

----
Regards,


Rick Smith (Nitro)

... Hmmm ... isn't BBSing just full of surprises?
--- GoldED+/LNX 1.1.5-b20180707
* Origin: ----> Abacus Sysop Point --->>>>bbs.abon.us:2323 (1:105/10.1)

Michael wrote (2021-04-13):

I see. There is a possibility that some bug in the code manifests itself only on a 64-bit ARM device. If some owner of such a device finds the
bug, it can be fixed. But I am not inclined to buy a Raspberry Pi just to see whether there is such a bug or not, sorry.

In the meantime, you may use a 32-bit OS.

You could use qemu to emulate an (virtual) aarch64 machine (or an raspberry)

https://futurewei-cloud.github.io/ARM-Datacenter/qemu/how-to-launch-aarch64-vm/

or someone could set it up and give you remote access.

---
* Origin: . (2:280/464.47)