Re: Bookworm Fasttrack and Virtualbox
Greg Wooledge () wrote: > It's not just you. The use of three "b" names in a row (buster, > bullseye, bookworm) was in my opinion a poor decision. I've taken > to calling the releases by their numbers (10, 11, 12) instead of > their codenames to avoid confusion wherever possible. > I feel ashamed. I tried to force the time_t 64 transition in unstable because the system was in a state no longer usable: more than 700 kept back packages and were unable to install anything new so after manually forcing the transition it became unusable, with a lot of broken things. So I set up a stable installation and I wanted Virtualbox for some virtual machines I use. I saw the fasttrack wiki page, I saw the instructions were for bullseye, and I assumed changing bullseye with bookworm will work, but I do not know why bullseye survived in the sources.list even after changing it manually at least two times. I do not know if KDE's klipper was playing with me pasting the incorrect data or what, but bullseye survived... I also didn't notice the error message trying to install virtualbox was referencing bullseye... Bullseye and bookworm are totally different words but maybe for no english native speakers they are easily mistaken I don't know. Now my sources.list is with bookworm, and virtualbox is working just fine, like it was in unstable. Thank you and sorry for the inconveniences.
Re: Bookworm Fasttrack and Virtualbox
Well... it seems my brain can't distinguish Bookworm from Bullseye. Virtualbox is now installed from Fasttrack repository and is working fine. Sorry for the inconveniences
Re: Bookworm Fasttrack and Virtualbox
Xiyue Deng () wrote: > See the version of virtualbox-qt has `fto11' instead of `fto12', while > looking at the repo I can find the build for fto12[1]. What's also > weird is that for all fasttrack supported releases I can only see > version 7.0.12 but no 7.0.6. Anyway, can you paste your sources.list > entries for fasttrack? > This is getting weird, my entries for fasttrack are: deb https://fasttrack.debian.net/debian-fasttrack/ bullseye-fasttrack main contrib non-free deb https://fasttrack.debian.net/debian-fasttrack/ bullseye-backports-staging main contrib non-free Anyway, I added virtualbox's own repository and it worked just fine
Re: Bookworm Fasttrack and Virtualbox
Xiyue Deng wrote: > > If you followed the fasttrack instruction on the website[1] literally, > you may be adding bullseye-fasttrack instead of bookworm-fasttrack where > the latter is what you wanted. > > [1] https://fasttrack.debian.net/ > > -- > Xiyue Deng > Of course I did it for Bookworm. The fact that the website is not yet updated for Bookworm is another thing that makes me think that fasttrack is abandoned. >
Bookworm Fasttrack and Virtualbox
Hello! This evening I tried to install virtualbox into a fresh Bookworm install. I followed the steps in Virtualbox's Debian Wiki entry. After set up fasttrack repository successfully and issue a apt install virtualbox command I get: Reading package lists... Done Building dependency tree... Done Reading state information... Done Some packages could not be installed. This may mean that you have requested an impossible situation or if you are using the unstable distribution that some required packages have not yet been created or been moved out of Incoming. The following information may help to resolve the situation: The following packages have unmet dependencies: virtualbox: Depends: python3 (< 3.10) but 3.11.2-1+b1 is to be installed Depends: python3.9 but it is not installable Depends: libgsoap-2.8.104 but it is not installable Depends: libssl1.1 (>= 1.1.1) but it is not installable Depends: libvpx6 (>= 1.6.0) but it is not installable Recommends: virtualbox-qt (= 7.0 .6-dfsg-1~fto11+1) but it is not going to be installed Recommends: libqt5opengl5 (>= 5.0.2) but it is not going to be installed E: Unable to correct problems, you have held broken packages. I was surprised by the python 3.9 dependency, it leads me to think fasttrack repos are abandoned? What is the right way to install virtualbox in Debian Bookworm? Thanks in advance Miguel A. Vallejo
Re: How can I check (and run) if an *.exe is a DOS or a Windows program?
If I remember correctly, all Windows EXE have an string saying: This program cannot be run in DOS mode. It is displayed when loaded under MS-DOS, so it should be easy to identify an EXE for DOS from an EXE for Windows. To run it, dosbox should work. Greetings
Re: e-mail with line in body beginning with "From"
I remember some e-mail programs automatically add an extra space in front of a From in the message body if any line starts with From. Probably Thunderbird is one of them. to...@tuxteam.de wrote: > > On Sat, Dec 10, 2022 at 08:36:39AM -0500, Greg Wooledge wrote: > > [...] > > > Technically, it's the 5-character sequence "From " (including the space) > > that matters in mbox formats. If you begin a line with "Fromage" [...] > > I thusly propose to drop the '>' escaping of "From" and change every From > at a line start to "Fromage". > > Hmmm. Now I'm hungry. > > (Thanks for the laugh :-) > > Cheers > -- > t
Re: intel-media-va-driver segmentation fault in unstable
Karthik () wrote: > > Same here I took a look at the syslog and found something: Mar 23 13:21:51 kernel: [ 347.474189] vlc[2625]: segfault at 30200 ip 7f75b88659ae sp 7f75 b8d23b00 error 4 in libigdgmm.so.12.1.0[7f75b87fc000+78000] Mar 23 13:21:51 kernel: [ 347.474197] Code: ff 4c 8b 55 c8 31 c0 b9 40 00 00 00 4c 8b 45 b8 4c 89 fe 49 8d 92 b0 33 00 00 48 89 d7 f3 41 0f 6f 00 ba c4 06 00 00 f3 48 ab <49> 8b 06 49 8d 7a 44 41 c7 82 b8 35 00 00 0a 00 00 00 f3 41 0f 6f Mar 23 13:23:11 kernel: [ 427.398576] vainfo[2646]: segfault at 30200 ip 7fc7348059ae sp 7 ffd46cdc950 error 4 in libigdgmm.so.12.1.0[7fc73479c000+78000] Mar 23 13:23:11 kernel: [ 427.398584] Code: ff 4c 8b 55 c8 31 c0 b9 40 00 00 00 4c 8b 45 b8 4c 89 fe 49 8d 92 b0 33 00 00 48 89 d7 f3 41 0f 6f 00 ba c4 06 00 00 f3 48 ab <49> 8b 06 49 8d 7a 44 41 c7 82 b8 35 00 00 0a 00 00 00 f3 41 0f 6f So the culprit is libigdgmm. With this info and a quick search I found this: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1007992 A downgrade of libigdgmm12 (both amd64 and i386) to a previous version fixes it temporarily
intel-media-va-driver segmentation fault in unstable
Hello! Since the lass update last night in Sid / Unstable, I'm getting segmentation faults from some programs, like VLC: vlc video.mkv VLC media player 3.0.17.3 Vetinari (revision 3.0.13-8-g41878ff4f2) [55c0d2e29460] main libvlc: Running vlc with the default interface. Use 'cvlc' to use vlc without interface. [7f758c004380] gl gl: Initialized libplacebo v4.192.1 (API v192) libva info: VA-API version 1.14.0 libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/iHD_drv_video.so libva info: Found init function __vaDriverInit_1_14 Segmentation fault Even utilities like vainfo fails: vainfo libva info: VA-API version 1.14.0 libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/iHD_drv_video.so libva info: Found init function __vaDriverInit_1_14 Segmentation fault Any ideas? Thanks in advance
Re: Wine fails because of missing libgcc_s_dw2-1.dll
Hello! Opened a bug report against Wine, because I was not sure about what package was the culprit. #1006725 It was reasigned to libz-mingw-w64 and a few moments ago I downloaded the updated package from the repository. Everything works fine now. Thanks to everyone who made this possible.
Re: Wine fails because of missing libgcc_s_dw2-1.dll
Thank you! I filled a bug report. Let's see what happens now.
Re: Wine fails because of missing libgcc_s_dw2-1.dll
> > I'm pretty sure the problem is the Wine DLL files were built without > > the -static-libgcc flag > > > > Time to file a bug? > > > > I think it is I'm looking for the exact package that contains zlib1.dll but I can't find it. Libwine depends on libz-mingw-w64 but the zlib1.dll there is not the same as the one wine puts in the drive_c directory. The one in drive_c directory is a PE32 DLL and the one provided with libz-mingw-w64 is a 64 bit one. I confirmed zlib1.dll needs libgcc_s_dw2-1.dll and this one needs libwinpthread-1.dll and... I can see a DLL hell in advance. Where does this zlib1.dll come from? It appears everytime wine creates a drive_c directory, even for a new user.
Re: Wine fails because of missing libgcc_s_dw2-1.dll
> Not entirely sure, but I think this is the issue: > More and more Wine DLL files are built with mingw. > The libwine package already depends on libz-mingw. > Probably the 'gcc-mingw-w64-i686-win32-runtime' package should now be > installed as well. > > Does the program work when you install this package? Installed gcc-mingw-w64-i686-win32-runtime (and gcc-mingw-w64-base because of dependencies) but still not working. These packages install some DLLs in /usr/lib/gcc/i686-w64-mingw32 so win32 programs can't load them there. Also note libgcc_s_dw2-1.dll is not provided with these packages. I'm pretty sure the problem is the Wine DLL files were built without the -static-libgcc flag Time to file a bug?
Re: Wine fails because of missing libgcc_s_dw2-1.dll
Hello! Finding and placing missing DLLs into the Windows directory of Wine is easy. I think the problem is the windows executables in wine have not been compiled with the -static-libgcc option of MinGW, so not including all these libgcc DDLs can be considered a bug itself. But compiling it without -static-libgcc can be considered a bug also. This is why I'm asking.
Wine fails because of missing libgcc_s_dw2-1.dll
Hello After the recent update of Wine packages in Sid, some programs do not run because of a missing libgcc_s_dw2-1.dll. They all worked just fine with the previous version. ~$ wine SpaceEngine.exe 014c:err:module:import_dll Library libgcc_s_dw2-1.dll (which is needed by L"Z:\\home\\eoz\\Programs\\SpaceEngine\\system\\zlib1.dll") not found 014c:err:module:import_dll Loading library zlib1.dll (which is needed by L"Z:\\home\\eoz\\Programs\\SpaceEngine\\system\\libtiff3.dll") failed ( error c07b). 014c:err:module:import_dll Library libtiff3.dll (which is needed by L"Z:\\home\\eoz\\Programs\\SpaceEngine\\system\\SpaceEngine.exe") not found 014c:err:module:LdrInitializeThunk Importing dlls for L"Z:\\home\\eoz\\Programs\\SpaceEngine\\system\\SpaceEngine.exe" failed, status c135 I looked in my backups and I never had any libgcc_s_dw2-1.dll installed in my system, but now it seems to be needed. Any ideas? Thanks in advance
apt automatically upgrading packages in unstable
A few days ago I noticed my debian unstable started to update packages automatically. A quick inspection showed apt was updated, and also the configuration files in /etc/apt/apt.conf.d, including a 20auto-upgrades file with all options enabled. Because auto update in sid is at least dangerous, what is the right way to keep apt configuration between apt updates? Thank you
Re: Telegram Systray Icon
> Oh. You Are running UNSTABLE version of Debian. And just installed a > program outside of Debian, from third party website. You must know what > you are doing. I know what I'm doing, thank you. > Also you must then know how to report bugs like this, I > understand. Otherwise you shouldn't be running UNSTABLE version of Debian. > I know how to report bugs, but sometimes it is difficult to guess from what package the bug comes from. That's the reason I ask in a user list. Anyway I found the problem. It is a bug in KDE: https://bugs.kde.org/show_bug.cgi?id=418996 Good night
Re: Telegram Systray Icon
I thought it was clear, but I was wrong. Telegram from website ( version 3.0.1 ), running in KDE under Unstable. Everything was ok until Bullseye came out and a tsunami of new packages arrived to Unstable. Since that moment the icon disappeared. Everything seems to work fine, except for that icon. No errors, nothing suspicious, neither in logs or running Telegram from command line. Just no icon. Thanks in advance
Telegram Systray Icon
Hello! Since the new Debian release and after a ton of new packages arrived to Unstable I noticed the system tray icon for Telegram desktop does not appear... or at least it appear completely black. See attached image. Is it a Debian problem or a Telegram problem? Any ideas? Thank you
Re: /bin/tempfile missing in debianutils 5.0.1-1
Thanks Greg https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=992385 It is xorg-common still using tempfile. So... if you need X your choices are downgrade debian-utils or wait until a new version of xorg-common is available. Greetings
/bin/tempfile missing in debianutils 5.0.1-1
Hello. After the new release a tsunami of new packages arrived to unstable. They all installed ok but after the upgrade the graphical interface didn't show up. A quick look in .xsession-errors showed tempfile command was missing. tempfile must be in debian-utils package but it isn't https://packages.debian.org/sid/amd64/debianutils/filelist Is this a bug or a change in the policy? Thanks in advance
Re: What is the command to access the temp sensors in a rpi4?
Do not forget for sensors to work you must first run sensors-detect Greetings El lun, 18 ene 2021 a las 20:11, Eduardo M KALINOWSKI () escribió: > > On 18/01/2021 15:35, Gene Heskett wrote: > > On Monday 18 January 2021 12:15:00 Stefan Monnier wrote: > > > >>> ISTR there is a "vcg*" sort of thing to read that stuff, but have > >>> lost it in the fog of time. > >> > >> If the kernel is sane, then I think that the following command should > >> give you this info: > >> > >> sensors > > This kernel isn't "sane": 4.19.71-rt24-v7l+ #1 SMP PREEMPT RT > > > >> it comes from the `lm-sensors` package. > > Returns a null. > >> > >> > > In my RPI4, running stock Debian from https://raspi.debian.net/, sensors > works: > > $ sensors > cpu_thermal-virtual-0 > Adapter: Virtual device > temp1:+43.8°C (crit = +90.0°C) > > > > -- > Linux é user-friendly e não idiota friendly. Se você não entende isso, > use o Windows! > > Eduardo M KALINOWSKI > edua...@kalinowski.com.br >
Re: copy/paste in vim (in terminal)
Richard Hector () wrote I used to be able to use my mouse to select/paste 'normally' (for X), > when using vim in a terminal. More recently (a few years?), it doesn't > seem to work. Yes, since a few years(?) you must use shift and the mouse for copy / paste in vim using Debian's default configuration.
Re: Trying to understand why WiFi gets disconnected in Debian Buster
Hi! I have been able to get the WiFi connected continuously to the AP for 112 hours and counting (almost five days!). Now the WiFi follows the AP frequency changes seamlessly without any drop. What did I do? I removed the crda and wireless-regdb packages. I have been suspecting of them long time ago and after many experiments I discovered wireless-regdb is Debian was (is) terribly outdated, so I filled a bug ( https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=949553 ), and the bug was corrected some days later, at least on testing and unstable, because stable has still the outdated package. I installed the new version in stable, but drops still happened, so I decided to completely remove any wireless restriction/configuration removing wireless-regdb and crda. Now my system has a wireless regdomain 00: DFS-UNSET, which is a sort of lowest common denominator for all wireless regdoms... but now it works just fine. I have seen in syslog dozens and dozens of AP frequency changes and the system followed all of them without any single problem. Now it is cristal clear there is a big issue with WiFi in Debian. I don't know if the kernel does not understand the info contained in the wireless-regdb database, or the info in the wireless-regdb is just wrong. This will be the next step.
Re: Trying to understand why WiFi gets disconnected in Debian Buster
didier gaumet wtote: > (sorry Miguel, again I clicked too quickly on > answer instead of answer to the group and > sent tu you a private e-mail) Don't worry. Thanks for the suggestion. I'll check current versions in buster backports to see if I'll be bitten by the i915 bugs... Meanwhile I've tested a Linksys dual band wireless N usb dongle that uses a ralink module and this problem when AP changes the channel in the 5 GHz band still persists so I must conclude I'm dealing with a kernel bug and not a mt76 module bug. Being affected by several WiFi bugs at the same times sucks!
Re: Trying to understand why WiFi gets disconnected in Debian Buster
Alexander V. Makartsev wrote: > This snippet is suggesting that AP wanted to switch channel\bandwidth to > dual channel, but your wifi adapter couldn't comply with the change, hence > forced disconnect. But the adapter is a 802.11ac one, it should be able to comply with the change > I suggest you to reconfigure your AP in a way it supports single channel > operation only (set it to either 20Mhz or 40Mhz, not both) and also if > possible, disable auto-channel function of your AP and manually select > operating channel\bandwidth, so your wifi adapter won't have to switch > between ever changing channels. The problem is any of these measures have a great impact on wifi performance, and there are other devices in the same network working just fine with these 80 MHz AC mode channels so I prefer to debug the problem and find a solution instead of cripple the performance of all connected devices. Thanks
Trying to understand why WiFi gets disconnected in Debian Buster
Hello everyone. In my syslog I have some entries like this every day: Mar 9 17:35:28 waterhole wpa_supplicant[524]: CTRL-EVENT-CHANNEL-SWITCH freq=5540 ht_enabled=1 ch_offset=1 ch_width=80 MHz cf1=5530 cf2=0 Mar 9 17:35:28 waterhole kernel: [18873.586006] AP a0:64:8f:xx:xx:xx tries to chanswitch to same channel, ignore Mar 9 17:35:28 waterhole kernel: [18873.586011] AP VHT information is invalid, disable VHT Mar 9 17:35:28 waterhole kernel: [18873.586016] AP a0:64:8f:xx:xx:xx changed bandwidth, new config is 5540 MHz, width 2 (5550/0 MHz) Mar 9 17:35:28 waterhole kernel: [18873.586018] AP a0:64:8f:xx:xx:xx changed bandwidth in a way we can't support - disconnect Mar 9 17:35:28 waterhole kernel: [18873.586020] failed to follow AP a0:64:8f:dc:b1:81 bandwidth change, disconnect and every time I get a wifi disconnect, which is very annoying and counterproductive. The only difference between entries is the "AP a0:64:8f:xx:xx:xx tries to chanswitch to same channel, ignore" line. Some times it appears, some times not, but all of them I got disconnected. I'm trying to understand the lines: CTRL-EVENT-CHANNEL-SWITCH freq=5540 ht_enabled=1 ch_offset=1 ch_width=80 MHz cf1=5530 cf2=0 and AP a0:64:8f:xx:xx:xx changed bandwidth, new config is 5540 MHz, width 2 (5550/0 MHz) Because except fo the "width 2", which I'm not sure of the meaning, I can't figure out what parameter is conflicting with my regulatory domain to produce the disconnection. ~# iw reg get global country ES: DFS-ETSI (2400 - 2483 @ 40), (N/A, 20), (N/A) (5150 - 5250 @ 80), (N/A, 23), (N/A), NO-OUTDOOR, AUTO-BW (5250 - 5350 @ 80), (N/A, 20), (0 ms), NO-OUTDOOR, DFS, AUTO-BW (5470 - 5725 @ 160), (N/A, 26), (0 ms), DFS (5725 - 5875 @ 80), (N/A, 13), (N/A) (57000 - 66000 @ 2160), (N/A, 40), (N/A) The system is a Debian Buster Desktop with a Mediatek MT7612u adapter connected to an Askey 802.11ac access point. Can anybody point me in the right direction? What am I missing? Thanks in advance.
Re: Conectivity loss because gateway's arp entry is incomplete
> Try to add your gateway arp with static arp entry and check whether the > problem persists. It does not work because the desktop machine still is not able to answer gateway's arp requests. Well actually it answer the request but they do not get outside the machine.
Re: Recommendations wanted: usb to WiFi dongle
My advice: Keep away from mt76 devices with Buster. I'm using Mediatek MT7612U and it does not work well
Re: Conectivity loss because gateway's arp entry is incomplete
> This is not necessarily true. I recently bought a (chinese) wifi > adapter (atheros chip) that was recognized as mediatek. The moment > I linked the correct drivers with the device it worked properly. Not my case. They all have the same usbID (0e8d:7612) and I physically opened them and I saw the MT7612U chip in the four devices.
Re: Conectivity loss because gateway's arp entry is incomplete
Nektarios Katakis wrote: > I would check the `dmesg` command output. If there are erros there then > they re coming from the kernel. I forgot to mention not a single message in syslog, kern.log, or any other log I could find. Nothing. Nada. Zero. > Have you made sure that the wifi is not the problem? Before going into > wild software investigation? No, that's what I want to discover to fill an appropriate bug report against the right component. I must admit MT7612u support is buggy as hell. I suffer many other bugs with these devices, all of them related to the inhability to follow AP frequency changes in the 5 GHz band, but this one does not seems related to the wifi itself. When the bug triggers I can send and receive packets. Only sending arp messages seems to be missing. > For example have you tried in a Windows > laptop with the correct device drivers? Because you even might not be > using the correct device drivers! I can't test the devices in Windows because I have no Windows here, but I have four dongles, from two different manufacturers, all of them with the MT7612U chip in it and all of them load the mt76x2u module when plugged, so I must conclude they are using the correct device driver/module. The problem exists with all the four dongles.
Conectivity loss because gateway's arp entry is incomplete
Hello! Diagnosing one of my several problems with WiFi in Debian I found a problem this time not related to the WiFi connection itself, but I don't know how to get the culprit... Anyone can point me in the right direction? Symptoms: From time to time I lost Internet connection for a period of about 5 to 30 minutes. It happens at least two or three times a day. Setup: Desktop computer (192.168.1.2) running Debian Buster amd64 with current kernel 4.19.98-1 connected via WiFi to router / gateway (192.168.1.1). Wifi Adapter is MT7612U from Mediatek. While connection is lost: - I can launch wavemon and I can confirm the WiFi adapter is still connected and sending / receiving packets and beacons. - Arp command shows the gateway HWaddress as incomplete. - Launching Wireshark in the desktop computer I can see incoming packets from other devices, but I can also see a "Who has 192.168.1.1? Tell 192.168.1.2" ARP requests from my desktop every second. At first glance it seems the gateway isn't responding to my arp requests, but... - I can see arp queries from the gateway to my desktop, and I can see the responses, but they seem they never reach the gateway. - Forcing network activity with my smartphone, I can see also arp requests from the smartphone answered by the desktop, but a ping from the smartphone to the desktop machine gives no response. Of course the phone can use internet without any problem using the same gateway. - If I unplug and replug the WiFi adapter everything works again. If I left the WiFi adapter plugged, eventually I get an arp response from the gateway, and everything works normally again. This may take up to 20-30 minutes. Only once the wifi connection crashed after several minutes. It seems that when I lost connectivity I can't sent arp messages but I can receive them. They are sent, I can see them with Wireshark, but it seems they will never get outside the machine. I searched for known bugs but I wasn't able to find anything, only a message from an Ubuntu user aparently having the same problem, but with a different WiFi adapter and kernel version (4.18 I think): https://askubuntu.com/questions/1122683/no-internet-access-after-random-intervals-in-ubuntu-18-04-2-lts The problem is now I don't know how to get the culprit of this bug. Is the kernel? Is the Mediatek kernel module? Any help will be appreciated.
Re: Understanding the two-year release cycle as a desktop user (and a Debian newcomer)
I have a friend who works as a system administrator that always says: Debian Stable --> Debian Server Debian Testing --> Debian Desktop Moderate Debian Unstable --> Debian Desktop Fast Being Moderate and Fast the speed at you will eat new bugs. As time goes by, I'm more and more agree with him. Debian Stable should be enough for everyone, but... in practice only works for servers and very low demanding desktops / workstations. if you use fast moving software (for example, ham radio applications), or developing new software, or helps developing new software, your only option is to run ahead to testing / unstable. Packages versions in Stable are obsolete even on the very same launch day. The developer puts a new version, you download the sources, try to compile them, compilation failed, asks the developer why and the answer is always the same: Your put_any_library_here version is obsolete, please upgrade. So you are forced to move to testing / unstable. Another case is if you suffer a serious bug with stable, usually hardware support. Then, also your only option is to run ahead to testing / unstable and cross your fingers. In my case, suffering the big problems with Intel video cards the only option is to move to unstable and pray for every new kernel to be the one that solves the problem. I've read there will be no chace to get a working kernel with Intel cards until next 5.7 so now I'm on stable, which seems to be somewhat stable (only one chrash in the last 7 days). I activelly follow the development of some programs in diverse areas (ham radio, astronomy, emulators, etc), and new versions usually appears in unstable months after release. The more prominent example is the kernel itself. Current version is 5.5.7, but unstable still have 5.4.19. When the statibily of your system depends on a feature or bug solved in a new kernel version, this can be distressing. Even with all this, I still prefer to use Debian.
Re: Poll about DVD drive behavior
DVD A DH16AASH, the tray has been open now for more than an hour. Hope this helps.
Re: The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
Jörg-Volker Peetz wrote: > Do you also use the skylake GPU firmware which should be located in > /lib/firmware/i915/ > (from package firmware-linux-nonfree, I think; file names beginning with > "skl_")? Of course. I even installed the missing files in current testing / unstable https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=931930 But no difference. Once you start to use 3D accelerated graphics, the bomb starts ticking to explode when you least expect it.
Re: Re: The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
> > Have you monitored your CPU temperatures? Yes. Always under 40°C... 45°C max. Temperature is not the problem.
Re: The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
Geoff Reidy wrote: > I've never had an Intel GPU but always had the impression they > were pretty solid, but my opinion is changing. As a user of Intel GPUs for the last 5 years I can tell you Intel GPUs and Linux are just a nightmare. A truly pain in the ass. > Note that kernel 5.5 may still have issues Yeah, I read it. I have been playing with the xserver-xorg-video-intel in Buster with kernel 4.19.0-8 and trying different options in xorg.conf I didn't get any hang since 10 hours ago, almost a new record for me. Also text console corruption has disappeared, so I will use this configuration to try to get pending work done, and then I will see. I still don't rule out buying a new graphics card. Thank you for all your comments and suggestions. They are really helping me.
Re: Hardware for Debian
I have a MSI B150M PRO-VDH motherboard. Never had any problem booting Linux on it. It can even disable UEFI stuff. Good luck.
Re: The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
Felix Miata wrote: > If you want a solution with absence of reverse-engineering or kernel > tainting, I suggest keeping after the devs and making sure to answer > any requests you might get there promptly in > https://gitlab.freedesktop.org/drm/intel/issues/673, and offer additional > testing and logs on the intel-...@lists.freedesktop.org mailing list > referencing that bug report. This has obviously been going on much > too long for (against) you. :-( The problem is the pattern always seems to be the same: 1.- Your system freezes because a GPU hang. 2.- Upload bug report to bug trackers 3.- After some time, the problem is solved in the next kernel version 4.- Wait for the new kernel version to arrive to unstable 5.- Test the new kernel and... your system freezes again, now with a totally different error / problem, but with the same effect: unreliable computer to get your job done 6.- Go to 2 This has been my workflow for the last five years. I'm getting tired. > Until this thread and that bug report, I was completely unaware of the > extent of recent Intel GPU complaints that that bug represents. https://linuxreviews.org/Linux_Kernel_5.5_Will_Not_Fix_The_Frequent_Intel_GPU_Hangs_In_Recent_Kernels Looking for solutions I found this article and it pretty much resumes the last years of Intel GPU experiences under any Linux distribution. elvis wrote: > Have you tried a live CD of a different distribution? Fedora, > Mint, Suse etc? Yes. I tried Mint, (K)Ubuntu, and almost any deb based distribution. The same problems with all of them. mlnl wrote: > I have the GPU hang bug too, but rare. After reading a lot > about it and saw a lot different ways to "configure" the system > for the Intel GPU, i have choosen the follwing steps (with > Debian Buster, Xfce, 5.4 backports kernel before and > 5.5.3 vanilla kernel now) Thank you for the detailed report. I didn't tried a lot the Xorg Inter "driver" because it seems to do not work at all with my HD530 GPU: Screen flashes, big rectangles turning into black, scroll corruption totally unusable. But I must admit I didn't play too much with the options for that driver. Maybe I should try it again. Andrei POPESCU wrote: > buster-backports has 5.4.8. I know. Kernel 4.19.x has console corruption but low hang rate, but 5.4.x has no console corruption but really high hang rate, at least on my machine. Thank you for all your suggestions.
Re: The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
> My advice: put an nvidia card in there. That's what I did, and have had no > problems since. Thank you for the advice. That's my plan if kernel 5.5 doesn't work well.
Re: The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
> I think it is important to find the appropriate kernel version for your > system - the one that has all the bits and bolts for your hardware. I doubt > I will move to newer kernel. This one seems to have all the fixes at least > for the hardware I am using now and especially the gpu part. I also had > freezes few years ago and followed the kernel releases until 4.9.25. That makes me wonder: What is the right way to do that in Debian? In stable releases there is no way to use another kernel except from backports (and they are newer, not older). And in testing / unstable the kernels disapears as new kernels get into. How to install and keep an old kernel in a new instalation? Thanks
Re: The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
> Are you sure this is not coming from a software you are using? For every problem I had / have with Intel GPU I can find other users on any distribution's bug tracker with exactly the very same errors in the kernel log, so I'm pretty sure I'm not the only one suffering from Intel GPUs. In fact I am surprised by the small number of people here with that problem. But it's true there are some programs that are very prone to hang the GPU when used, like Chrome, Chromium, Firefox, Google Earth, KDE's taskbar, Virtualbox... all of them use some kind of 3D acceleration
Re: The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
Interesting very interesting. > (On what GPU?) The same as yours: VGA compatible controller: Intel Corporation HD Graphics 530 (rev 06) in a I5-6500 CPU. > For whatever it's worth, I do *not* see any problems like this. Not > even close. Intel integrated graphics of this generation have been > rock solid for me on Debian 10, with non-free firmware + microcode. You are lucky. Very lucky. For the last 5 years this have been a constant pain, one problem after problem. When a problem got fixed by a new kernel a new problem appeared, for every single kernel version I tested. It even affected my ability to get the work done: yesterday I had 3 complete system freezes in less than an hour. But I use KDE, I use 3D acceleration, and I use programs who uses 3D acceleration so I constantly hit these kind of bugs. One of the less buggy kernels for me in the one in Debian Buster (4.19.0-8-amd64). Hangs are not so frequent with this kernel but the text consoles are corrupted: https://bugs.freedesktop.org/show_bug.cgi?id=107951 This capture is from my own computer: https://bugs.freedesktop.org/attachment.cgi?id=143903 The problem was solved in later kernels but the fix never reached Debian Buster. > The web page you linked also talks about "transition to idle", which > I'm guessing is related to CPU power management...? I'm using a desktop, so power saving is not a problem. Of course one of the multiple tests I made was disable GPU power saving, limit the C states...all the proposed workarounds that works for some users but none of them worked for me. I will try kernel 5.5 when available in unstable and if I still get hangs, I will try a different graphics card. Thank you.
Re: The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
> Hi, Miguel > The computer I'm typing on is a Lenovo Z570. I7-2670 processor, 8 gigs > ram, 480 gig ssd. Snappy as hell. Running Debian Bullseye. I've made > it usable by: > Adding a 20-intel.config file to /etc/X11/xorg.conf.d > Adding the Ubuntu graphics PPA > Modifying the /etc/default/grub file > Modifying the grub command line to disable modesetting. This one you > have to get just right or > you'll end up with a black screen. This is exactly the kind of things I was afraid of... > If you are OK with much more fan running and slightly poorer 2D and 3D > graphics, with substantially poorer battery life, just do the > 20-intel.config with, instead of "intel" as the driver, make the > driver line "modesetting". I always used modesetting, because the xserver-xorg-video-intel doesn't work at all for me: Flashing screens, big rectangles blinking... totally unusable. Anyway I will look for the Ubuntu graphics PPA and how it differs from the Debian packages. Thank you!
Re: The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
> I have had perfectly reasonable experiences with: > > nvidia 8400 > nvidia 720 > nvidia 730 > nvidia 1030 > nvidia 1050 I made a quick search and some of these cards are under 50 euros, so I will try if I do not get a solution. Thank you for the advice!
Re: The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
> I was under impression that Intel GPUs are supported quite well. > > How exactly did you reach the conclusion that the freezes are related to > the GPU? Because everytime my system hangs / freezes I found something like this in syslog: [ 135.116721] i915 :00:02.0: GPU HANG: ecode 9:1:0x, hang on rcs0 [ 135.116724] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 135.116725] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [ 135.116726] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 135.116727] The GPU crash dump is required to analyze GPU hangs, so please always attach it. [ 135.116729] GPU crash dump saved to /sys/class/drm/card0/error [ 135.117739] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 135.118508] [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 0001, RESET_CTL: 0001} [ 135.118620] i915 :00:02.0: Resetting chip for hang on rcs0 [ 135.120388] [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 0001, RESET_CTL: 0001} [ 135.121147] [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 0001, RESET_CTL: 0001} [ 159.118357] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 159.119128] [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 0001, RESET_CTL: 0001} [ 159.119221] i915 :00:02.0: Resetting chip for hang on rcs0 [ 159.120990] [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 0001, RESET_CTL: 0001} [ 159.121748] [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 0001, RESET_CTL: 0001} [ 163.341488] Asynchronous wait on fence i915:kwin_x11[909]:1326 timed out (hint:intel_atomic_commit_ready+0x0/0x50 [i915]) [ 167.118543] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 167.119314] [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 0001, RESET_CTL: 0001} [ 167.119412] i915 :00:02.0: Resetting chip for hang on rcs0 [ 167.121181] [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 0001, RESET_CTL: 0001} [ 167.121938] [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 0001, RESET_CTL: 0001} [ 175.118708] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 183.118916] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 185.102909] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 187.118943] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 189.102987] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 191.119032] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 193.103069] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 195.123083] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 197.103126] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 199.123155] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 201.103201] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 203.119230] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 205.103272] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 207.119305] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 209.103334] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 211.119366] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 213.103400] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 215.119433] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 217.103471] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 219.119496] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 221.103531] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 223.119560] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 225.103594] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 227.119629] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 229.103659] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 231.123662] i915 :00:02.0: Resetting rcs0 for hang on rcs0 [ 233.103737] i915 :00:02.0: Resetting rcs0 for hang on rcs0 ... etc, etc, etc, until I got tired and shutdown the computer using the power button. More people with my very same experience here: https://gitlab.freedesktop.org/drm/intel/issues/673 > Other things to try: > - update the microcode (package intel-microcode in non-free) > - try with different, known good RAM > - try with a different, known good and generously sized power source > - update the motherboard BIOS I always ran the intel microcode package, tried every bios update, and I even changed the whole computer, so this is not the solution. Thank you
The nightmare of Intel Integrated GPUs under Linux in general and Debian in particular
Around the end of 2014 I moved to Debian as my primary operating system. In that year my computer was a Dual Core CPU with integrated GPU. It worked just fine with the previous operating system but as soon I started to use Debian every day I got system freezes that I quickly diagnose as GPU bugs. Those days Wheezy was the current stable release. Few months later Jessie was out, and the system freezes still where there. And they were related to the Intel GPU. Someone told me my CPU was too old and nobody will look for a solution in such old hardware, so in 2018 I upgraded my computer to a i5 with HD530 Intel GPU. Things worked fine for one or two weeks, until a new Intel GPU system freeze happend. I tried testing and unstable, but the freezes where still happening. Stretch went out and the freezes continued. Buster was out, and the problems with Intel GPUs are still causing complete system freezes. My last complete system freeze was yesterday: three complete system freezes in less than an hour. My first bug report about a system freeze was in 2015 (see Debian bug #787267) and my last report was only a few days ago. I went from Wheezy to Buster, from kernel 3.16 up to kernel 5.4, and with all of them I had serious GPU issues (system hangs/freezes/display corruptions). Almost five years suffering this day after day is just too much for me. I need a solution. Kernels 5.5 just seems as buggy as all the previous ones at least related to the Intel GPUs and according to the new bug reports at Freedesktop's Gitlab. What are my alternatives? nVidia cards? I've never used an nVidia card but I have read also tons of problems with them in the past. How about now? And how about AMD cards? What are your recommendations / experiences? Thanks in advance.
Re: Buster 10.2 kernel problems
> Downgrades are not supported, did you re-install? Yes, full reinstall over a wiped hard disk. > Try the kernel from backports, currently 5.4. That was exactly the problem I had. Current kernels 5.4.x both on Testing and Sid produce constant GPU hangs on my Intel HD Graphics 530, The ones that outputs "i915: Resetting rcs0 for hang on rcs0" (https://bugzilla.kernel.org/show_bug.cgi?id=205545). It is supposed to be solved on 5.4.x but it is still present. So... no current Debian kernel is bug free and works needs to be done. This is the real problem.
Buster 10.2 kernel problems
Today I downgraded my main computer to Buster 10.2 (current Stable) and found some problems. First: Syslog is constantly filled with messages like these: xhci_hcd :00:14.0: WARN Wrong bounce buffer write length: 630 != 318 xhci_hcd :00:14.0: WARN Wrong bounce buffer write length: 234 != 0 xhci_hcd :00:14.0: WARN Wrong bounce buffer write length: 284 != 630 xhci_hcd :00:14.0: WARN Wrong bounce buffer write length: 208 != 284 xhci_hcd :00:14.0: WARN Wrong bounce buffer write length: 386 != 234 Second: I suffer text console screen corruption when I press Ctrl-Alt-Fn, screenshot here: https://bugs.freedesktop.org/attachment.cgi?id=143903 It makes text console use very unpleasant. Graphics are Intel Corporation HD Graphics 530 (rev 06). I know these bugs have been solved in later kernels (I was in sid until yesterday) but since buster is tied to 4.19.0-6 I don't know how to preceed. Any ideas? Thanks in advance.
TP-LINK Archer T1U USB WiFi device only works on pure N networks
Hi Playing with different 802.11ac usb devices I noticed the small 5 GHz only TP-Link Archer T1U (usbid 2357:0105) only works if you try to connect to a pure 5 GHz 802.11n network. In this case no problem at all. But if you try to connect to a 802.11ac/n network, it does not connect, falling in a endless connecting loop. Tested with kernel 5.3.0-2-amd64 (5.3.9-3) and firmware-misc-nonfree 20190717-2 over the same WiFi router, only changing the 5 GHz band mode from N to AC/N. Sorry if this is not the correct mailing list but I think there is not a dedicated wireless Debian mailing list. Please contact me if someone need more data. Greetings Syslog: Nov 22 22:41:58 waterhole kernel: [ 5604.108872] usb 1-6: new high-speed USB device number 9 using xhci_hcd Nov 22 22:41:58 waterhole kernel: [ 5604.272184] usb 1-6: New USB device found, idVendor=2357, idProduct=0105, bcdDevice= 1.00 Nov 22 22:41:58 waterhole kernel: [ 5604.272189] usb 1-6: New USB device strings: Mfr=1, Product=2, SerialNumber=3 Nov 22 22:41:58 waterhole kernel: [ 5604.272193] usb 1-6: Product: WiFi Nov 22 22:41:58 waterhole kernel: [ 5604.272195] usb 1-6: Manufacturer: MediaTek Nov 22 22:41:58 waterhole kernel: [ 5604.272198] usb 1-6: SerialNumber: 1.0 Nov 22 22:41:59 waterhole kernel: [ 5604.405191] usb 1-6: reset high-speed USB device number 9 using xhci_hcd Nov 22 22:41:59 waterhole kernel: [ 5604.558249] usb 1-6: ASIC revision: 7612 MAC revision: 76502000 Nov 22 22:41:59 waterhole kernel: [ 5604.559434] usb 1-6: firmware: direct-loading firmware mediatek/mt7610e.bin Nov 22 22:41:59 waterhole systemd[1]: NetworkManager-dispatcher.service: Succeeded. Nov 22 22:42:00 waterhole kernel: [ 5605.559757] usb 1-6: EEPROM ver:02 fae:01 Nov 22 22:42:00 waterhole kernel: [ 5605.801741] ieee80211 phy11: Selected rate control algorithm 'minstrel_ht' Nov 22 22:42:00 waterhole NetworkManager[654]: [1574458920.4396] device (wlan0): driver supports Access Point (AP) mode Nov 22 22:42:00 waterhole NetworkManager[654]: [1574458920.4407] manager: (wlan0): new 802.11 Wi-Fi device (/org/freedesktop/NetworkManager/Devices/13) Nov 22 22:42:00 waterhole systemd[1]: Starting Load/Save RF Kill Switch Status... Nov 22 22:42:00 waterhole mtp-probe: checking bus 1, device 9: "/sys/devices/pci:00/:00:14.0/usb1/1-6" Nov 22 22:42:00 waterhole mtp-probe: bus: 1, device: 9 was not an MTP device Nov 22 22:42:00 waterhole systemd-udevd[10423]: Using default interface naming scheme 'v243'. Nov 22 22:42:00 waterhole systemd-udevd[10423]: ethtool: autonegotiation is unset or enabled, the speed and duplex are not writable. Nov 22 22:42:00 waterhole kernel: [ 5605.860583] usb 1-6 wlxc025e96f9d55: renamed from wlan0 Nov 22 22:42:00 waterhole systemd[1]: Started Load/Save RF Kill Switch Status. Nov 22 22:42:00 waterhole NetworkManager[654]: [1574458920.5049] rfkill11: found Wi-Fi radio killswitch (at /sys/devices/pci:00/:00:14.0/usb1/1-6/ieee80211/phy11/rfkill11) (driver usb) Nov 22 22:42:00 waterhole NetworkManager[654]: [1574458920.5176] device (wlan0): interface index 13 renamed iface from 'wlan0' to 'wlxc025e96f9d55' Nov 22 22:42:00 waterhole NetworkManager[654]: [1574458920.5274] device (wlxc025e96f9d55): state change: unmanaged -> unavailable (reason 'managed', sys-iface-state: 'external') Nov 22 22:42:00 waterhole mtp-probe: checking bus 1, device 9: "/sys/devices/pci:00/:00:14.0/usb1/1-6" Nov 22 22:42:00 waterhole mtp-probe: bus: 1, device: 9 was not an MTP device Nov 22 22:42:00 waterhole systemd-udevd[10405]: Using default interface naming scheme 'v243'. Nov 22 22:42:01 waterhole dbus-daemon[653]: [system] Activating via systemd: service name='org.freedesktop.nm_dispatcher' unit='dbus-org.freedesktop.nm-dispatcher.service' requested by ':1.8' (uid=0 pid=654 comm="/usr/sbin/NetworkManager --no-daemon ") Nov 22 22:42:01 waterhole systemd[1]: Starting Network Manager Script Dispatcher Service... Nov 22 22:42:01 waterhole dbus-daemon[653]: [system] Successfully activated service 'org.freedesktop.nm_dispatcher' Nov 22 22:42:01 waterhole systemd[1]: Started Network Manager Script Dispatcher Service. Nov 22 22:42:01 waterhole wpa_supplicant[664]: dbus: fill_dict_with_properties dbus_interface=fi.w1.wpa_supplicant1.Interface.P2PDevice dbus_property=P2PDeviceConfig getter failed Nov 22 22:42:01 waterhole NetworkManager[654]: [1574458921.4166] sup-iface[0x56310c3b5a20,wlxc025e96f9d55]: supports 4 scan SSIDs Nov 22 22:42:01 waterhole NetworkManager[654]: [1574458921.4171] device (wlxc025e96f9d55): supplicant interface state: starting -> ready Nov 22 22:42:01 waterhole NetworkManager[654]: [1574458921.4172] device (wlxc025e96f9d55): state change: unavailable -> disconnected (reason 'supplicant-available', sys-iface-state: 'managed') Nov 22 22:42:03 waterhole NetworkManager[654]: [1574458923.5728] policy: auto-activating connection 'Wireless2' (b7f9451b-48dc-4011-8305-649c4fd808b8) Nov 22 22:42:03 waterhole NetworkManager[654]:
Re: USB WIFI Adapters?
If WiFi N is fine for you I recommend the Cisco WUSB600N V2. It's a dual band USB stick based on the RaLink RT3572 chip and works like a charm with Debian. I used some of them in the past for several years with no issues. If you need 802.11ac I guess the support is not very good yet with Debian. I "upgraded" my WiFi sticks to new AC ones based on the Mediatek MT7612 and I have frequent disconnects. They almost work, but still not at 100% Hope this helps. Miguel A. Vallejo
WiFi: changed bandwidth in a way we can't support - disconnect
Hi I'm experiencing some weird WiFi problems since the arrival of Kernels 5.x. Currently I run 5.3.0-2-amd64 and MT7612 MediaTek Inc. 802.11ac WLAN USB3 dongle. I suffer frequent disconects, and everytime this happens, I see this in the system log: wpa_supplicant[651]: wlx00e0bb330276: CTRL-EVENT-STARTED-CHANNEL-SWITCH freq=5540 ht_enabled=1 ch_offset=1 ch_width=80 MHz cf1=5530 cf2=0 wpa_supplicant[651]: wlx00e0bb330276: CTRL-EVENT-CHANNEL-SWITCH freq=5540 ht_enabled=1 ch_offset=1 ch_width=80 MHz cf1=5530 cf2=0 wlx00e0bb330276: AP a0:64:8f:dc:b1:81 tries to chanswitch to same channel, ignore wlx00e0bb330276: AP VHT information is invalid, disable VHT wlx00e0bb330276: AP a0:64:8f:dc:b1:81 changed bandwidth, new config is 5540 MHz, width 2 (5550/0 MHz) wlx00e0bb330276: AP a0:64:8f:dc:b1:81 changed bandwidth in a way we can't support - disconnect wlx00e0bb330276: failed to follow AP a0:64:8f:dc:b1:81 bandwidth change, disconnect wlx00e0bb330276: CTRL-EVENT-DISCONNECTED bssid=a0:64:8f:dc:b1:81 reason=3 locally_generated=1 NetworkManager[642]: [1574118371.9594] sup-iface[0x560817a7a130,wlx00e0bb330276]: connection disconnected (reason -3) Different 5 GHz channels but always the same sequence. After disconnection Network Manager keep asking for the WiFi password (Network Manager already knows this password) but this is a different problem... Other devices connected to the same WiFi router using 802.11ac mode does not have this problem, only my Debian Sid box. Any idea on how to solve / fix / track down this problem? Unfortunately I can't use a wired connection. Thanks in advance. Migue A. Vallejo
Re: Sid broke my wireless
Hi! Kernels 5.2.x are a disaster for wifi. Several wifi modules doesn't work well or doesn't work at all. I'm experiencing myself too much frequent disconnects with USB wifi cards (Mediatek and Realtek). Then add USB 3.0 problems (Wrong bounce buffer write length warnings) and the result is a barely usable system. Downgrade to 4.19 kernels is not possible for me because with them I have 100% cpu usage from ksoftirqd kernel process. I'm really waiting for a new kernel upgrade in Sid.
Re: kernel 5.2.x and firmware-iwlwifi: incompatibilities?
Actually runing linux-image-4.19.0-5-amd64 (4.19.37-6) without any wifi issues (this is a desktop), but with linux-image-5.2.0-2-amd64 wifi is unstable with too frequent drops. WiFi card is a MediaTek MT-7612 USB dongle. Hope this helps basti () wrote: > > Migual has you update to 4.19.37-5+deb10u1 ? > > With previews version of 4.19 I had problems after suspend to disk. > Now it seems to be fixed. > > There are several Bug Reports with "iwlwifi 4.19 suspend" when you > search in on google. > > Best Regards
Re: kernel 5.2.x and firmware-iwlwifi: incompatibilities?
Same problem here. Kernel 5.2.x makes my MT76x2 wifi very unstable, with bad performance and lot of dropouts. Running 4.19.x for now. > I am using Sid on my laptop (HP ZBook 15) and since first linux-image-5.2.0-1 > and then linux-image-5.2.0-2 > the wifi is not stable and goes down after few seconds.
Re: Buster: do_IRQ: 1.35 No irq handler for vector
I added this line as very first line in /etc/rsyslog.conf :msg, contains, "No irq handler for vector" ~ Basically it means: if message contains "No irq handler for vector" then discard. Once you add the line, restart rsyslog: systemctl restart rsyslog It's only a temporary solution until the progem gets fixed, but at least it allows you to continue doing your work. Hope this helps. Bruce Halco () wrote: > > Can you share what changes you made to rsyslog.conf? That's not > something I've ever messed with. > > Thanks. > > Bruce
Re: Buster: do_IRQ: 1.35 No irq handler for vector
Same problem here. For me the messages appear *only* when I start the Arduino IDE and stops as soon as I close it. I suspect is something related to the pool of serial ports the Arduino IDE do to detect the plug of a new Arduino board. Adding pci=nomsi,noaer didn't work for me so I ended tweaking /etc/rsyslog.conf to get rid of them and continue to work as usual. El lun., 18 mar. 2019 a las 15:47, Bruce () escribió: > > I upgraded from Squeeze to Buster last week and have been having a > problem with "No irq handler for vector" messages appearing in all the > console windows. > > There were originally 3 different numeric values in the messages. After > adding 'pci=nomsi,noaer' to the grub boot options, 2 of them stopped. > > I still get this more than 100 times per day.: > > Message from syslogd@penguin at Mar 18 10:11:08 ... > kernel:[161132.371343] do_IRQ: 1.35 No irq handler for vector > > The time between messages can be as little as 9 seconds, or more than 2 > hours. The messages happen throughout the day, whether I'm at the > computer or not. > > I've tried 'dmesg -n 2', 'dmesg -n 1', and even 'dmesg -D'. The messages > keep coming. I especially don't understand how 'dmesg -D' doesn't help, > as it's documented purpose is to 'Disable the printing of messages to > the console'. > > The messages don't correlate to any other events I can find in the log > files. > > On a long shot, I disconnected the optical drives, unplugged the UPS > from the USB port, and removed the TV capture card. None of that made a > difference. > > Other than the messages, the system is running fine. > > The system includes a Phemon II CPU and 16 GB RAM in a Gigabyte > 890GPA-UD3H motherboard. > > Any suggestions? > > Bruce >
Re: PKCS#7 signature not signed with a trusted key
Yesterday the kernel was upgraded to 4.19.0-4 (4.19.28-2) and the problem disappeared.
Mediatek 802.11ac USB devices in kernel 4.19
Hello. Some time ago I read the kernel 4.19 include support for some Mediatek 802.11ac USB devices. When I saw the kernel 4.19 arrived into unstable/testing I ordered one of these devices. The device arrived but it does not work at all. A quick look at the kernel configuration shows those modules are not compiled in the Debian kernel: grep -i mt76x /boot/config-4.19.0-1-amd64 # CONFIG_MT76x0U is not set # CONFIG_MT76x2E is not set # CONFIG_MT76x2U is not set I noticed there is an open bug about this: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=918331 Meanwhile, is there a quick way to compile only those modules and install them in a working kernel? Thank you.
Mediatek 802.11ac USB devices in kernel 4.19
Hello. Some time ago I read the kernel 4.19 include support for some Mediatek 802.11ac USB devices. When I saw the kernel 4.19 arrived into unstable/testing I ordered one of these devices. The device arrived but it does not work at all. A quick look at the kernel configuration shows those modules are not compiled in the Debian kernel: grep -i mt76x /boot/config-4.19.0-1-amd64 # CONFIG_MT76x0U is not set # CONFIG_MT76x2E is not set # CONFIG_MT76x2U is not set I noticed there is an open bug about this: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=918331 Meanwhile, is there a quick way to compile only those modules and install them in a working kernel? Thank you.
Re: Disabling sound fade-in in Debian Buster
> I am not sure it can be of use, but maybe have a look at this bug report > solution: > > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=904822#17 > > Regards, Yes! That was the problem. Disabling power saving in Intel HDA codec makes sound works normal, as it should be. Thank you!
Disabling sound fade-in in Debian Buster
Since I moved my desktop to Buster I noticed each time a sound is played there is a fade-in effect at the begining. The problem is short duration sounds, like notifications, does not sound at all, or you hear only the very end at low volume. A second play of the same sound just after the first sounds fine. This happens with any program playing sounds: DragonPlayer, VLC, Audacity, System notifications... It seems there is a tmeout somewhere and everytime the sound stars a fade-in is produced. I revised pulseaudio configuration files, KDE preferences and googled a lot but I found nothing. Any idea how to disable this automatic fade-in? Thanks in advance.
Re: Most reliable dual band driver/chipset
I'm using Ralink RT3572 (Linksys WUSB600N v2) with no problems after disabling mac randomization in Network Manager. The only drawback is it is only a 802.11n device. I'm looking for a 802.11ac upgrade with no success.
Re: Audacity 2.2.2 + Pulseaudio in Debian Sid
After a lot of googling, trial and error I solved the issue downgrading libasound2 and related packages to version 1.1.6: libasound2_1.1.6-1_amd64.deb libasound2_1.1.6-1_i386.deb libasound2-data_1.1.6-1_all.deb libasound2-dev_1.1.6-1_amd64.deb libasound2-plugins_1.1.6-1+b1_amd64.deb libasound2-plugins_1.1.6-1+b1_i386.deb Something is really bad with current version 1.1.7 of libasound2. Hope this helps someone. El lun., 5 nov. 2018 a las 13:55, Miguel A. Vallejo () escribió: > > Hello. > > I upgraded a machine to sid. Everything seems to work just fine but I > noticed a problem with Audacity 2.2.2 and Pulseaudio 12.2-2. > > I can't select pulse neither as recording or playback device. > > I can select "default" on playing device, and it appears in > pavucontrol and I can manage it. > > But audacity doesn't appear in the recording section of pavucontrol. > In audacity there is no pulse option in record devices, only hardware > devices. > > Any idea? I need to record from pulseaudio monitor devices. > > Thanks in advance.
Audacity 2.2.2 + Pulseaudio in Debian Sid
Hello. I upgraded a machine to sid. Everything seems to work just fine but I noticed a problem with Audacity 2.2.2 and Pulseaudio 12.2-2. I can't select pulse neither as recording or playback device. I can select "default" on playing device, and it appears in pavucontrol and I can manage it. But audacity doesn't appear in the recording section of pavucontrol. In audacity there is no pulse option in record devices, only hardware devices. Any idea? I need to record from pulseaudio monitor devices. Thanks in advance.
Re: USB2 or 3 WiFi dual band adapters
My experience is similar. Currently I'm using Linksys WUSB600N v2 USB dongles based on Ralink RT3572 and they work just fine. They can handle my 100 MBit conection using 2x2 N mode in 5 GHz band without too much problems. But I have also a Linksys WUSB6300 802.11ac dongle, based on RTL8812AU. Although I made it to work, all drivers/modules I tried seems to be in very early stage to my taste, with many missing functions: a lot of missing information while running monitoring programs (like wavemon). So, I'm also looking for a Linux friendly 802.11ac USB dongle. Greetings
Predictable Network Interface Names prevents WiFi connections.
Hello. I'm having a weird problem with my Debian testing installation. I use KDE Plasma 5 and NetworkManager in a up-to-date system. The problem is I can't connect to any wifi if Predictable Network Interface Names are in use. I can see all wireless networks in Network Manager, so I select mine, I enter the network password and I can see this in syslog Dec 6 14:47:31 waterhole NetworkManager[562]: [1481032051.7586] keyfile: add connection in-memory (7c83331f-74b7-4ef8-98a5-67bc03e24b36,"EOZWireless2") Dec 6 14:47:31 waterhole NetworkManager[562]: [1481032051.7954] device (wlx1c7ee55ecc0a): Activation: starting connection 'EOZWireless2' (7c83331f-74b7-4ef8-98a5-67bc03e24b36) Dec 6 14:47:31 waterhole NetworkManager[562]: [1481032051.8158] keyfile: update /etc/NetworkManager/system-connections/EOZWireless2 (7c83331f-74b7-4ef8-98a5-67bc03e24b36,"EOZWireless2") and persist connection Dec 6 14:47:31 waterhole NetworkManager[562]: [1481032051.8160] audit: op="connection-add-activate" uuid="7c83331f-74b7-4ef8-98a5-67bc03e24b36" name="EOZWireless2" pid=1137 uid=1000 result="success" Dec 6 14:47:31 waterhole NetworkManager[562]: [1481032051.8213] device (wlx1c7ee55ecc0a): state change: disconnected -> prepare (reason 'none') [30 40 0] Dec 6 14:47:31 waterhole NetworkManager[562]: [1481032051.8221] manager: NetworkManager state is now CONNECTING Dec 6 14:47:31 waterhole NetworkManager[562]: [1481032051.8611] device (wlx1c7ee55ecc0a): set-hw-addr: set-cloned MAC address to 1C:7E:E5:5E:CC:0A (permanent) Dec 6 14:47:32 waterhole kernel: [ 205.121018] IPv6: ADDRCONF(NETDEV_UP): wlx1c7ee55ecc0a: link is not ready Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0483] device (wlx1c7ee55ecc0a): state change: prepare -> config (reason 'none') [40 50 0] Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0494] device (wlx1c7ee55ecc0a): Activation: (wifi) access point 'EOZWireless2' has security, but secrets are required. Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0505] device (wlx1c7ee55ecc0a): state change: config -> need-auth (reason 'none') [50 60 0] Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0575] device (wlx1c7ee55ecc0a): state change: need-auth -> prepare (reason 'none') [60 40 0] Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0585] device (wlx1c7ee55ecc0a): state change: prepare -> config (reason 'none') [40 50 0] Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0592] device (wlx1c7ee55ecc0a): Activation: (wifi) connection 'EOZWireless2' has security, and secrets exist. No new secrets needed. Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0598] Config: added 'ssid' value 'EOZWireless2' Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0602] Config: added 'scan_ssid' value '1' Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0605] Config: added 'key_mgmt' value 'WPA-PSK' Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0605] Config: added 'auth_alg' value 'OPEN' Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0605] Config: added 'psk' value '' Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0625] sup-iface[0x55680e8fe970,wlx1c7ee55ecc0a]: config: set interface ap_scan to 1 Dec 6 14:47:32 waterhole NetworkManager[562]: [1481032052.0976] device (wlx1c7ee55ecc0a): supplicant interface state: inactive -> scanning Dec 6 14:47:36 waterhole wpa_supplicant[614]: wlx1c7ee55ecc0a: SME: Trying to authenticate with 4c:60:de:6b:c3:0c (SSID='EOZWireless2' freq=5200 MHz) Dec 6 14:47:36 waterhole kernel: [ 209.921607] wlx1c7ee55ecc0a: authenticate with 4c:60:de:6b:c3:0c Dec 6 14:47:36 waterhole kernel: [ 210.000312] wlx1c7ee55ecc0a: send auth to 4c:60:de:6b:c3:0c (try 1/3) Dec 6 14:47:36 waterhole kernel: [ 210.001329] wlx1c7ee55ecc0a: authenticated Dec 6 14:47:36 waterhole NetworkManager[562]: [1481032056.9302] device (wlx1c7ee55ecc0a): supplicant interface state: scanning -> authenticating Dec 6 14:47:41 waterhole wpa_supplicant[614]: wlx1c7ee55ecc0a: SME: Deauth request to the driver failed Dec 6 14:47:41 waterhole NetworkManager[562]: [1481032061.9310] device (wlx1c7ee55ecc0a): supplicant interface state: authenticating -> disconnected Dec 6 14:47:42 waterhole NetworkManager[562]: [1481032062.0313] device (wlx1c7ee55ecc0a): supplicant interface state: disconnected -> scanning Dec 6 14:47:46 waterhole wpa_supplicant[614]: wlx1c7ee55ecc0a: SME: Trying to authenticate with 4c:60:de:6b:c3:0c (SSID='EOZWireless2' freq=5200 MHz) Dec 6 14:47:46 waterhole kernel: [ 219.838794] wlx1c7ee55ecc0a: authenticate with 4c:60:de:6b:c3:0c Dec 6 14:47:46 waterhole kernel: [ 219.916673] wlx1c7ee55ecc0a: send auth to 4c:60:de:6b:c3:0c (try 1/3) Dec 6 14:47:46 waterhole NetworkManager[562]: [1481032066.8465] device (wlx1c7ee55ecc0a): supplicant interface state: scanning -> authenticating Dec 6 14:47:46 waterhole kernel: [ 219.920717]
Re: formato .DAT
Son MPEG, más concretamente los MPEG de los videoCD. Cualquier reproductor que admita mpeg, funcionará con los .dat. Saludos! Primera noticia que tengo de que dat sea un formato de video. Lo tengo por formato de datos generales. ¿Estás seguro? Saludos -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Freeze de Potato
Barbwired wrote: ¿Qué pasó en el SIMO? Lo de siempre: Mucho Windows. Y muy poco Linux. Yo pensé que iba a palpar el cambio en el ambiente. Pero me topé con lo de siempre. Y no me hagáis recordarlo, que me vuelvo a encabronar :-) Un saludo! A mi me han dicho que poco Linux, mucho Windows y mucho mas telefonito... :-) Saludines!! Miguel A.Vallejo, EA4EOZ [EMAIL PROTECTED]
Re: Freeze de Potato
Me referia a que ahora mismo es mas rapido un ordena con win95 que el mismo ordenador con kde o gnome... Disiento. Pues yo disiento contigo ;-) En mi pentium 133 con 38 MB de RAM windows 95 es o parece más rápido que el KDE 1.1. Por ejemplo el explorador de archivos del kde es bastante más lento (sí ya sé que por lo visto hace más cosas, pero el caso es que es más lento), tarda bastante en aparecer algunos diálogos o en cargarse alguno programas, etc. y eso sin hablar de lo que tarda en cargarse. No creo que sea por falta de memoria, sino porque el procesador se me está quedando pequeño... o que el KDE está integramente programado en C++ ¿? Estas en la misma situacion que yo, P133 pero el mio con 64 Mb de ram, w95 vuela, kde se arrastra, tanto que su uso ya se hace incomodo... :-( , de ahi el utilizar unicamente un gestor de ventanas: fvwm95 :-) Saludines! Miguel Angel Vallejo, EA4EOZ [EMAIL PROTECTED]
Re: Freeze de Potato
Hola Andres y compañia! Yo, como ves, no. De sencillo lo justito. Pero de facil de programar, una leche. ¿Ejemplo? Mira lo sencillo que es un CGI filosoficamente: un script sin más que recibe una entrada de texto y envia una salida formateada en texto ¡y lo flexible que es! Mira lo complejo que es una pagina .asp que se apoya en .dlls compiladas para generar mediante un enlace nativo u ODBC sobre una base de datos una salida lo más complicada posible. :-D, me referia a la creacion de ejecutables... Ya se que programar algo que tenga que correr bajo un ISS o un primo suyo, es de locos! Microsoft se esfuerza en complicar las cosas, y programando mucho más, sin entrar en temas de seguridad y posibilidad de hacer que un pobre usuario sepa que se le puede ejecutar al ver el correo o navegar sin desactivar la mitad de las funcionalidades de dichos programas. Cierto al 100%, por eso existe Netscape! :-D Microsoft pasa de nosotros. Cierto! Y desde luego, su politica de multitud de versiones y entornos de desarrollo no va encaminada a facilitarnos la vida, si no a hacernos depender cada vez más (MSDN, TechNet, reuniones, foros, etc.), pelas al fin y al cabo. Tu lo has dicho, ni mas ni menos, todo es por el $$$ Sinceramente, creo que seria imposible viendo los caminos que estan cogiendo tanto gnome como kde... y seria una lastima... ¿Te parecen malos? No, me parecen muy buenos, pero a mi ver, seria mejor centrar los esfuerzos en un gestor de escritorio que sea mucho mas simple, que no tenga tanta chorradita (entendedme) y que este orientado simplemente a ser eficaz. Realmente un gestor de escritorio con que gestione el cut paste, el drag drop entre aplicaciones (de todo tipo de datos, como texto, texto con formato, imagenes, ficheros, etc), el manejo de ventanas (un buen wm sin demasiados artificios) y poco mas, es todo lo que se necesita, Como realmente hace windows, yo con las facilidades que da win en este tipo de cosas, me sobra, realmente no hace falta nada mas... Pues fijate la suerte que tienes, a mi me parecen los más. Lástima que encima me toque ayudarles, si fueran menos, tendria menos carga de trabajo. Si, sera que he tenido suerte... :-) Ahí quizá lleves razón, en las X tambien faltan un monton de detalles para los que sufren sin raton. Pero con Windows, te aseguro que hay demasiados programas así, y sobre todo comerciales de mucha pasta. Lo que mas me gusta de windows es la aparente uniformidad entre las aplicaciones, al menos los shortcuts del teclado suelen ser los mismos en todos los programas, pero en linux, tenemos los X-cosas que van a la bola del programador, por otro lado los K-cosas que van por otro y los Gnome-cosas que van por otro... Eso es lo que no me gusta... En la variedad esta el gusto, pero esto creo que impide que un programador decida libremente como hacer su programa. Pues realmente, si quiere convencer a todos los usuarios, debera hacer tres programas distintos (Un X, un K y un G) con el trabajo que ello lleva... Yo mismo tengo pensado portar una aplicacion de win a linux, y no se ni siquiera que libreria escojer (lesstif, Qt, GTK...), Podria haber una sola, al estilo de la libc (que aunque vaya cambiando de version, sigue siendo ella)... A mi ver, Gnome y KDE estan haciendo una batalla tan absurda por ver quien incluye mas caracteristicas en sus gestores (engordandolos en una palabra) que se pierde el autentico significado de un gestor de escritorio: La interaccion rapida y comoda para el usuario entre todas la aplicaciones del sistema... Al menos, asi lo veo, por supuesto puedo estar equivocado ( y probablemente lo este)... Saludos y gracias por responder a mi mensaje! Miguel Angel Vallejo, EA4EOZ [EMAIL PROTECTED]
Re: Porqué no uso tal o pascual (Era: Re: Freeze de Potato)
J. Carlos Muro decía: A mí lo que me gustaría saber es por qué hay al parecer tanta gente que se niega a utilizar 'el WindowMaker/KDE/Gnome' como F. José Avila decía. Mi caso particular: .- No hay quien lo instale correctamente (nisiquiera siguiendo el README al pie de la letra) .- No hay quien lo configure para que no este constantemente sacando mensajes de error .- Son lentos de narices (muucho mas que un 98 con vistas web de las carpetas .- Dejan el ordenador en estado catatonico... Donde estan aquellos tiempos donde se proclamaba a los cuatro vientos que Linux no consumia apenas recursos y que funcionaba en maquinas de juguete (386 + 4Mb) ?? debería hacerse una especie de encuesta en la web de Debian, para que realmente se sepa qué es lo que quieren los usuarios. Yo creo que habrían sorpresas ;-) Pero mchas sorpresas... Miguel A.Vallejo [EMAIL PROTECTED]
Re: [Off-Topic] Fonética (Re: Acentos y similares conteclado yanki)
Yo tambien tengo un teclado con distribucion inglesa, sin embargo, aunque en modo texto fue mas o menos facil modificar el mapa de teclado, cada aplicacion lee lo que le sale de los bytes de el... En las X, mejor no hablamos... Acentos?? Enyes??? Desde que no uso windows, han desaparecido... :'( Alguien ha conseguido sacar caracteres buenos con un teclado US?? Como?? Miguel A.Vallejo [EMAIL PROTECTED]
Re: Freeze de Potato
Andres Seco Hernandez wrote: Mira, Windows tendrá lo que tu quiera pero es muy sencillo y agradable de usar. Y también de programar. No quiere decir esto que esté de acuerdo con la filosofía de Microsoft ni mucho menos. Estoy totalmente de acuerdo con este hombre... Ni tampoco que haya que sacrificar facilidad por estabilidad. Pero imagina algo tan estable como Linux con una interfaz tan buena como la de Windows. Sinceramente, creo que seria imposible viendo los caminos que estan cogiendo tanto gnome como kde... y seria una lastima... Esto si que lo discuto. De sencillo nada. Yo doy soporte a un monton de gente que usa Windows y están tan perdidos como si lo hubieran visto ayer por primera vez, y eso que llevan años viendolo. Siempre hay gente por todos los sitios que no se les da bien eso de los ordenadores, pero te aseguro que son los menos, es mas, a todas las personas que conozco en este caso que comentas, ni siquiera tenian el concepto de fichero o programa, a si que imaginate... ¿Agradable? Es una cuestión de gustos. 100% de acuerdo... A mi no me parece agradable usar aplicaciones que no han tenido en cuenta su uso sin raton el dia que se te ha caido de la mesa y no rula, ??, he utilizado mucho los w9x sin raton y te aseguro que se manejan bastante bien, mcho mejor que las X... ni que se cuelgue cuando le de la gana (limpito ¿eh? recien instaladito en un equipo que Linux lleva 8 meses sin parar, sin problemas hardware). Si, esto le da por hacerlo muy amenudo... pero realmente pocas veces lo hace cuando le da la gana... siempre es por algo... lo malo es que muy pocas veces puedes solucionarlo, como ocurre en linux... Saludines!!! Miguel A.Vallejo [EMAIL PROTECTED]
Re: Freeze de Potato
Ricard wrote: Que programar es mas agradable, COMO? CUANDO? QUE? me he perdido algo? puf! lo siento, quiza no sea parcial pero despues realizar durante 3 meses un proyecto en C++ bajo win32 ... nose ... pruebalo Veo que has tenido dolores de cabeza con esto :-) y si lo has hecho EN QUE PUNYETAS TE BASAS, programar en win32 comparado con Linux esta a años luz de distancia. Si, en las tripas, si, pero desde el lado del programador, yo al menos me siento mejor alli, quizas ya la costumbre... Sinceramente, creo que seria imposible viendo los caminos que estan cogiendo tanto gnome como kde... y seria una lastima... Por que, tan fantastica es la interface de Windows? nose, poneos un Wine. Me referia a que ahora mismo es mas rapido un ordena con win95 que el mismo ordenador con kde o gnome... Es dificil que alguin sistema se cuelgue por nada, tu afirmacion es un simple amago. Cojete el quickime 3 para win y mira una peli medio-bajada de internet (de las que se corta a mitad de fichero), pete total!, pero claro, como lo cambias?? No se puede... Por ejemplo, cargas una pagina de internet con un ActiveX y el windows se cuelga. Que me digas que es porque has cargado un ActiveX es una tonteria ... que debe hacer el usuario, no cargarlos? Evitar el explorer... Que es active x? :-) Netscape power!! :-D Llegara el dia que Linux sera tan o mas grafico que Windows, pero aun no ha llegado, porque las cosas aun fallan demasiado. De aqui a un año tendremos un panorama muy diferente. Yo creo que ahora mismo, linux es mas grafico que cualquier windows, al menos de cara al usuario, pero, a que coste?? Cuanta maquina se desperdicia en el gestor de escritorio?? Demasiada para mi gusto... Seguire duranto mucho tiempo con el fvwm95... Saludines!! Miguel A.Vallejo [EMAIL PROTECTED]