#579 flashed Libreboot to ga-g41m-es2l motherboard and couple issues raised

Open
opened 4 months ago by mattic · 5 comments
mattic commented 4 months ago

Hello

i have several Libreboot machines and they behave mostly well - with ga-g41m-es2l i stumbled to following;

1) wonder why this "ga-g41m-es2l_usqwerty_txtmode.rom" Libreboot binary is "txtmode" because of following

2) my GNU/Linux systems are aschetic, i build a lot from source, and have no binary package installation system - Libreboot i've installed using binary packages but Linux kernels i've built to my aschetic taste since 1.2.n - with this motherboard i tried Linux kernel 3.18.132, 3.19.8 and 4.20.2 and they all have the same problem of flipping from text mode to framebuffer;

4.20.2 dmesg says;


[ 0.446451] [drm] Replacing VGA console driver [ 0.447177] Console: switching to colour dummy device 80x25 [ 0.447294] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [ 0.447299] [drm] Driver supports precise vblank timestamp query. [ 0.447310] resource sanity check: requesting [mem 0x000c0000-0x000dffff], which spans more than PCI Bus 0000:00 [mem 0x000c0000-0x000c3fff window] [ 0.447323] caller pci_map_rom+0x63/0x1a0 mapping multiple BARs [ 0.447716] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem [ 10.566522] [drm:drm_atomic_helper_wait_for_flip_done] ERROR [CRTC:39:pipe A] flip_done timed out [ 10.568782] [drm] RC6 disabled, disabling runtime PM support


3.18.120 dmesg says;


[ 0.099718] [drm] Memory usable by graphics device = 1024M [ 0.099773] [drm] Replacing VGA console driver [ 0.100468] Console: switching to colour dummy device 80x25 [ 0.101214] i915 0000:00:02.0: irq 26 for MSI/MSI-X [ 0.101225] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [ 0.101229] [drm] Driver supports precise vblank timestamp query. [ 0.101329] vgaarb: device changed decodes: PCI:0000:00:02.0,olddecodes=io+mem,decodes=io+mem:owns=io+mem [ 0.104632] [drm] GPU HANG: ecode -1:0x00000000, reason: Command parser error, iir 0x00008000, action: continue [ 0.104632] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 0.104632] [drm] Please file a new bug report on bugs.freedesktop.org against DRI -> DRM/Intel [ 0.104632] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 0.104632] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. [ 0.104632] [drm] GPU crash dump saved to /sys/class/drm/card0/error [ 0.104632] i915: render error detected, EIR: 0x00000010 [ 0.104632] i915: IPEIR: 0x00000000 [ 0.104632] i915: IPEHR: 0x00000000 [ 0.104632] i915: INSTDONE_0: 0xfffffffe [ 0.104632] i915: INSTDONE_1: 0xffffffff [ 0.104632] i915: INSTDONE_2: 0x00000000 [ 0.104632] i915: INSTDONE_3: 0x00000000 [ 0.104632] i915: INSTPS: 0x00000000 [ 0.104632] i915: ACTHD: 0x00000000 [ 0.104632] i915: page table error [ 0.104632] i915: PGTBL_ER: 0x00000003 [ 0.104632] [drm:i915_report_and_clear_eir] ERROR EIR stuck: 0x00000010, masking [ 0.105202] [drm] Initialized i915 1.6.0 20140905 for 0000:00:02.0 on minor 0


and further;


[ 2.173404] fbcon: inteldrmfb (fb0) is primary device [ 2.225019] ------------[ cut here ]------------ [ 2.225027] WARNING: CPU: 1 PID: 25 at drivers/gpu/drm/i915/intel_display.c:901 i9xx_crtc_disable+0x8a/0x6e0() [ 2.225029] vblank wait on pipe A timed out [ 2.225032] CPU: 1 PID: 25 Comm: kworker/u4:1 Not tainted 3.18.132 #2 [ 2.225033] Hardware name: GIGABYTE GA-G41M-ES2L/GA-G41M-ES2L, BIOS 41ebf41 09/07/2016 [ 2.225040] Workqueue: events_unbound async_run_entry_fn [ 2.225045] 0000000000000286 0000000000000000 ffffffff81509bdd 0000000000009393 [ 2.225047] ffff88013b30b908 0000000000000009 ffffffff8103d348 ffff88013b2a9000 [ 2.225050] ffff8800b7848800 ffff8800b7860000 0000000000000000 ffff8800b7848b28 [ 2.225051] Call Trace: [ 2.225057] [] ? dump_stack+0x48/0x5c [ 2.225062] [] ? warn_slowpath_common+0x68/0x80 [ 2.225065] [] ? warn_slowpath_fmt+0x45/0x50 [ 2.225069] [] ? intel_wait_for_vblank+0x196/0x1f0 [ 2.225073] [] ? i9xx_crtc_disable+0x8a/0x6e0 [ 2.225076] [] ? __intel_set_mode+0x7a2/0x15c0 [ 2.225080] [] ? intel_set_mode+0xd/0x30 [ 2.225084] [] ? intel_crtc_set_config+0x94d/0xe70 [ 2.225088] [] ? drm_mode_set_config_internal+0x5a/0xe0 [ 2.225092] [] ? restore_fbdev_mode+0xc3/0xe0 [ 2.225095] [] ? drm_fb_helper_restore_fbdev_mode_unlocked+0x1b/0x60 [ 2.225098] [] ? drm_fb_helper_set_par+0x1d/0x50 [ 2.225101] [] ? intel_fbdev_set_par+0x11/0x60 [ 2.225106] [] ? fbcon_init+0x56a/0x5f0 [ 2.225111] [] ? visual_init+0x80/0xe0 [ 2.225114] [] ? do_bind_con_driver+0xf9/0x2b0 [ 2.225117] [] ? do_take_over_console+0xcf/0x170 [ 2.225121] [] ? do_fbcon_takeover+0x5b/0xc0 [ 2.225124] [] ? notifier_call_chain+0x44/0x70 [ 2.225127] [] ? __blocking_notifier_call_chain+0x44/0x60 [ 2.225131] [] ? register_framebuffer+0x1de/0x310 [ 2.225134] [] ? drm_fb_helper_initial_config+0x265/0x3a0 [ 2.225138] [] ? async_run_entry_fn+0x2d/0xe0 [ 2.225141] [] ? process_one_work+0x124/0x310 [ 2.225145] [] ? worker_thread+0x114/0x450 [ 2.225148] [] ? process_one_work+0x310/0x310 [ 2.225151] [] ? kthread+0xbc/0xe0 [ 2.225154] [] ? __kthread_parkme+0xa0/0xa0 [ 2.225158] [] ? ret_from_fork+0x58/0x90 [ 2.225161] [] ? __kthread_parkme+0xa0/0xa0 [ 2.225162] ---[ end trace eeb8f65a4846f302 ]--- [ 2.235316] Console: switching to colour frame buffer device 210x65


system runs smoothly after boot though, and framebuffer seems ok

because of the hang system boot takes long

3) Libreboot also stays long rolling disk before anything appear on screen

note that i would like to use 3.n kernel because w 4.n kernel says when flipping to inteldrmfb;

[ 10.568782] [drm] RC6 disabled, disabling runtime PM support

which means exactly that = no runtime power management - no matter what kernel configuration options i use, or if i have Libreboot or not, as long as i use Intel chipset 945 or even 965 on any system i own... i know that something has changed a lot w Intel chipset code somewhere between 3.n and 4.n - and i would like you to confirm this symptom.

Thanks for your comments

/mc

Hello i have several Libreboot machines and they behave mostly well - with ga-g41m-es2l i stumbled to following; 1) wonder why this "ga-g41m-es2l_usqwerty_txtmode.rom" Libreboot binary is "txtmode" because of following 2) my GNU/Linux systems are aschetic, i build a lot from source, and have no binary package installation system - Libreboot i've installed using binary packages but Linux kernels i've built to my aschetic taste since 1.2.n - with this motherboard i tried Linux kernel 3.18.132, 3.19.8 and 4.20.2 and they all have the same problem of flipping from text mode to framebuffer; 4.20.2 dmesg says; ---------------------------------------- [ 0.446451] [drm] Replacing VGA console driver [ 0.447177] Console: switching to colour dummy device 80x25 [ 0.447294] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [ 0.447299] [drm] Driver supports precise vblank timestamp query. [ 0.447310] resource sanity check: requesting [mem 0x000c0000-0x000dffff], which spans more than PCI Bus 0000:00 [mem 0x000c0000-0x000c3fff window] [ 0.447323] caller pci_map_rom+0x63/0x1a0 mapping multiple BARs [ 0.447716] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem [ 10.566522] [drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:39:pipe A] flip_done timed out [ 10.568782] [drm] RC6 disabled, disabling runtime PM support ---------------------------------------- 3.18.120 dmesg says; ---------------------------------------- [ 0.099718] [drm] Memory usable by graphics device = 1024M [ 0.099773] [drm] Replacing VGA console driver [ 0.100468] Console: switching to colour dummy device 80x25 [ 0.101214] i915 0000:00:02.0: irq 26 for MSI/MSI-X [ 0.101225] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [ 0.101229] [drm] Driver supports precise vblank timestamp query. [ 0.101329] vgaarb: device changed decodes: PCI:0000:00:02.0,olddecodes=io+mem,decodes=io+mem:owns=io+mem [ 0.104632] [drm] GPU HANG: ecode -1:0x00000000, reason: Command parser error, iir 0x00008000, action: continue [ 0.104632] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 0.104632] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [ 0.104632] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 0.104632] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. [ 0.104632] [drm] GPU crash dump saved to /sys/class/drm/card0/error [ 0.104632] i915: render error detected, EIR: 0x00000010 [ 0.104632] i915: IPEIR: 0x00000000 [ 0.104632] i915: IPEHR: 0x00000000 [ 0.104632] i915: INSTDONE_0: 0xfffffffe [ 0.104632] i915: INSTDONE_1: 0xffffffff [ 0.104632] i915: INSTDONE_2: 0x00000000 [ 0.104632] i915: INSTDONE_3: 0x00000000 [ 0.104632] i915: INSTPS: 0x00000000 [ 0.104632] i915: ACTHD: 0x00000000 [ 0.104632] i915: page table error [ 0.104632] i915: PGTBL_ER: 0x00000003 [ 0.104632] [drm:i915_report_and_clear_eir] *ERROR* EIR stuck: 0x00000010, masking [ 0.105202] [drm] Initialized i915 1.6.0 20140905 for 0000:00:02.0 on minor 0 ---------------------------------------- and further; ---------------------------------------- [ 2.173404] fbcon: inteldrmfb (fb0) is primary device [ 2.225019] ------------[ cut here ]------------ [ 2.225027] WARNING: CPU: 1 PID: 25 at drivers/gpu/drm/i915/intel_display.c:901 i9xx_crtc_disable+0x8a/0x6e0() [ 2.225029] vblank wait on pipe A timed out [ 2.225032] CPU: 1 PID: 25 Comm: kworker/u4:1 Not tainted 3.18.132 #2 [ 2.225033] Hardware name: GIGABYTE GA-G41M-ES2L/GA-G41M-ES2L, BIOS 41ebf41 09/07/2016 [ 2.225040] Workqueue: events_unbound async_run_entry_fn [ 2.225045] 0000000000000286 0000000000000000 ffffffff81509bdd 0000000000009393 [ 2.225047] ffff88013b30b908 0000000000000009 ffffffff8103d348 ffff88013b2a9000 [ 2.225050] ffff8800b7848800 ffff8800b7860000 0000000000000000 ffff8800b7848b28 [ 2.225051] Call Trace: [ 2.225057] [<ffffffff81509bdd>] ? dump_stack+0x48/0x5c [ 2.225062] [<ffffffff8103d348>] ? warn_slowpath_common+0x68/0x80 [ 2.225065] [<ffffffff8103d3d5>] ? warn_slowpath_fmt+0x45/0x50 [ 2.225069] [<ffffffff812af4a6>] ? intel_wait_for_vblank+0x196/0x1f0 [ 2.225073] [<ffffffff812b9bda>] ? i9xx_crtc_disable+0x8a/0x6e0 [ 2.225076] [<ffffffff812b3b32>] ? __intel_set_mode+0x7a2/0x15c0 [ 2.225080] [<ffffffff812bc1bd>] ? intel_set_mode+0xd/0x30 [ 2.225084] [<ffffffff812bd14d>] ? intel_crtc_set_config+0x94d/0xe70 [ 2.225088] [<ffffffff81259caa>] ? drm_mode_set_config_internal+0x5a/0xe0 [ 2.225092] [<ffffffff81248f13>] ? restore_fbdev_mode+0xc3/0xe0 [ 2.225095] [<ffffffff8124b48b>] ? drm_fb_helper_restore_fbdev_mode_unlocked+0x1b/0x60 [ 2.225098] [<ffffffff8124b4ed>] ? drm_fb_helper_set_par+0x1d/0x50 [ 2.225101] [<ffffffff812c7f11>] ? intel_fbdev_set_par+0x11/0x60 [ 2.225106] [<ffffffff811dd4ca>] ? fbcon_init+0x56a/0x5f0 [ 2.225111] [<ffffffff81237b00>] ? visual_init+0x80/0xe0 [ 2.225114] [<ffffffff81239b59>] ? do_bind_con_driver+0xf9/0x2b0 [ 2.225117] [<ffffffff8123aa0f>] ? do_take_over_console+0xcf/0x170 [ 2.225121] [<ffffffff811dd5ab>] ? do_fbcon_takeover+0x5b/0xc0 [ 2.225124] [<ffffffff81055234>] ? notifier_call_chain+0x44/0x70 [ 2.225127] [<ffffffff81055584>] ? __blocking_notifier_call_chain+0x44/0x60 [ 2.225131] [<ffffffff811e56de>] ? register_framebuffer+0x1de/0x310 [ 2.225134] [<ffffffff8124b255>] ? drm_fb_helper_initial_config+0x265/0x3a0 [ 2.225138] [<ffffffff810566ad>] ? async_run_entry_fn+0x2d/0xe0 [ 2.225141] [<ffffffff8104f5a4>] ? process_one_work+0x124/0x310 [ 2.225145] [<ffffffff8104f8a4>] ? worker_thread+0x114/0x450 [ 2.225148] [<ffffffff8104f790>] ? process_one_work+0x310/0x310 [ 2.225151] [<ffffffff8105444c>] ? kthread+0xbc/0xe0 [ 2.225154] [<ffffffff81054390>] ? __kthread_parkme+0xa0/0xa0 [ 2.225158] [<ffffffff8150fc88>] ? ret_from_fork+0x58/0x90 [ 2.225161] [<ffffffff81054390>] ? __kthread_parkme+0xa0/0xa0 [ 2.225162] ---[ end trace eeb8f65a4846f302 ]--- [ 2.235316] Console: switching to colour frame buffer device 210x65 ---------------------------------------- system runs smoothly after boot though, and framebuffer seems ok because of the hang system boot takes long 3) Libreboot also stays long rolling disk before anything appear on screen note that i would like to use 3.n kernel because w 4.n kernel says when flipping to inteldrmfb; [ 10.568782] [drm] RC6 disabled, disabling runtime PM support which means exactly that = no runtime power management - no matter what kernel configuration options i use, or if i have Libreboot or not, as long as i use Intel chipset 945 or even 965 on any system i own... i know that something has changed a lot w Intel chipset code somewhere between 3.n and 4.n - and i would like you to confirm this symptom. Thanks for your comments /mc
Swift Geek commented 4 months ago
Collaborator

txtmode/vesa(corebootfb) have nothing to do with linux, just display mode in grub. Since memtest is only compatible with txtmode we need to provide both

txtmode/vesa(corebootfb) have nothing to do with linux, just display mode in grub. Since memtest is only compatible with txtmode we need to provide both
mattic commented 4 months ago
Poster

yes, i do understand that, and it is ok - did not find vesafb binary for this board though ( which would be nice as i once a year boot the thing and it shows me a nice image )

the other issues are of importance, mostly for the project

system takes 18 seconds before 'Loading operating system...' shows

and 48 seconds before it begins loading kernel

and 1 minute before framebuffer shows first text - and there are the errors mentioned earlier

( compared to my macbook2.1 w libreboot and similar GNU+Linux where booting to first text on screen takes about 2s and first Linux booting messages less than 8s )

yes, i do understand that, and it is ok - did not find vesafb binary for this board though ( which would be nice as i once a year boot the thing and it shows me a nice image ) the other issues are of importance, mostly for the project system takes 18 seconds before 'Loading operating system...' shows and 48 seconds before it begins loading kernel and 1 minute before framebuffer shows first text - and there are the errors mentioned earlier ( compared to my macbook2.1 w libreboot and similar GNU+Linux where booting to first text on screen takes about 2s and first Linux booting messages less than 8s )
mattic commented 4 months ago
Poster

on https://libreboot.org/docs/hardware/ga-g41m-es2l.html is a comment about ide drives 'IDE on the board is untested, but it might be possible to use a SATA HDD using an IDE SATA adapter. The SATA ports do work'

i tried and ide drives do not work w ide cable connected to ide connector of the board ( above comment is somewhat unclear ) - not that it matters now-a-days

on https://libreboot.org/docs/hardware/ga-g41m-es2l.html is a comment about ide drives 'IDE on the board is untested, but it might be possible to use a SATA HDD using an IDE SATA adapter. The SATA ports do work' i tried and ide drives do not work w ide cable connected to ide connector of the board ( above comment is somewhat unclear ) - not that it matters now-a-days
mattic commented 4 months ago
Poster

now i have tried all pci wlan nics i have and they don't work ( rt2500, rtl8185, ... ) = errors or just no connection

of course it is possible that all my cards are dead - but seems odd - anyway i try to get some more to test further ( all these have worked w many machines w similar system over the years )

i won't use cards that need proprietary firmware and thus the only usb nic i have just now is buffalo wli-u2-kg125s ( the only rndis usb wlan nic ) and it works as expected

now i have tried all pci wlan nics i have and they don't work ( rt2500, rtl8185, ... ) = errors or just no connection of course it is possible that all my cards are dead - but seems odd - anyway i try to get some more to test further ( all these have worked w many machines w similar system over the years ) i won't use cards that need proprietary firmware and thus the only usb nic i have just now is buffalo wli-u2-kg125s ( the only rndis usb wlan nic ) and it works as expected
Ghost commented 3 months ago

The "hang" issue at boot is caused by GRUB payload with those chipsets. (G41) The issue is still present even with latest GRUB (built from source). The seabios payload for example do not have this issue. (Linux kernel boots in less than 5 seconds) For now, it is still unsolved. Concerning PCI slot, Libreboot "latest" stable rom do not include some patches to enable these.* I'm using coreboot 4.9 and the PCI ports (including pcie-x16) works flawlessly. I recommend you to use coreboot with seabios payload if you want something functional.

*I only use pcie-x1 and pcie-x16 but I guess good old PCI should work.

The "hang" issue at boot is caused by GRUB payload with those chipsets. (G41) The issue is still present even with latest GRUB (built from source). The seabios payload for example do not have this issue. (Linux kernel boots in less than 5 seconds) For now, it is still unsolved. Concerning PCI slot, Libreboot "latest" stable rom do not include some patches to enable these.* I'm using coreboot 4.9 and the PCI ports (including pcie-x16) works flawlessly. I recommend you to use coreboot with seabios payload if you want something functional. *I only use pcie-x1 and pcie-x16 but I guess good old PCI should work.
Sign in to join this conversation.
Loading...
Cancel
Save
There is no content yet.