Page 5 of 5

Re: The new network stack

Posted: Mon Sep 11, 2017 11:17 pm
by hidnplayr
Fixed in #6971

Re: The new network stack

Posted: Fri Feb 16, 2018 4:19 pm
by 0CodErr
Still have a question is it an issue or not? http://board.kolibrios.org/viewtopic.ph ... 232#p69046 Maybe I just don't understand something?

Re: The new network stack

Posted: Sat Apr 21, 2018 4:05 pm
by hidnplayr
0CodErr wrote:

Code: Select all

======================================================================
========== Function 74, Subfunction 8, Get TX byte counter. ==========
======================================================================
Parameters:
  * eax = 74 - function number
  * bl = 8 - subfunction number
  * bh = device number
Returned value:
  * eax = Number of bytes sent since device start (lower dword)
                  -1 on error
  * ebx = Number of bytes sent since device start (higher dword)
======================================================================
========== Function 74, Subfunction 9, Get RX byte counter. ==========
======================================================================
Parameters:
  * eax = 74 - function number
  * bl = 9 - subfunction number
  * bh = device number
Returned value:
  * eax = Number of bytes received since device start (lower dword)
                  -1 on error
  * ebx = Number of bytes received since device start (higher dword)
What if number of bytes is for example 0x1ffffffff?
Yes, it is an issue.
0CodErr wrote: Also the function below may have the same problem.

Code: Select all

======================================================================
==== Function 76, Protocol 0 - Ethernet, Subfunction 0, Read MAC. ====
======================================================================
Parameters:
  * eax = 76 - function number
  * high half of ebx = 0 (Ethernet)
  * bh = device number
  * bl = 0 (Read MAC)
Returned value:
  * eax = -1 on error, otherwise lower bits of MAC
  * bx = upper bits of MAC
Edit: Yes, it is also an issue, fix could be to use only lower 2 bytes of EAX for MAC on success.

Re: The new network stack

Posted: Sat Apr 21, 2018 4:37 pm
by 0CodErr
hidnplayr wrote: upper 2 bytes of EAX will be 0 when there is no error
I don't understand why.
Since the MAC address is a 6-byte value.
According to the description of the function EAX contains lower(not higher) bits of MAC.

Re: The new network stack

Posted: Sat Apr 21, 2018 5:44 pm
by hidnplayr
0CodErr: You are correct, I did not read my own documentation. See previous post.

Re: The new network stack

Posted: Wed Oct 31, 2018 6:26 pm
by Leency
Yesterday hidnplayr and I have made couple of tests at my main PC. This computer has several issues working with Kolibri. For example, I must use keyboard only in Kolibri (USB mouse and keyboard don't work).
As a result:
- lan card Intel I218-V was added to list of supported devices (integrated eth at Asus Z97-e motherboard)
- found an speed calculation issue in NETSTAT which was fixed by hidnplayr (thanks a lot)
- http.obj buffer was increased which must increase download speed (but it didn't help me much)

The average download speed in Kolibri was about 180 Kb.

Now I have bought 4 lan cards to test them in Kolibri. Not sure the all are in working state. But lets check!