Stage2 starts successfully, but there is no mouse cursor whatsoever. I tried to get as far as I could using navigation keys, but gave up in diskdrake. Some screens have the basic "X" cursor, but it does not move when the touchpad is used. Possibly this happened as part of an attempt to enable tap-to-click ? Reproducible: Steps to Reproduce:
Hm, this is probably because of: - make psmouse modular (mga#15660) that I changed in kernel-4.4.0-0.rc8.1.mga6 I guess we need to start preloading psmouse in installer the same way we do for evdev
CC: (none) => thierry.vignaud, tmb
I wonder if adding /lib/udev/rules.d/70-mouse.rules to stage2 would make udev to autoload psmouse?
CC: (none) => mageiaSource RPM: (none) => drakx-installer-stage2
Ah, didn't even think about that... I just submitted drakx-installer-stage2-17.13-5.mga6 that preloads psmouse...
Doesn't do the trick, I'm afraid. Same symptom - no mouse cursor. Is the psmouse driver enough for a touchpad, or is something else needed, possibly with the addition of some xorg.conf stuff ?
What driver is used in the installed system? Was drakx working OK with your touchpad in mga5? Please try again then after a couple steps: - plug a USB key - go to tty2 - run the "bug" command - attach the "report.bug" file you'll find on this USB key to this bug report
Keywords: (none) => NEEDINFO
(In reply to Thierry Vignaud from comment #5) > What driver is used in the installed system? According to harddrake2 it's a SynPS/2 Synaptics Touchpad vendor id 0002 device id 0007 module mouse0 > Was drakx working OK with your touchpad in mga5? Yes, since MGA forked and right up through Dec 21 with cauldron. > > Please try again then after a couple steps: > - plug a USB key > - go to tty2 > - run the "bug" command > - attach the "report.bug" file you'll find on this USB key to this bug report I'll try to get to this today.
Hmm, one interesting thing... I ran mcc from the command line in order to run harddrake2, and got the following in stdout/stderr: Warning: skip "Elantech" field => "" Warning: skip "Numpad" field => "" Warning: skip "Handlers" field => "HASH(0x2159470)" Warning: skip "Synaptics" field => "1" Warning: skip "bustype" field => "0011" Warning: skip "sysfs_path" field => "/devices/platform/i8042/serio4/input/input16" Warning: skip "location" field => "isa0060/serio4/input0" Warning: skip "ALPS" field => "" modinfo: ERROR: Module mouse0 not found. [
However: [root@ftglap ~]# lsmod | grep mouse psmouse 122880 0
Should also mention re: bug#17471 that I'm using the nonfree isolinux and I *do* have the keyboard recognized while the 17471 OP does not. This may have more to do with the ancillary files (e. g. .rdz) than with the kernel itself.
I'm having some trouble trying to locate a USB key, but I retested and noted that 1) Neither tty3 nor tty4 have any messages that pertain to any mouse driver I can identify 2) From tty2, "lsmod | grep mouse" gives nothing. I visually checked the lsmod output to be sure, and psmouse does not appear.
Oops, my bad... It does not help to probe a module that is not available,... the drakx list_modules.pm needs to be updated to add the module to boot images I will fix up and release new drakx-installer-images
commit 3ecd99eeabe9ac660bf181ab49c363c679540395 Author: Thomas Backlund <tmb@...> Date: Mon Jan 11 00:20:06 2016 +0200 psmouse is now modular, so add it to list_modules.pm (mga#17470) --- Commit Link: http://gitweb.mageia.org/software/drakx/commit/?id=3ecd99eeabe9ac660bf181ab49c363c679540395
drakx-installer-images-2.31-2.mga6 with the fix should be on mirrors soonish... Please try again with the new boot(-nonfree).iso when it's available
CC: (none) => wilcal.int
*** Bug 17471 has been marked as a duplicate of this bug. ***
I really hate to report this, but we're not there yet. The symptom is still the same - keyboard OK, no mouse. However, lsmod verifies that both psmouse and evdev are loaded, but have a use count of 0. The following message appears on tty4, and it did before. Actually, there are 4 such messages with different devices, but all the same error - no such file or directory: systemd-udevd: failed to execute '/usr/lib/udev/libinputdevice-group' 'libinput-device-group/sys/devices/platform/i8042/serio4/input/input9/event2' : no such file or directory The only reason this sticks out to me is the fact that it seems to deal with input devices.
commit 8639b390a5052bed277db6a0db6ce2b2e0bac0b3 Author: Thierry Vignaud <thierry.vignaud@...> Date: Mon Jan 11 19:58:09 2016 +0100 add bug ID (mga#17470) --- Commit Link: http://gitweb.mageia.org/software/drakx/commit/?id=8639b390a5052bed277db6a0db6ce2b2e0bac0b3
Interesting. Bug#17486 is pretty much a duplicate, but the OP claims that the mouse cursor is really there, just invisible, and that if you get lucky and click it while it's over something clickable, the click takes effect. Maybe this is another case of buttons and sliders being there but not getting drawn ?
*** Bug 17486 has been marked as a duplicate of this bug. ***
CC: (none) => jyri2000
I can reproduce the issue as it's described in bug 17486 using the current stage2: the mouse pointer does not appear at all (in VirtualBox). I know barjac also had no mouse pointer on real hw. I'll have to check if the pointer is actually working but invisible as mentioned in bug 17486, if so it might be a GTK+3 issue as there are other themeing issues (that I'll try report this evening).
CC: (none) => zen25000
(In reply to Rémi Verschelde from comment #19) > I know > barjac also had no mouse pointer on real hw. > > I'll have to check if the pointer is actually working but invisible That was the case for me - it was invisible, but could be used by watching for button icon changes on mouseover.
That might well have been the case for me. It never occurred to me to check for an invisible mouse pointer.
Tried cauldron network install today morning and when got past bug 17485 the situation was the same - mouse cursor was there but invisible - scrollbar appeared when the cursor happened to be on it. But when I selected the installation language using keyboard, the cursor appeared. Not a usual "arrow" cursor but the "cross" cursor. Managed to finish the install, at the following start the DM window (tested with xdm and sddm) was totally frozen, even ctl+alt+Fx keys did not work. Not sure if that is related to the installation bug though...
It looks like newer gtk+ needs some cursor file We probably should include some files from /usr/share/icons/Adwaita/cursors/ (from adwaita-cursor-theme). When those do not exists, eg: gedit tries /usr/share/fonts/Type1/cursor.pfa /usr/share/fonts/misc/cursor.pcf.gz /usr/share/fonts/misc/olcursor.pcf.gz
CC: (none) => olav
*** Bug 17547 has been marked as a duplicate of this bug. ***
CC: (none) => ennael1
Renaming the report to match the current bug that is being worked on (even if the initial report about missing synaptics support was probably different and should now be fixed as per comment 13). The new focus of this bug is as described in comments 22 and 23.
Summary: Today's stage2 does not recognize a synaptics tochpad as any sort of mouse => Mouse pointer is not visible in the early steps of stage2
Keywords: NEEDINFO => (none)Status: NEW => ASSIGNEDAssignee: bugsquad => thierry.vignaud
Using /mga/x86_64/misc/drakx-in-chroot /mga/x86_64/ /T --strace 2>&1|grep cursor, excluding libraries, we can see gtk+3 tries to access: stat("/usr/share/fonts/cursor.pcf.gz", {st_mode=S_IFREG|0644, st_size=5225, ...}) = 0 open("/usr/share/fonts/cursor.pcf.gz", O_RDONLY) = 6 open("/.icons/default/cursors/col-resize", O_RDONLY) = -1 ENOENT open("/usr/share/icons/default/cursors/col-resize", O_RDONLY) = -1 ENOENT open("/usr/share/pixmaps/default/cursors/col-resize", O_RDONLY) = -1 ENOENT open("/.icons/default/cursors/h_double_arrow", O_RDONLY) = -1 ENOENT open("/usr/share/icons/default/cursors/h_double_arrow", O_RDONLY) = -1 ENOENT open("/usr/share/pixmaps/default/cursors/h_double_arrow", O_RDONLY) = -1 ENOENT open("/.icons/default/cursors/left_ptr", O_RDONLY) = -1 ENOENT open("/usr/share/icons/default/cursors/left_ptr", O_RDONLY) = -1 ENOENT open("/usr/share/pixmaps/default/cursors/left_ptr", O_RDONLY) = -1 ENOENT (I've removed " (No such file or directory)" after ENOENT for readability) Olav, any idea why it doesn't try to access /usr/share/icons/Adwaita/? Do we've to explicitly set the icon theme?
*** Bug 17572 has been marked as a duplicate of this bug. ***
CC: (none) => tarazed25
For those who want to play: Go to your local mirror and uncompress stage2 in install/stage2/live: $ misc/mdkinst_stage2_tool --uncompress install/stage2 Creating install/stage2/live from install/stage2/mdkinst.sqfs Then you can add files to the install/stage2/live chroot. Then you can try with drakx-in-chroot: # misc/drakx-in-chroot . /tmp/test-root
I'll need to check. Either it doesn't have that as icon theme. Could also be that it is expecting this in a "gresource". A file containing other files basically.
Having just finished a test script to run an install-and-reboot test on ver 228 using openqa I have been finding similar problems from the opposite point of view. The easiest way to drive an openqa test is to use keyboard navigation. I can confirm that the mouse cursor comes and goes, while the field focus indicator is mostly there but not always. In the latest log-in screen, the only way to change the session type is with a mouse click! It does not respond to any keyboard strokes. I am commenting on my experiences here because it may be part of the same problem. No intention to hijack the bug report. If I should be opening a separate bug report please tell me and I will do so. The sometimes unpredictable form focus and mouse cursor are complicating preparation of openqa tests.
CC: (none) => vzawalin1
well if the regression happens at the same time, there're quite a lot chances it's the same... For the record, I even installed basesystem-minimal+gtk3.0+adwaita_theme in the installer chroot to no avail
commit ec82f61f29017ac28d203370df8bbb3b5a7003c6 Author: Thierry Vignaud <thierry.vignaud@...> Date: Fri Jan 29 07:06:47 2016 +0100 fix hidden cursor (mga#17470) broken in commit 4998a6d4c66c42d8522ddaf36cda5ccb9dd964e9 --- Commit Link: http://gitweb.mageia.org/software/drakx/commit/?id=ec82f61f29017ac28d203370df8bbb3b5a7003c6
Actually it was as side effect of Pascal's cleaning...
Status: ASSIGNED => RESOLVEDResolution: (none) => FIXED
commit 2161b7f17001075b4e4001b7164f6fd01a968d8e Author: Thomas Backlund <tmb@...> Date: Mon Jan 11 00:20:06 2016 +0200 psmouse is now modular, so add it to list_modules.pm (mga#17470) (cherry picked from commit 3ecd99eeabe9ac660bf181ab49c363c679540395) --- Commit Link: http://gitweb.mageia.org/software/drakx/commit/?id=2161b7f17001075b4e4001b7164f6fd01a968d8e