why 74.-1 returns -1 when there are exists Active Network Device?
Code and screen attached.
Tested in QEMU.
The new network stack
-
- Attachments
-
-
74.-1.7z (461 Bytes)Downloaded 460 times
-
74.png (40.33 KiB)Viewed 10900 times
-
Из хаоса в космос
I used ebx in kernel instead of bl, like for the rest of the functions. It was illogical so I fixed it.
"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius -- and a lot of courage -- to move in the opposite direction." Albert Einstein
Now function returns 1 if there is no device and 2 if one.
Clarify, please, is this behaviour OK?
Clarify, please, is this behaviour OK?
- Attachments
-
-
2.png (39.16 KiB)Viewed 10877 times
-
1.png (53.75 KiB)Viewed 10877 times
-
Из хаоса в космос
To view the status of the stack, please look at netstat program instead
You'll notice that there is always a loopback device when stack is active. It is counted as one device.
You'll notice that there is always a loopback device when stack is active. It is counted as one device.
"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius -- and a lot of courage -- to move in the opposite direction." Albert Einstein
Using netstat program is clear. But I need to check network card status for some programs, like HTMLv and Liza. If there is active network card no need to run downloader in HTMLv and try to connect to server in Liza.hidnplayr wrote:To view the status of the stack, please look at netstat program instead
Из хаоса в космос
You should try to connect, and work with the returned error codes insteadIf there is active network card no need to run downloader in HTMLv and try to connect to server in Liza.
"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius -- and a lot of courage -- to move in the opposite direction." Albert Einstein
OK, I'll try. Thanks
Из хаоса в космос
On one of my parents' computers, the network card is not always found/initialized:PCIDEV finds the card on resolution 1024x768, but doesn't find it on resolution 1280x1024. Boardlog shows that the card completes initialization on resolution 1024x768. On resolution 1280x1024 the initialization is started, but never completes (NETCFG is stuck with 99% CPU usage, and cannot be killed). Night build r3840, kernel SVN r3837.
Output of boardlog + PCIDEV on 1024x768: Output of boardlog + PCIDEV on 1280x1024: Photo of NETCFG stuck with 99% CPU usage:
Code: Select all
10ec 8136 03 00 00 02 02 00 00 0b REALTEK SEMICONDUCTOR CORP. Network - Ethernet
Output of boardlog + PCIDEV on 1024x768: Output of boardlog + PCIDEV on 1280x1024: Photo of NETCFG stuck with 99% CPU usage:
Spoiler:
The bug is possibly fixed in latest build, I should however take the time to review all code someday...Leency wrote:Kolibri build #3870, Qemu 0.15.1
1. Open /sys/HTMLv
2. Click "Home" button from toolbar
3. Click link "Cтатьи"
Now you can not close any program. You may open a new one, but there is no way to close them.
See PCIDEV.png and Boardlog.png for more information. This is a very serious bug, could you please, fix it?
"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius -- and a lot of courage -- to move in the opposite direction." Albert Einstein
>> The bug is possibly fixed in latest build
Sorry, not fixed.
Kolibri build #3886, Qemu 0.15.1.
Sorry, not fixed.
Kolibri build #3886, Qemu 0.15.1.
Из хаоса в космос
Not fixed in VirtualBox too #4016
- Attachments
-
-
Screenshot_3.png (55.05 KiB)Viewed 10623 times
-
Из хаоса в космос
Leency, have patience
I'll let you know when it is fixed.
PS: I have gained some insight on what the exact problem could be but have not made any attempt to fix it yet.
PSS: If it annoys you too much while testing, edit SOCKET_process_end in socket.inc, to ret immedately. This way sockets will not be closed, but system will not hang either
I'll let you know when it is fixed.
PS: I have gained some insight on what the exact problem could be but have not made any attempt to fix it yet.
PSS: If it annoys you too much while testing, edit SOCKET_process_end in socket.inc, to ret immedately. This way sockets will not be closed, but system will not hang either
"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius -- and a lot of courage -- to move in the opposite direction." Albert Einstein
Sorry, for my molestation.
Из хаоса в космос
Hi, Jeffrey
Some pages are not fully loaded in HTMLv. You can see it at our homepage (see scr).
Please, contact me if you will need additional information or testing.
Another strange thing: I saw pages, where was strlen(bufpointer) > bufsize.
For example, http://websvn.kolibrios.org/ page. Difference was about 500 byte. Why is it so?
Some pages are not fully loaded in HTMLv. You can see it at our homepage (see scr).
Please, contact me if you will need additional information or testing.
Another strange thing: I saw pages, where was strlen(bufpointer) > bufsize.
For example, http://websvn.kolibrios.org/ page. Difference was about 500 byte. Why is it so?
- Attachments
-
-
Screenshot_1.png (19.49 KiB)Viewed 10330 times
-
Из хаоса в космос
Netstat program GUI fix, reduce window redraw. Waiting for a new system colors.
- Attachments
-
-
Screenshot_1.png (4.61 KiB)Viewed 10228 times
-
Из хаоса в космос
Who is online
Users browsing this forum: No registered users and 1 guest