[Intel-gfx] Help-On frame Buffer

[Bruno Prémont]

Hi Sanjay,

On Mon, 13 December 2010 sanjay anvekar wrote:
> Linux Kernel: uname -a info
> Linux (none) 2.6.31.6 #1 SMP Tue Sep 7 16:47:45 IST 2010 i686 unknown

That's an oldish kernel (and probably quite buggy KMS related)!
Are you seeing issues with more recent ones preventing their use?
Otherwise it definitely is worth trying something recent like 2.6.36.x
or maybe even latest 2.6.37-rc if 2.6.36.x didn't help.

> Output of following command is attached in vbios.bin
>   echo 1 > /sys/class/drm/card0/device/rom
>  cat /sys/class/drm/card0/device/rom > vbios.bin
> 
> 
>   "Could you try disabling vesafb in kernel config?"
> I tried disabling vesafb and still getting the same display.
> dmesg output after disabling the vesafb is as follows.
> 
> [    1.891] ACPI: SSDT 3f7ce0c0 00CC (v0  PmRef  Cpu1Ist 0030 INTL 2011)
> [    1.892] ACPI: SSDT 3f7ce3d0 005 (v0  PmRef  Cpu1Cst 0030 INTL 2011)
> [    1.801] ACPI: CPU1 (power states: C1[C1] C2[C2] C3[C3])
> [    1.814] processor LNXCPU:0: registered as cooling_device1
> [    1.864] Non-volatile memory driver v1.3
> [    1.863] intel_rng: FWH not detected
> [    1.872] Linux agpgart interface v0.13
> [    1.876] agpgart-intel 00:0:0.0: Intel 95GME Chipset
> [    1.878] agpgart-intel 00:0:0.0: detected 73K stolen memory
> [    1.800] agpgart-intel 00:0:0.0: AGP aperture is 26M @ 0xd0000
> [    1.803] [drm] Initialized drm 1.1.0 2001
> [    1.806] i95 00:0:0.0: PCI INT A -> GSI 1 (level, low) -> IRQ 1
> [    1.806] i95 00:0:0.0: setting latency timer to 6
> [    2.222] i2c-adapter i2c-2: unable to read EDID block.
> [    2.222] i95 00:0:0.0: LVDS-1: no EDID data
> [    2.290] i2c-adapter i2c-3: unable to read EDID block.
> [    2.299] i95 00:0:0.0: LVDS-1: no EDID data
> [    2.204] i2c-adapter i2c-2: unable to read EDID block.
> [    2.203] i95 00:0:0.0: LVDS-1: no EDID data
> [    1.806] i95 00:0:0.0: PCI INT A -> GSI 1 (level, low) -> IRQ 1
> [    1.806] i95 00:0:0.0: setting latency timer to 6
> [    2.222] i2c-adapter i2c-2: unable to read EDID block.
> [    2.222] i95 00:0:0.0: LVDS-1: no EDID data
> [    2.290] i2c-adapter i2c-3: unable to read EDID block.
> [    2.299] i95 00:0:0.0: LVDS-1: no EDID data
> [    2.204] i2c-adapter i2c-2: unable to read EDID block.
> [    2.203] i95 00:0:0.0: LVDS-1: no EDID data
> [    2.577] [drm] LVDS-8: set mode 12x78 c
> [    2.777] Console: switching to colour frame buffer device 18x4
> [    2.777] Console: switching to colour frame buffer device 18x4
> [    2.769] [drm] fb0: inteldrmfb frame buffer device
> [    2.762] [drm] Initialized i95 1.6.0 2003 for 00:0:0.0 on minor 0
> [    2.762] Serial: 85/150 driver, 4 ports, IRQ sharing enabled
> [    2.902] serial85: ttyS0 at I/O 0x3f8 (irq = 4) is a 150A
> [    3.241] serial85: ttyS1 at I/O 0x2f8 (irq = 3) is a 150A
> [    3.481] serial85: ttyS2 at I/O 0x3e8 (irq = 1) is a 150A
> [    3.722] serial85: ttyS3 at I/O 0x2e8 (irq = 1) is a 150A
> [    3.724] 0:0: ttyS0 at I/O 0x3f8 (irq = 4) is a 150A
> [    3.735] 0:0: ttyS1 at I/O 0x2f8 (irq = 3) is a 150A
> [    3.734] 0:0: ttyS2 at I/O 0x3e8 (irq = 1) is a 150A
> [    3.744] 0:0a: ttyS3 at I/O 0x2e8 (irq = 1) is a 150A
> [    3.796] brd: module loaded

Chris, do you remember some odd-looking output like:
  'i95 00:0:0.0'
or
  'Intel 95GME Chipset'
in the above kernel log?

In addition the PCI slot looks a bit too much 00:0:0.0.

> When I tried with "grep /sys/module/i915/parameters/*" command system didn't 
> return anything and it was stuck there. I had to restart the system.

A litte '.' is missing in the grep command (my bad), thus if there were
less than 2 matches for the glob grep would be waiting for input.

grep command should have been:
  grep . /sys/module/i915/parameters/*

> Please suggest me further.
> 
> Thanks & Best Regards,
> Sanjay
> ________________________________
> From: Bruno Prémont <bonbons at linux-vserver.org>
> To: sanjay anvekar <sanjayanvekar at yahoo.com>
> Cc: intel-gfx at lists.freedesktop.org
> Sent: Fri, 10 December, 2010 1:06:05 PM
> Subject: Re: [Intel-gfx] Help-On frame Buffer
> 
> Hi Sanjay,
> 
> Note, please keep list CCed and avoid HTML e-mails.
> In addition, inlining (without line wrapping) the log snippled would
> have been preferred (easier to replying).
> 
> On Thu, 9 Dec 2010 07:43:18 sanjay anvekar wrote:
> > Hi Bruno,
> >     Thanks for your reply.
> > Attached is the ouput of  ("dmesg | grep -A 10 -B 10 drm") as "drm.txt".
> >
> >
> > [    1.836] ACPI: SSDT 3f7ce0c0 00CC (v0  PmRef  Cpu1Ist 0030 INTL 2011)
> > [    1.841] ACPI: SSDT 3f7ce3d0 005 (v0  PmRef  Cpu1Cst 0030 INTL 2011)
> > [    1.857] ACPI: CPU1 (power states: C1[C1] C2[C2] C3[C3])
> > [    1.851] processor LNXCPU:0: registered as cooling_device1
> > [    1.801] Non-volatile memory driver v1.3
> > [    1.814] intel_rng: FWH not detected
> > [    1.878] Linux agpgart interface v0.13
> > [    1.846] agpgart-intel 00:0:0.0: Intel 95GME Chipset
>                                             ^^^^^
> This device name looks suspicious (as do all the i95 on the following lines...
> 
> What exact kernel version are you using?
> 
> > [    1.815] agpgart-intel 00:0:0.0: detected 73K stolen memory
> > [    1.908] agpgart-intel 00:0:0.0: AGP aperture is 26M @ 0xd0000
> > [    1.970] [drm] Initialized drm 1.1.0 2001
> > [    1.932] i95 00:0:0.0: PCI INT A -> GSI 1 (level, low) -> IRQ 1
> > [    1.998] i95 00:0:0.0: setting latency timer to 6
> > [    2.217] i2c-adapter i2c-2: unable to read EDID block.
> > [    2.239] i95 00:0:0.0: LVDS-1: no EDID data
> > [    2.319] i2c-adapter i2c-3: unable to read EDID block.
> > [    2.336] i95 00:0:0.0: LVDS-1: no EDID data
> > [    2.357] i2c-adapter i2c-2: unable to read EDID block.
> > [    2.381] i95 00:0:0.0: LVDS-1: no EDID data
> 
> As there does not seem to be a working EDID around, driver will have to
> find LVDS display details somewhere in VBIOS blob. (which might fail,
> in which case dumping VBIOS will be needed for someone the have a look
> at it)
> 
> echo 1 > /sys/class/drm/card0/device/rom
> cat /sys/class/drm/card0/device/rom > vbios.bin
> 
> And attach the vbios. (Well hoping you don't get the same error as
> for the other sysfs files)
> 
> > [    2.362] fb: conflicting fb hw usage inteldrmfb vs VESA VGA - removing 
> >generic driver
> 
> Could you try disabling vesafb in kernel config?
> 
> > [    2.381] Console: switching to colour dummy device 8x2
> > [    2.381] Console: switching to colour dummy device 8x2
> > [    2.582] [drm] LVDS-8: set mode 12x78 c
> > [    2.794] Console: switching to colour frame buffer device 18x4
> > [    2.794] Console: switching to colour frame buffer device 18x4
> 
> Unless a specific character go eaten your console size definitely looks
> damn small!
> 
> > [    2.786] [drm] fb0: inteldrmfb frame buffer device
> > [    2.780] [drm] Initialized i95 1.6.0 2003 for 00:0:0.0 on minor 0
> > [    2.788] Serial: 85/150 driver, 4 ports, IRQ sharing enabled
> > [    3.021] serial85: ttyS0 at I/O 0x3f8 (irq = 4) is a 150A
> > [    3.261] serial85: ttyS1 at I/O 0x2f8 (irq = 3) is a 150A
> > [    3.501] serial85: ttyS2 at I/O 0x3e8 (irq = 1) is a 150A
> > [    3.741] serial85: ttyS3 at I/O 0x2e8 (irq = 1) is a 150A
> > [    3.742] 0:0: ttyS0 at I/O 0x3f8 (irq = 4) is a 150A
> > [    3.752] 0:0: ttyS1 at I/O 0x2f8 (irq = 3) is a 150A
> > [    3.751] 0:0: ttyS2 at I/O 0x3e8 (irq = 1) is a 150A
> > [    3.762] 0:0a: ttyS3 at I/O 0x2e8 (irq = 1) is a 150A
> > [    3.718] brd: module loaded
> >
> > Unfortunatley I could not get "sysfs" files ,it returned some error.
> 
> What error are you getting?
> 
> > Also 
> > 
> >    "grep /sys/module/i915/parameters/*" and  "uname -a"  returned nothing.
> 
> Weird...
> 
> > Also I am attaching my basic graphic display file where I amusing VESA frame 
> > buffer.
> 
> As long as framebuffer console is not properly setup it makes no sense to look
> at your program, the issue lying in i915 driver or data it can obtain from 
> vbios.
> 
> > Please suggest me something.
> > 
> > Thanks & Best Regards,
> > Sanjay
> > 
> > 
> > 
> > 
> > ________________________________
> > From: Bruno Prémont <bonbons at linux-vserver.org>
> > To: sanjay anvekar <sanjayanvekar at yahoo.com>
> > Cc: intel-gfx at lists.freedesktop.org
> > Sent: Mon, 6 December, 2010 6:35:41 PM
> > Subject: Re: [Intel-gfx] Help-On frame Buffer
> > 
> > Hi Sanjay,
> > 
> > Most useful would be the pertinent part of your kernel log
> > (dmesg | grep -A 10 -B 10 drm) as well as some of the content
> > of the following sysfs files (edid is binary):
> >   /sys/class/drm/card0/card0-*/{dpms,edid,enabled,modes,status}
> > 
> > Eventually also:
> >   grep /sys/module/i915/parameters/*
> > and
> >   uname -a
> > 
> > 
> > Bruno
> > 
> > 
> > 
> > On Mon, 6 Dec 2010 03:49:43 sanjay anvekar wrote:
> > > Hi Bruno,
> > >     Thanks for your reply. I have recompiled the Linux kernel with 
> > > CONFIG_DRM_I915
> > > enabled, with the updated kernel file my display is looking strange. i.e. I 
> > >have 
> > >
> > > vga=791
> > > option in my boot configuration file and with this actual graphics screen is 
> 
> 
> > > displayed only
> > > in quarter area of screen. Which is reapting twice, covering left and right 
> > > quarter of screen,
> > > rest screen is black(i.e.bottom half of the screen). 
> > > How can I solve this problem? Please help me. 
> > > With the latest Kernel file with CONFIG_DRM_I915 enabled, how do I set 
> > >different 
> > >
> > > screen resolutions. Please let me know.
> > > 
> > > Thanks & Best Regards,
> > > Sanjay
> > > 
> > > 
> > > 
> > > 
> > > ________________________________
> > > From: Bruno Prémont <bonbons at linux-vserver.org>
> > > To: sanjay anvekar <sanjayanvekar at yahoo.com>
> > > Cc: intel-gfx at lists.freedesktop.org
> > > Sent: Sat, 4 December, 2010 2:12:57 AM
> > > Subject: Re: [Intel-gfx] Help-On frame Buffer
> > > 
> > > Hi,
> > > 
> > > You should look at KMS (Kernel Mode Setting) with i915 driver that you
> > > find under DRI. (CONFIG_DRM_I915 + CONFIG_DRM_I915_KMS)
> > > KMS can also be manually enabled/disabled with i915.modeset kernel
> > > parameter when i915 is built-in.
> > > 
> > > VESA only knows a few standard predefined modes but not the wide
> > > resolutions currently seen. It's based on some BIOS code and can only
> > > set resolution at early system boot, does not support s2ram (S3) nor any
> > > acceleration.
> > > 
> > > Bruno
> > > 
> > > 
> > > 
> > > On Fri, 03 December 2010 sanjay anvekar <sanjayanvekar at yahoo.com> wrote:
> > > > Hi,
> > > > 
> > > >     I am new to Linux and frame buffer. I am using Intel 945GSE / ICH7M 
> > > > chipset Panel PC with integrated Intel 945GSE
> > > > display chipset. Different size and resolution LCD screens are interfaced 
> > > > to the hardware (it supports Dual Channel 24-bit LVDS).
> > > > I am using Linux VESA frame buffer for various graphic object 
> > > > implementation.I am able to get 1024x768 screen resolution using
> > > > VESA frame buffer, but I am not able to get the full screen resolution of 
> > > > 1366x768 pixels supported my LCD screen.
> > > > Currenlty for various resolution I am changing boot configuration file as 
> > > > below
> > > >     e.g. vga=791 for 1024x768.
> > > > 
> > > > My existing BIOS does not list the 1366x768 resolution. So I would like to 
> > > > know how I could get the full resolution using the exsiting  hardware.
> > > > Please help me.
> > > > 
> > > > Thanks & Best Regards,
> > > > Sanjay