Re: [Intel-gfx] i810 forgets configured rows columns on ttys on startx shutdown
On 2013-10-29 19:19 (GMT-0400) Felix Miata composed: On 2013-10-29 13:54 (GMT+0100) Daniel Vetter composed: Userspace modesetting is known to be racy and broken, so people don't really want to waste time on digging into bugs which are most likely flukes. But if you can bisect this issue to a specific patch a fix should be possible. Again I got no clear answer to the question asked, so I guess now it's fair to assume this must be the right place for discussion after all. I've installed openSUSE 12.3. Its 3.7.10-1.16-default and server 1.13.2 behave nicely. Beyond installing Mageia 3 to get 3.8.13 and 1.13.4, and as one who does not build software even via proxy, I'm open to suggestions how to get kernels and/or other xorg versions needed to perform bisection. This problem is differently present with Mageia 3's desktop kernel 3.8.13.4. The ttys are left usable on startx exit, but the exit log X prints to screen is scrambled from mismatch between character size and line length. I don't know whether to try older kernels, or newer, even if I knew where to find any. -- The wise are known for their understanding, and pleasant words are persuasive. Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
[Intel-gfx] i810 forgets configured rows columns on ttys on startx shutdown
Embedded Dell motherboard/video BIOS has very limited VESA support!!! (F00:80x25=VGA, F01:80x50=VGA, F02:80x43=VGA, F03:80x28=VGA, F05:80x30=VGA, F07:80x60=VGA, 309:132x25=VESA, 30A:132x43=VESA, 30B:132x50=VESA, 30C:132x60=VESA, (8)120:132x25=BIOS, (8)121:132x43=BIOS, (8)122:132x50=BIOS, (8)123:132x60=BIOS) Pre-KMS it used to be that common vga= modes were useless, so I took to using 0x121 to escape from 80x25 screens. Using 0x8121 required specifying an appropriate font via config file. Even after KMS began, vga=0x8121 continued to work, and still does at least through kernel 3.4.47. Now with kernel 3.11.x (openSUSE 13.1) and 3.12.rc6 (Mageia 4): boot: ttys OK, whatever font, rows and columns kernel determines unless specified otherwise on cmdline via vga=. video= and resolution= on cmdline have no effect. Vga= still works too if font is correctly specified by config file, but only for 0x12# BIOS modes. After startx is started, before being stopped, ttys remain functional. Once startx is stopped, ttys may or may not remain functional. Sometimes after startx has been run exactly once since booting they continue to be useful. Other times, the rows and columns are no longer remembered, switching to fonts appropriate to an 80x25 screen, but outputting more than 80 columns and 25 rows, scrambling the output in part via line breaks absent when needed at column 80. -- The wise are known for their understanding, and pleasant words are persuasive. Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] i810 forgets configured rows columns on ttys on startx shutdown
On Tue, Oct 29, 2013 at 9:37 AM, Felix Miata mrma...@earthlink.net wrote: Pre-KMS it used to be that common vga= modes were useless, so I took to using 0x121 to escape from 80x25 screens. Using 0x8121 required specifying an appropriate font via config file. Even after KMS began, vga=0x8121 continued to work, and still does at least through kernel 3.4.47. This is still pre-kms, there's no kms driver for i810 chipsets. So whatever bug there is might be just bad luck or the userspace X driver not restoring stuff correctly. But I'm pretty sure it's not a kernel bug, at least no in the i810 driver. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] i810 forgets configured rows columns on ttys on startx shutdown
On 2013-10-29 09:51 (GMT+0100) Daniel Vetter composed: Felix Miata wrote: Pre-KMS it used to be that common vga= modes were useless, so I took to using 0x121 to escape from 80x25 screens. Using 0x8121 required specifying an appropriate font via config file. Even after KMS began, vga=0x8121 continued to work, and still does at least through kernel 3.4.47. This is still pre-kms, there's no kms driver for i810 chipsets. So whatever bug there is might be just bad luck or the userspace X driver not restoring stuff correctly. But I'm pretty sure it's not a kernel bug, at least no in the i810 driver. I was hoping for a (what appears to be a not our bug) response that was more helpful WRT whose bug it likely is. To me it's clearly a bug report that either can be found or needs to be made at bugs.freedesktop.org, but in what component? https://bugs.freedesktop.org/describecomponents.cgi?product=xorg has a long list, among which there is only one that includes the string intel, Driver/intel, which is what this list seems to be about. Does it look like Server/General, Server/DDX/Xorg, or something else? As it manifest in more than one distro, it surely isn't likely to be downstream. -- The wise are known for their understanding, and pleasant words are persuasive. Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] i810 forgets configured rows columns on ttys on startx shutdown
On Tue, Oct 29, 2013 at 08:00:20AM -0400, Felix Miata wrote: On 2013-10-29 09:51 (GMT+0100) Daniel Vetter composed: Felix Miata wrote: Pre-KMS it used to be that common vga= modes were useless, so I took to using 0x121 to escape from 80x25 screens. Using 0x8121 required specifying an appropriate font via config file. Even after KMS began, vga=0x8121 continued to work, and still does at least through kernel 3.4.47. This is still pre-kms, there's no kms driver for i810 chipsets. So whatever bug there is might be just bad luck or the userspace X driver not restoring stuff correctly. But I'm pretty sure it's not a kernel bug, at least no in the i810 driver. I was hoping for a (what appears to be a not our bug) response that was more helpful WRT whose bug it likely is. To me it's clearly a bug report that either can be found or needs to be made at bugs.freedesktop.org, but in what component? https://bugs.freedesktop.org/describecomponents.cgi?product=xorg has a long list, among which there is only one that includes the string intel, Driver/intel, which is what this list seems to be about. Does it look like Server/General, Server/DDX/Xorg, or something else? As it manifest in more than one distro, it surely isn't likely to be downstream. Userspace modesetting is known to be racy and broken, so people don't really want to waste time on digging into bugs which are most likely flukes. But if you can bisect this issue to a specific patch a fix should be possible. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] i810 forgets configured rows columns on ttys on startx shutdown
On 2013-10-29 13:54 (GMT+0100) Daniel Vetter composed: Felix Miata wrote: On 2013-10-29 09:51 (GMT+0100) Daniel Vetter composed: Felix Miata wrote: Pre-KMS it used to be that common vga= modes were useless, so I took to using 0x121 to escape from 80x25 screens. Using 0x8121 required specifying an appropriate font via config file. Even after KMS began, vga=0x8121 continued to work, and still does at least through kernel 3.4.47. This is still pre-kms, there's no kms driver for i810 chipsets. So whatever bug there is might be just bad luck or the userspace X driver not restoring stuff correctly. But I'm pretty sure it's not a kernel bug, at least no in the i810 driver. I was hoping for a (what appears to be a not our bug) response that was more helpful WRT whose bug it likely is. To me it's clearly a bug report that either can be found or needs to be made at bugs.freedesktop.org, but in what component? https://bugs.freedesktop.org/describecomponents.cgi?product=xorg has a long list, among which there is only one that includes the string intel, Driver/intel, which is what this list seems to be about. Does it look like Server/General, Server/DDX/Xorg, or something else? As it manifest in more than one distro, it surely isn't likely to be downstream. Userspace modesetting is known to be racy and broken, so people don't really want to waste time on digging into bugs which are most likely flukes. But if you can bisect this issue to a specific patch a fix should be possible. Again I got no clear answer to the question asked, so I guess now it's fair to assume this must be the right place for discussion after all. I've installed openSUSE 12.3. Its 3.7.10-1.16-default and server 1.13.2 behave nicely. Beyond installing Mageia 3 to get 3.8.13 and 1.13.4, and as one who does not build software even via proxy, I'm open to suggestions how to get kernels and/or other xorg versions needed to perform bisection. -- The wise are known for their understanding, and pleasant words are persuasive. Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
