Re: [arch-general] Arch Linux PC as a Remote Desktop Node
On 07/30/2018 12:32 PM, Ralph Corderoy wrote: > Hi David, > >> Foxtrot Mike wrote: >>> I want to use the Arch system as a Remote Desktop node. The Arch >>> system would use a login manager (such as lightdm) to authenticate >>> users from Windows Domain. Once the user has been authenticated, the >>> system is supposed to automatically open a RDP connection (using >>> freerdp) to the Windows Server (if possible, using the credentials >>> provided to lightdm so the user doesn't have to enter his password >>> twice). As soon as the user quits the RDP session, his X-session >>> should also be closed automatically. >> I am totally confused as to what and why you are wanting to put your >> Arch server between the employees and the RDP session to the windows >> server?? Where is the code they produce stored? Windows? Arch Server? > Mike didn't mention an Arch server. He has a Windows server and ten > PCs. He wants each PC to be a dumb RDP terminal to the server. He is > considering running `the Arch system' on each dumb PC to provide the > RDP-terminal functionality, e.g. an X server and RDP client. > Thanks for summarizing it so elegantly.
Re: [arch-general] Arch Linux PC as a Remote Desktop Node
On 07/29/2018 10:01 PM, Chris Warrick via arch-general wrote: > On Fri, 27 Jul 2018 at 19:07, Foxtrot Mike via arch-general > wrote: >> Hi all, >> >> Currently we have around 10 employees who develop software using Visual >> Studio. The idea is to install the development tools on the Windows >> Server system, and to have all the developers connect to the server over >> RDP using low-end low-power computers. The server is pretty beefy >> though. The low end client PCs will save up-front cost as well as power >> bills. The network backend will not have any issue with the increased >> RDP traffic. > If I were one of the employees, I wouldn’t be very happy about the > idea. RDP will never be as reliable and snappy as working on a > physical machine, even if the server was more powerful. Developers > aren’t the best employees to make savings on. Do note that the savings > will be limited due to eg. Windows Server licensing. Also, have you > tested it and made sure that all the software is compatible with > concurrent use via RDP? Well, the decision to use RDP comes from higher management. There was a suggestion to use proprietary thin clients (such as from HP), but I thought I could make the 'system' myself with some tinkering. The network backend is enough for these RDP sessions. May be the lag won't be that bad. If the worst happens, I'll write all this effort and time off as 'experience'. > But if you have to, here’s an idea: > > One Linux user account, auto-login into X. That user account runs > Openbox (as something more user-friendly), which auto-starts a RDP > client. > After the user ends their Windows session, a dialog box (eg. from > Zenity) appears, asking to restart the Windows session or shut down > the computer, perhaps with a timeout. > Thanks. That's the approach I have decided would be the best, based on the very helpful comments from other users. One thing though. How can I automatically detect that FreeRDP session has been closed so that I can run a small Zenity script? Regards.
Re: [arch-general] Arch Linux PC as a Remote Desktop Node
On 07/28/2018 02:52 AM, cyelae via arch-general wrote: > On 2018-07-27 19:07, Foxtrot Mike via arch-general wrote: >> Hi all, >> >> Currently we have around 10 employees who develop software using Visual >> Studio. The idea is to install the development tools on the Windows >> Server system, and to have all the developers connect to the server over >> RDP using low-end low-power computers. The server is pretty beefy >> though. The low end client PCs will save up-front cost as well as power >> bills. The network backend will not have any issue with the increased >> RDP traffic. > > I'm not sure about the auth part given my little experience with it, > but if you're going to log into a server via RDP, can't you simply > have your lightweight machine automatically open a default session > [1], connect to the windows server, and authenticate users there? You mean I should have something like a guest account on Linux for opening an X session, and then the RDP application would connect to the Windows Server and ask for Windows Domain password? I think it could be done, but I'll have to severely limit the guest account from security point of view. > > If you're going to only have one app running on the client machines, > you don't need a window manager; xinit do that [2] [3] I tried opening Firefox using init, without any window manager. Firefox did open but there were some issues with graphics. Such as upon right clicking, the options menu wouldn't show. Also, youtube for some reason didn't show any thumbnail. After some research it seemed that directly running an application from xinit without a window manager is not a good idea if the application is graphically complex. > > [1] https://wiki.archlinux.org/index.php/Xinit#Autostart_X_at_login > [2] > https://wiki.archlinux.org/index.php/Xinit#Starting_applications_without_a_window_manager > [3] https://bbs.archlinux.org/viewtopic.php?id=107319
Re: [arch-general] Arch Linux PC as a Remote Desktop Node
On 07/27/2018 11:17 PM, Giancarlo Razzolini wrote: > Em julho 27, 2018 14:46 Foxtrot Mike via arch-general escreveu: >> >> The issue with x2go and ltsp is that I'll have to separately manage >> username and passwords for local Linux login. The solution that I'd >> rather prefer would use Active directory authentication so the >> current system administrator won't have to do anything extra. The >> group policies are already there. Once the Arch system is properly >> configured, I'd disable local logins so there will be very limited >> chance for a user to corrupt/modify Arch system. And ideally, the >> user would have no way to interact with the local system. Thats why I >> want to limit the user to freeRDP. Anything else, and the X-session >> expires. > > You have more than one option to authenticate to windows AD servers > [0] . You > have PAM Ldap, winbind, making a samba server the secondary > controller, etc. I thought these options worked together, i.e, I'd to use samba PAM and winbing all together. Thanks for the info. I'll look deeper into it. > > You will probably need a local home dir for storing session data, but > this can > be created/destroyed on demand. > >> >> Plus, I am very much into embedded linux systems (routers, SBCs, >> etc). I think putting the various pieces together would be give me a >> lot more to learn as compared to using a third party specialized >> software such as a kiosk script. >> > > Why reinvent the wheel here? I understand the need for learning, but I > wouldn't > do this on something that is intended as a production system. Again, > don't use > plain X protocol over the network, it's very wasteful. I plan to use RDP. I think it's not the same as using 'plain X protocol over the network' since RDP includes encryption and compression, afaik. Please correct me if I'm wrong. > > Regards, > Giancarlo Razzolini > > [0] https://wiki.archlinux.org/index.php/Active_Directory_Integration
Re: [arch-general] Arch Linux PC as a Remote Desktop Node
On 07/27/2018 10:16 PM, Giancarlo Razzolini wrote: > Em julho 27, 2018 14:07 Foxtrot Mike via arch-general escreveu: >> >> Here are the major tasks: >> >> 1- Ask LightDM to use Windows Domain (Kerberos) authentication. I am >> a little confused. There are supposedly many different ways with >> little changes to do this. [1] is one solution. LDAP is also a >> possibility. I need advice from someone who knows this field better >> than me :p >> >> 2- How to ask i3-wm (my default wm) to run freerdp at login? I guess >> [2] will get this done. >> >> 3- How to ask freerdp to authenticate using the ticket received from >> TGT during LightDM Domain authentication? If I could somehow >> configure freerdp to use Kerberos Tickets then the user won't have to >> enter his Domain password again. >> >> 4- How to ask i3-wm to close the X-session when freeRDP quits? I read >> something a while ago about .xsession files to achieve this >> functionality, but can't find it now. >> > Hi Mike, > > You have some options here. I suggest you look into x2go and ltsp for > starters. > I don't suggest you use plain X over the network. > > With those 2 options you can have this kiosk mode you want, for the > users to only > be able to access windows. > > Regards, > Giancarlo Razzolini Thanks for the reply. The issue with x2go and ltsp is that I'll have to separately manage username and passwords for local Linux login. The solution that I'd rather prefer would use Active directory authentication so the current system administrator won't have to do anything extra. The group policies are already there. Once the Arch system is properly configured, I'd disable local logins so there will be very limited chance for a user to corrupt/modify Arch system. And ideally, the user would have no way to interact with the local system. Thats why I want to limit the user to freeRDP. Anything else, and the X-session expires. Plus, I am very much into embedded linux systems (routers, SBCs, etc). I think putting the various pieces together would be give me a lot more to learn as compared to using a third party specialized software such as a kiosk script. Regards.
[arch-general] Arch Linux PC as a Remote Desktop Node
Hi all, Currently we have around 10 employees who develop software using Visual Studio. The idea is to install the development tools on the Windows Server system, and to have all the developers connect to the server over RDP using low-end low-power computers. The server is pretty beefy though. The low end client PCs will save up-front cost as well as power bills. The network backend will not have any issue with the increased RDP traffic. I have been using Arch Linux for some years now, so I undertook this as a research project. I want to use the Arch system as a Remote Desktop node. The Arch system would use a login manager (such as lightdm) to authenticate users from Windows Domain. Once the user has been authenticated, the system is supposed to automatically open a RDP connection (using freerdp) to the Windows Server (if possible, using the credentials provided to lightdm so the user doesn't have to enter his password twice). As soon as the user quits the RDP session, his X-session should also be closed automatically. Here are the major tasks: 1- Ask LightDM to use Windows Domain (Kerberos) authentication. I am a little confused. There are supposedly many different ways with little changes to do this. [1] is one solution. LDAP is also a possibility. I need advice from someone who knows this field better than me :p 2- How to ask i3-wm (my default wm) to run freerdp at login? I guess [2] will get this done. 3- How to ask freerdp to authenticate using the ticket received from TGT during LightDM Domain authentication? If I could somehow configure freerdp to use Kerberos Tickets then the user won't have to enter his Domain password again. 4- How to ask i3-wm to close the X-session when freeRDP quits? I read something a while ago about .xsession files to achieve this functionality, but can't find it now. Any help would be appreciated! Regards. [1] https://wiki.archlinux.org/index.php/Active_Directory_Integration#Join_the_domain [2] https://i3wm.org/docs/userguide.html#_automatically_starting_applications_on_i3_startup
Re: [arch-general] ddcutil not working with nvidia card
On 01/23/2018 09:43 PM, David C. Rankin wrote: On 1/23/2018 8:02 AM, Foxtrot Mike via arch-general wrote: I tried googling this error, but didn't find anything helpful. So I tried using xbacklight instead. It didn't give any positive results either, even when I provided different kernel paramters as discribed in [1]. No matter what I do with acpi and kernel parameters, "/sys/class/backlight/acpi_video0/brightness" doesn't give actual values. Also, 'xbacklight set' has no effect on screen brightness at all. I also tried adding `Option "RegistryDwords" "EnableBrightnessControl=1"` in "/usr/share/X11/xorg.conf.d/10-nvidia-drm-outputclass.conf". Still keyboard brigthness keys couldn't change the brightness. Can someone help me get ddcutil working with nvidia proprietary driver? I doubt I can held get ddcutil working, but the problem is a lack of an nvidia kernel module interface with /sys/class/backlight (which can be frustrating). You do not want to use xbacklight -- which ironically does not control actual backlight (display power), but instead controls pixel darkness (the amount of black/white per-pixel). Reducing backlight does not reduce display power, it just makes the pixels darker giving the impression of reduced brightness. (this is really bad for laptop battery life where reducing the display power is important) You want xrandr. Query with `xrandr -q` and find your interface and you can then set the desired brightness with something similar to xrandr --output LVDS-0 --brightness 42 (where LVDS-0 is your monitor connections shown by --query) Modern desktops generally integrate xrandr control (if not, you can simply implement a short script that allows you to check and set the desired brightness. There were also several AUR packages (nvidia-bl, and nvidiabl or similar) that provided a /sys/class/backlight kernel module that would would work with most laptops, but I don't think they are maintained any longer. (there was an example script on the X11 wiki some time ago for doing this) That is another option. I tried using xrandr to adjust brightness, but it doesn't modify brightness on monitor. Then I stumbled upon [1] to learn that backlight can't be actually controlled with xrandr on LCD/LED monitors: --brightness brightness Multiply the gamma values on the crtc currently attached to the output to specified floating value. Useful for overly bright or overly dim outputs. However, this is a software only modification, if your hardware has support to actually change the brightness, you will probably prefer to use xbacklight. And as described in the initial question, xbacklight doesn't work with my NVIDIA card despite all the kernel parameters. However, with ddcutil I was able to set contrast and brightness at monitor hardware level so that even at low brightness levels the screen remained perfectly legible. With xrandr, modifying the brightness distorts the colors of text. The text is readable but not white anymore. I'll see if there's some other solution. Regards. [1] https://askubuntu.com/questions/62249/how-do-you-change-brightness-color-and-sharpness-from-command-line
[arch-general] ddcutil not working with nvidia card
Hi all, With my AMD card, I was able to set my monitor's (ViewSonic VA2249, connected via DVI) brightness and contrast from command line using: ddcutil --display 1 setvcp 12 20 Later on, I swapped AMD card with an NVIDIA one (using nvidia 387.34-20). I tried using ddcutil with this new card as well, but I got this error: ~$ sudo ddcutil detect Failure getting EDID for /dev/i2c-5: status code=DDCRC_EDID(-3016): invalid EDID No displays found ~$ I tried googling this error, but didn't find anything helpful. So I tried using xbacklight instead. It didn't give any positive results either, even when I provided different kernel paramters as discribed in [1]. No matter what I do with acpi and kernel parameters, "/sys/class/backlight/acpi_video0/brightness" doesn't give actual values. Also, 'xbacklight set' has no effect on screen brightness at all. I also tried adding `Option "RegistryDwords" "EnableBrightnessControl=1"` in "/usr/share/X11/xorg.conf.d/10-nvidia-drm-outputclass.conf". Still keyboard brigthness keys couldn't change the brightness. Can someone help me get ddcutil working with nvidia proprietary driver? Regards. [1] https://wiki.archlinux.org/index.php/Backlight#Kernel_command-line_options
Re: [arch-general] Why there is no NetworkManager in ArchISO
Wow. I could have sworn that email was right at the top of my archlinux folder when i fired up my email client earlier today. Now it's no where. Perhaps an unlucky coincident. Peace. On 01/12/2018 09:32 PM, Bruno Pagani via arch-general wrote: > Le 12/01/2018 à 17:25, Foxtrot Mike via arch-general a écrit : > >> Hi, >> >> I have asked quite a few questions here myself, and received excellent >> guidance and help. >> I was in no way trying to imply the superiority of one tool over >> another, or of keyboard over mouse. >> Was just trying to add something (helpful?) to this conversation based >> on what I knew. >> >> It turns out what I thought to be 'network manager is not actually the >> network manager, as exlained by Mr Eli. Thanks for the explanation. >> >> No offence meant. And none taken. >> >> Peace. > Well the sole issue with this is that the same Eli already answered that > exact point 5 months ago (on July 24th) in that thread you bumped. ;) > > Regards, > Bruno >
Re: [arch-general] Why there is no NetworkManager in ArchISO
Hi, I have asked quite a few questions here myself, and received excellent guidance and help. I was in no way trying to imply the superiority of one tool over another, or of keyboard over mouse. Was just trying to add something (helpful?) to this conversation based on what I knew. It turns out what I thought to be 'network manager is not actually the network manager, as exlained by Mr Eli. Thanks for the explanation. No offence meant. And none taken. Peace. On 01/12/2018 07:57 PM, Eli Schwartz via arch-general wrote: > On 01/12/2018 08:48 AM, Foxtrot Mike via arch-general wrote: >> Hi, >> >> I use arch with i3wm. Been using for around 3 years. Still lovin it. >> Dont really need a network manager when netctl can handle my wifi pretty >> nicely. Actually I prefer typing over mouse-ing, so I would always >> prefer a tool like netctl over mouse-driven tools such as network manager. > Please keep in mind that NetworkManager is primarily a backgrounded > daemon, with two CLI reference frontends -- nmtui and nmcli. > > network-manager-applet, which most people think of as "NetworkManager", > is actually a thirdparty GUI for NetworkManager, developed by some of > the same people as a separate project. Said GUI is not even installed > with networkmanager, it is a completely separate package that *depends* > on networkmanager. > > Please learn more about your system/tools before passing judgment on them. >
Re: [arch-general] Why there is no NetworkManager in ArchISO
Hi, I use arch with i3wm. Been using for around 3 years. Still lovin it. Dont really need a network manager when netctl can handle my wifi pretty nicely. Actually I prefer typing over mouse-ing, so I would always prefer a tool like netctl over mouse-driven tools such as network manager. Peace. On 07/24/2017 10:30 AM, Junayeed Ahnaf via arch-general wrote: > Hello, > > Why is there no NetworkManager in ArchISO? Isn't it widely accepted as > the go to method of connecting to internet in Linux? Is there any reason > for it not to be default? > > Thanks > > J >
Re: [arch-general] [Solved] Switch from AMD GPU to nvidia GPU
Hi, Thanks for the reply. So I re-installed the AMD graphics card. Updated the system with 'pacman -Syu'. Created a bootable arch USB. Shut the PC off and installed nvidia card. Booted up the system. Everything worked fine this time. Don't know what the issue was but perhaps updating the system resolved it. Regards On 12/30/2017 11:12 PM, Florijan Hamzic wrote: Hi, Afaik, you should at least be able to boot into the system, maybe you can try to but into a arch boot image and read the log files. If you have some more information you could check this guide: https://wiki.archlinux.org/index.php/NVIDIA/Troubleshooting Foxtrot Mike via arch-general mailto:arch-general@archlinux.org>> schrieb am Sa., 30. Dez. 2017, 18:54: Hi all, I was using amd r7 260x card without any problem with 'amdgpu' module. Now I want to install my new nvidia card (1050ti). I assumed that installing 'nvidia' package from pacman would be enough to be able to use my new GPU. (Apparently not). The nvidia related packages installed on my PC are: ~$ pacman -Qs nvidia local/libvdpau 1.1.1+3+ga21bf7a-1 Nvidia VDPAU library local/nvidia 387.34-15 NVIDIA drivers for linux local/nvidia-utils 387.34-2 NVIDIA drivers utilities Upon rebooting after switching the cards, arch boots fine until "Reached target graphical interface" appears on the display. It seemed the PC was stuck, but switching to tty2 resulted in flashing keyboard lights. Which, according to a quick Google search, means kernel panic. Moreover, [1] says that "the files are read in ASCII order, and by convention their names start with XX- (two digits and a hyphen, so that for example 10 is read before 20).". But I have multiple files in "/usr/share/X11/xorg.conf.d" starting with the same number 10: ~$ ls /usr/share/X11/xorg.conf.d/ 10-amdgpu.conf10-quirks.conf 10-nvidia-drm-outputclass.conf40-libinput.conf Please help me in getting my new card up and running. Regards. [1] https://wiki.archlinux.org/index.php/xorg#Using_.conf_files
[arch-general] Switch from AMD GPU to nvidia GPU
Hi all, I was using amd r7 260x card without any problem with 'amdgpu' module. Now I want to install my new nvidia card (1050ti). I assumed that installing 'nvidia' package from pacman would be enough to be able to use my new GPU. (Apparently not). The nvidia related packages installed on my PC are: ~$ pacman -Qs nvidia local/libvdpau 1.1.1+3+ga21bf7a-1 Nvidia VDPAU library local/nvidia 387.34-15 NVIDIA drivers for linux local/nvidia-utils 387.34-2 NVIDIA drivers utilities Upon rebooting after switching the cards, arch boots fine until "Reached target graphical interface" appears on the display. It seemed the PC was stuck, but switching to tty2 resulted in flashing keyboard lights. Which, according to a quick Google search, means kernel panic. Moreover, [1] says that "the files are read in ASCII order, and by convention their names start with XX- (two digits and a hyphen, so that for example 10 is read before 20).". But I have multiple files in "/usr/share/X11/xorg.conf.d" starting with the same number 10: ~$ ls /usr/share/X11/xorg.conf.d/ 10-amdgpu.conf 10-quirks.conf 10-nvidia-drm-outputclass.conf 40-libinput.conf Please help me in getting my new card up and running. Regards. [1] https://wiki.archlinux.org/index.php/xorg#Using_.conf_files
Re: [arch-general] Clone a block device using 'cat'
On 12/16/2017 01:03 AM, Maciek Borzecki wrote: > On Fri, Dec 15, 2017 at 8:37 PM, Foxtrot Mike via arch-general > wrote: >> Note: This is not an Arch related question. I tried asking this question >> elsewhere, but didnt get any useful response. In the past I have received >> very helpful responses on this mail list so I am posting this here in the >> hope that someone would be willing/able to help me out. >> >> Hi, >> >> I am able to telnet into a router that's running a customized linux >> firmware. It was compiled using 'buildroot'. >> >> I was able to dig out some specs of the router. I intend to clone the >> router's firmware for further study. 'dd' is not available so I tried using >> cat to throw the entire contents of the flash drive (mtdblock0) to my pc >> using netcat: >> >> Router: >> cat /dev/mtdblock0 | nc ip port >> >> PC: >> nc -lp port > routerFirmware >> >> Then on my PC I used 'binwalk' to examine the downloaded file. The result is: >> >> DECIMAL HEXADECIMAL DESCRIPTION >> >> 0 0x0 Squashfs filesystem, little endian, >> non-standard signature, version 4.0, compression:gzip, size: xxx bytes, >> xxx inodes, blocksize: xx bytes, created: -xx-xx xx:xx:xx >> >> Seems good so far. But '# unsquash routerFlash' results in: >> >> Can't find a SQUASHFS superblock on routerFlash >> >> The most probable reason I think is that 'cat' and 'nc' are not able to >> clone the entire flash as I intend. I am stuck here. Any help would be >> appreciated! > It's possible that squashfs used to build the image carried some > vendor modifications. Is not that uncommon actually. You can try your > luck with: https://github.com/devttys0/sasquatch or > https://github.com/rampageX/firmware-mod-kit > > Cheers, Thanks a lot for the very helpful pointer. I got it working with sasquatch. Turns out the file is actually compressed with lzma, not gzip as was reported by unsquash. Regards,
[arch-general] Clone a block device using 'cat'
Note: This is not an Arch related question. I tried asking this question elsewhere, but didnt get any useful response. In the past I have received very helpful responses on this mail list so I am posting this here in the hope that someone would be willing/able to help me out. Hi, I am able to telnet into a router that's running a customized linux firmware. It was compiled using 'buildroot'. I was able to dig out some specs of the router. I intend to clone the router's firmware for further study. 'dd' is not available so I tried using cat to throw the entire contents of the flash drive (mtdblock0) to my pc using netcat: Router: cat /dev/mtdblock0 | nc ip port PC: nc -lp port > routerFirmware Then on my PC I used 'binwalk' to examine the downloaded file. The result is: DECIMAL HEXADECIMAL DESCRIPTION 0 0x0 Squashfs filesystem, little endian, non-standard signature, version 4.0, compression:gzip, size: xxx bytes, xxx inodes, blocksize: xx bytes, created: -xx-xx xx:xx:xx Seems good so far. But '# unsquash routerFlash' results in: Can't find a SQUASHFS superblock on routerFlash The most probable reason I think is that 'cat' and 'nc' are not able to clone the entire flash as I intend. I am stuck here. Any help would be appreciated! Regards.
[arch-general] STM32 compilation on ArchLinuxARM
TL;DR; I need a way to install arm-none-eabi-gcc on Raspberry Pi (Arch Linux ARM) using pacman or from source. Hi guys, First of all, I know that this mailing list entertains questions and information related to Arch Linux, not Arch Linux ARM. Since I have received valuable suggestions on this list before, I am going to ask my question here in the hope that someone would be able and willing to help. I couldn't find the required info on the internet. I am running Arch on my main desktop. I installed arm-none-eabi-gcc package using pacman so I can compile code for my STM32F103 board. I switched from Raspbian to ArchLinuxARM on Raspberry pi yesterday*. I have found most of the my daily usage packages in ARchLinuxARM repo, but can't find arm-none-eabi-gcc package. There's a post on ArchLinuxARM forums [1] requesting the same info, but the post didn't get any reply. Every google search tells ways to develop for Raspberry Pi on x86 computers using arm-none-eabi-gcc. I couldn't find any meaningful info about developing for ARM Cortex-Mx on Raspberry pi. Any suggestions would be apreciated. Regards, Fulcrum [1] https://archlinuxarm.org/forum/viewtopic.php?f=15&t=10060 * Hoping to use Raspberry pi as my daily computer so I could run Windows on my main desktop for gaming
Re: [arch-general] Not being able to upgrade system because of catalyst-hook
[SOLVED] Thanks for your help. FINALLY, I got catalyst working. I took many hours though. Here is what I tried (giving as much detail as possible to help others): > 2016-09-13 10:39 GMT+02:00 G. Schlisio : > What about version 15.9-12? It adds a patch to make it compatible with > Linux 4.7 > > [1] > https://aur.archlinux.org/cgit/aur.git/commit/?h=catalyst-hook&id=1ed0d7e92f486d3b1bae49b9c43b34b29585d1c0 > 1- I tried this patch. Didn't make a difference. Still catalyst-hook was causing dependencies issue with linux4.7. > You should update the catalyst-hook package manually. The latest version > supports Linux 4.7. > > Regards, > Sebastiaan > 2- Removed all the catalyst packages. 'radeon' was blacklisted to fell back to console. Upgraded the system, and reinstalled catalyst packages. Tried Xorg 1.17 and 1.18. Result: Xorg won't start. Gave some errors. Googled those errors (Can't remember the error now). Tried a few things suggested on forums. Still no luck. 3- Tried 'amdgpu' package. After installing everything and setting up, found out that amdgpu doesn't support my card (R7 260x). I needed to rebuild kernel to enable support for my card. Settled on 'radeon' instead. 'radeon' didn't let me check my fan speed and temps so I kept looking around. On 09/14/2016 06:44 AM, Mrrob wrote: > I think the OP implied they are using the catalyst & xorg117 repos. This > comment on the catalyst-hook AUR package indicates the repo has changed > and updating your pacman.conf to reflect this new location may help: > > Vi0L0 commented on 2016-08-20 23:51 > > Atm I'm unable to update wirephire mirror. Don't know when I will > get access back. > Please use hactar mirror: > [catalyst] > Server = http://mirror.hactar.bz/Vi0L0/catalyst/$arch > > mrrob > --- 4- Not sure, but I think that updating the catalyst repo link in my pacman.conf did the trick. Removed 'amdgpu', blacklisted 'radeon', and then installed catalyst-hook and catalyst-utils packages (15.12.3, i guess). Again added 'nomodeset' to kernel parameters and generated grub.cfg. And after almost 5 hours of cursing linux, I finally got my GUI back. Thanks guys for your help!
Re: [arch-general] Not being able to upgrade system because of catalyst-hook
On 09/13/2016 01:42 PM, Sebastiaan Lokhorst via arch-general wrote: > 2016-09-13 10:39 GMT+02:00 G. Schlisio : > >>> You should update the catalyst-hook package manually. The latest version >>> supports Linux 4.7. >> >> this is misleading, as there is no upgrade. >> > > What about version 15.9-12? It adds a patch to make it compatible with > Linux 4.7 > > [1] > https://aur.archlinux.org/cgit/aur.git/commit/?h=catalyst-hook&id=1ed0d7e92f486d3b1bae49b9c43b34b29585d1c0 > Thanks for the suggestion. Could you tell me how to apply this patch? I downloaded this patch, extracted it using tar, and then used 'makepkg -sri' to install it like any other AUR package. It installed fine but now this slightly modified error occurs while upgrading: :: catalyst-hook: requires linux<4.7
[arch-general] Not being able to upgrade system because of catalyst-hook
Hi all, For almost a month, I have been unable to upgrade my system (# sudo pacman -Su), apparently because catalyst-hook won't work with Linux 4.7. I thought that in a few weeks atmost, catalyst-hook would be 'upgraded' to support Linux 4.7, but still no luck. I am a beginner, so may be I am missing something. I couldn't find any such info on archWiki. Here's how I have been upgrading my system for the last year: = = >sudo pacman -Syy [sudo] password for foxtrot: :: Synchronizing package databases... xorg117 25.0 KiB 60.3K/s 00:00 [##] 100 catalyst 12.3 KiB 616K/s 00:00 [##] 100 core 119.4 KiB 64.3K/s 00:02 [##] 100 extra 1760.5 KiB 70.5K/s 00:25 [##] 100 community 3.6 MiB 116K/s 00:32 [##] 100 multilib 172.6 KiB 119K/s 00:01 [##] 100 >sudo pacman -Su :: Starting full system upgrade... :: Replace libdbus with core/dbus? [Y/n] y resolving dependencies... looking for conflicting packages... error: failed to prepare transaction (could not satisfy dependencies) :: catalyst-hook: installing linux (4.7.2-1) breaks dependency 'linux<4.7' = = BTW I am using xorg117. Any help would be appreciated! Regards, Fulcrum
[arch-general] MSI AMD R7 260x fan speed with AMDGPU
Hi all, I have a little problem with my R7 260x because of which I must check if the fan is working at startup. If it isn't, I give the fan a little push with my finger so it starts working. The fan gets jammed every fortnight or so (I didn't send the fan for repair or warranty claim because I'm sooo lazy, and now I am in another country). Anyway, the point is that I must be able to see my fan speed. Currently, I am using AMD Catalyst just because it lets me find out my fan speed with: aticonfig --pplib-cmd "get fanspeed 0" I would like to get rid of this driver and install AMDGPU. Last time I tried AMDGPU it was around an year ago, and back then fan monitoring fucntion wasn't available. I would like to ask you guys if this function is available now? And if not, how can I get fan speed readings without the AMD Catalyst software? Any help would be appreciated! regards, Fulcrum
[arch-general] Problem with MATLAB libraries
Hi all, I have MATLAB 8.6 (2015b). If I install MATLAB in the default directory (/usr/local/MATLAB), everything works fine. However, I want to install MATLAB on a different partition (/mnt/Software). I installed MATLAB on Software parition successfully, but when I try to run matlab, I get this error: = >pwd /mnt/Software/LinuxMatlab/installed/bin >./matlab expr: error while loading shared libraries: libc.so.6: cannot open shared object file: No such file or directory /bin/uname: error while loading shared libraries: libc.so.6: cannot open shared object file: No such file or directory Sorry! We could not determine the machine architecture for your host. Please contact: MathWorks Technical Support for further assistance. trap: usage: trap [-lp] [[arg] signal_spec ...] = I think I need to link some libraries so that MATLAB can access them from the Software parition. I am not quite sure how to do it and which libraries do I need to symlink. Any help would be appreciated! Regards, Fulcrum