Author: blekot Date: Mon Apr 3 21:41:46 2006 GMT
Module: SOURCES Tag: LINUX_2_6
---- Log message:
- new patch from gentoo
- fb working on nvidia with nvidia binary packages
- swiching from X to console is working too
---- Files affected:
SOURCES:
linux-2.6-vesafb-tng.patch (NONE -> 1.1.2.1) (NEW)
---- Diffs:
================================================================
Index: SOURCES/linux-2.6-vesafb-tng.patch
diff -u /dev/null SOURCES/linux-2.6-vesafb-tng.patch:1.1.2.1
--- /dev/null Mon Apr 3 23:41:46 2006
+++ SOURCES/linux-2.6-vesafb-tng.patch Mon Apr 3 23:41:41 2006
@@ -0,0 +1,2980 @@
+diff --git a/Documentation/fb/vesafb.txt b/Documentation/fb/vesafb.txt
+index ee277dd..f768129 100644
+--- a/Documentation/fb/vesafb.txt
++++ b/Documentation/fb/vesafb.txt
+@@ -2,16 +2,18 @@
+ What is vesafb?
+ ===============
+
+-This is a generic driver for a graphic framebuffer on intel boxes.
++Vesafb is a generic framebuffer driver for x86 and x86_64 boxes.
+
+-The idea is simple: Turn on graphics mode at boot time with the help
+-of the BIOS, and use this as framebuffer device /dev/fb0, like the m68k
+-(and other) ports do.
+-
+-This means we decide at boot time whenever we want to run in text or
+-graphics mode. Switching mode later on (in protected mode) is
+-impossible; BIOS calls work in real mode only. VESA BIOS Extensions
+-Version 2.0 are required, because we need a linear frame buffer.
++VESA BIOS Extensions Version 2.0 are required, because we need access to
++a linear frame buffer. VBE 3.0 is required if you want to use modes with a
++higher (than the standard 60Hz) refresh rate.
++
++The VESA framebuffer driver comes in two flavors - the standard 'vesafb'
++and 'vesafb-tng'. Vesafb-tng is available only on 32-bit x86 due to the
++technology it uses (vm86). Vesafb-tng has more features than vesafb
++(adjusting the refresh rate on VBE3.0-compliant boards, switching the
++video mode without rebooting, selecting a mode by providing its
++modedb name, and more).
+
+ Advantages:
+
+@@ -29,26 +31,35 @@ Disadvantages:
+ How to use it?
+ ==============
+
+-Switching modes is done using the vga=... boot parameter. Read
+-Documentation/svga.txt for details.
+-
+-You should compile in both vgacon (for text mode) and vesafb (for
+-graphics mode). Which of them takes over the console depends on
+-whenever the specified mode is text or graphics.
+-
+-The graphic modes are NOT in the list which you get if you boot with
+-vga=ask and hit return. The mode you wish to use is derived from the
+-VESA mode number. Here are those VESA mode numbers:
++If you are running a 32-bit x86 system and you decide to use vesafb-tng,
++you can either compile the driver into the kernel or use it as a module.
++The graphics mode you want to use is in both cases specified using the
++standard modedb format.
++
++If your system doesn't support vm86 calls, things get a little more tricky.
++Since on such systems you can't do BIOS calls from protected mode in which
++kernel runs, you have to decide at boot time whenever you want to run in text
++or in graphics mode. Switching mode later on is impossible. Switching modes
++is done using the vga=... boot parameter. Read Documentation/svga.txt for
++details. Below is a more detailed description of what to do on systems using
++the standard vesafb driver.
++
++You should compile in both vgacon (for text mode) and vesafb (for graphics
++mode). Which of them takes over the console depends on whenever the
++specified mode is text or graphics.
++
++The graphic modes are NOT in the list which you get if you boot with vga=ask
++and hit return. The mode you wish to use is derived from the VESA mode number.
++Here are those VESA mode numbers:
+
+ | 640x480 800x600 1024x768 1280x1024
+ ----+-------------------------------------
+-256 | 0x101 0x103 0x105 0x107
+-32k | 0x110 0x113 0x116 0x119
+-64k | 0x111 0x114 0x117 0x11A
+-16M | 0x112 0x115 0x118 0x11B
++256 | 0x101 0x103 0x105 0x107
++32k | 0x110 0x113 0x116 0x119
++64k | 0x111 0x114 0x117 0x11A
++16M | 0x112 0x115 0x118 0x11B
+
+-The video mode number of the Linux kernel is the VESA mode number plus
+-0x200.
++The video mode number of the Linux kernel is the VESA mode number plus 0x200.
+
+ Linux_kernel_mode_number = VESA_mode_number + 0x200
+
+@@ -56,15 +67,15 @@ So the table for the Kernel mode numbers
+
+ | 640x480 800x600 1024x768 1280x1024
+ ----+-------------------------------------
+-256 | 0x301 0x303 0x305 0x307
+-32k | 0x310 0x313 0x316 0x319
+-64k | 0x311 0x314 0x317 0x31A
+-16M | 0x312 0x315 0x318 0x31B
+-
+-To enable one of those modes you have to specify "vga=ask" in the
+-lilo.conf file and rerun LILO. Then you can type in the desired
+-mode at the "vga=ask" prompt. For example if you like to use
+-1024x768x256 colors you have to say "305" at this prompt.
++256 | 0x301 0x303 0x305 0x307
++32k | 0x310 0x313 0x316 0x319
++64k | 0x311 0x314 0x317 0x31A
++16M | 0x312 0x315 0x318 0x31B
++
++To enable one of those modes you have to specify "vga=ask" in the lilo.conf
++file and rerun LILO. Then you can type in the desired mode at the "vga=ask"
++prompt. For example if you like to use 1024x768x256 colors you have to say
++"305" at this prompt.
+
+ If this does not work, this might be because your BIOS does not support
+ linear framebuffers or because it does not support this mode at all.
+@@ -72,11 +83,12 @@ Even if your board does, it might be the
+ Extensions v2.0 are required, 1.2 is NOT sufficient. You will get a
+ "bad mode number" message if something goes wrong.
+
+-1. Note: LILO cannot handle hex, for booting directly with
++1. Note: LILO cannot handle hex, for booting directly with
+ "vga=mode-number" you have to transform the numbers to decimal.
+ 2. Note: Some newer versions of LILO appear to work with those hex values,
+ if you set the 0x in front of the numbers.
+
++
+ X11
+ ===
+
+@@ -84,98 +96,164 @@ XF68_FBDev should work just fine, but it
+ another (accelerated) X-Server like XF86_SVGA might or might not work.
+ It depends on X-Server and graphics board.
+
+-The X-Server must restore the video mode correctly, else you end up
++The X-Server must restore the video mode correctly, or else you end up
+ with a broken console (and vesafb cannot do anything about this).
++With vesafb-tng chances are that the console will be restored properly
++even if the X server messes up the video mode.
+
+
+ Refresh rates
+ =============
+
+-There is no way to change the vesafb video mode and/or timings after
+-booting linux. If you are not happy with the 60 Hz refresh rate, you
+-have these options:
++With VBE3.0 compatible BIOSes and vesafb-tng it is possible to change
++the refresh rate either at boot time (by specifying the @<rr> part of
++the mode name) or later, using the fbset utility.
++
++If you want to use the default BIOS refresh rate while switching modes
++on a running system, set pixclock to 0.
++
++With VBE2.0 there is no way to change the mode timings after booting
++Linux. If you are not happy with the 60 Hz refresh rate, you have
++these options:
+
+- * configure and load the DOS-Tools for your the graphics board (if
+- available) and boot linux with loadlin.
++ * configure and load the DOS tools for your the graphics board (if
++ available) and boot Linux with loadlin.
+ * use a native driver (matroxfb/atyfb) instead if vesafb. If none
+ is available, write a new one!
+- * VBE 3.0 might work too. I have neither a gfx board with VBE 3.0
+- support nor the specs, so I have not checked this yet.
++ * use a BIOS editor to change the default refresh rate (such an
++ editor does exist at least for ATI Radeon BIOSes).
++ * if you're running a non-vm86 and VBE3.0-compatible system, you can
++ use a kernel patch (vesafb-rrc) to hard-code some mode timings in
++ the kernel and use these while setting the graphic mode at boot time.
++
++Note that there are some boards (nVidia 59**, 57** and newer models)
++claiming that their Video BIOS is VBE3.0 compliant, while ignoring the
++CRTC values provided by software such as vesafb-tng. You'll not be able
++to change the refresh rate if you're using one of these boards.
+
+
+ Configuration
+ =============
+
+-The VESA BIOS provides protected mode interface for changing
+-some parameters. vesafb can use it for palette changes and
+-to pan the display. It is turned off by default because it
+-seems not to work with some BIOS versions, but there are options
+-to turn it on.
+-
+-You can pass options to vesafb using "video=vesafb:option" on
+-the kernel command line. Multiple options should be separated
+-by comma, like this: "video=vesafb:ypan,invers"
+-
+-Accepted options:
+-
+-invers no comment...
+-
+-ypan enable display panning using the VESA protected mode
+- interface. The visible screen is just a window of the
+- video memory, console scrolling is done by changing the
+- start of the window.
+- pro: * scrolling (fullscreen) is fast, because there is
+- no need to copy around data.
+- * You'll get scrollback (the Shift-PgUp thing),
+- the video memory can be used as scrollback buffer
+- kontra: * scrolling only parts of the screen causes some
+- ugly flicker effects (boot logo flickers for
+- example).
+-
+-ywrap Same as ypan, but assumes your gfx board can wrap-around
+- the video memory (i.e. starts reading from top if it
+- reaches the end of video memory). Faster than ypan.
+-
+-redraw scroll by redrawing the affected part of the screen, this
+- is the safe (and slow) default.
+-
+-
+-vgapal Use the standard vga registers for palette changes.
+- This is the default.
+-pmipal Use the protected mode interface for palette changes.
+-
+-mtrr:n setup memory type range registers for the vesafb framebuffer
+- where n:
+- 0 - disabled (equivalent to nomtrr) (default)
+- 1 - uncachable
+- 2 - write-back
+- 3 - write-combining
+- 4 - write-through
++The VESA BIOS provides protected mode interface for changing some parameters.
++vesafb can use it for palette changes and to pan the display. It is turned
++off by default because it seems not to work with some BIOS versions, but
++there are options to turn it on.
++
++You can pass options to vesafb using "video=vesafb:option" on the kernel
++command line. Multiple options should be separated by comma, like this:
++"video=vesafb:ypan,[EMAIL PROTECTED]"
++
++Note that vesafb-tng still uses the "video=vesafb:option" format of the
++kernel command line video parameter. "video=vesafb-tng:xxx" is incorrect.
++
++Accepted options (both vesafb and vesafb-tng):
++
++ypan Enable display panning using the VESA protected mode interface
++ The visible screen is just a window of the video memory,
++ console scrolling is done by changing the start of the window.
++ pro: * scrolling (fullscreen) is fast, because there is
++ no need to copy around data.
++ * you'll get scrollback (the Shift-PgUp thing),
++ the video memory can be used as scrollback buffer
++ con: * scrolling only parts of the screen causes some
++ ugly flicker effects (boot logo flickers for
++ example).
++
++ywrap Same as ypan, but assumes your gfx board can wrap-around the video
++ memory (i.e. starts reading from top if it reaches the end of
++ video memory). Faster than ypan.
++
++redraw Scroll by redrawing the affected part of the screen, this is the
++ safe (and slow) default.
++
++vgapal Use the standard VGA registers for palette changes.
++ This is the default.
++
++pmipal Use the protected mode interface for palette changes.
++
++mtrr:n Setup memory type range registers for the vesafb framebuffer
++ where n:
++ 0 - disabled (equivalent to nomtrr) (default)
++ 1 - uncachable
++ 2 - write-back
++ 3 - write-combining
++ 4 - write-through
+
+- If you see the following in dmesg, choose the type that matches the
+- old one. In this example, use "mtrr:2".
++ If you see the following in dmesg, choose the type that matches
++ the old one. In this example, use "mtrr:2".
+ ...
+ mtrr: type mismatch for e0000000,8000000 old: write-back new: write-combining
+ ...
+
+-nomtrr disable mtrr
++nomtrr Do not use memory type range registers for vesafb.
+
+ vremap:n
+ remap 'n' MiB of video RAM. If 0 or not specified, remap memory
+- according to video mode. (2.5.66 patch/idea by Antonino Daplas
+- reversed to give override possibility (allocate more fb memory
+- than the kernel would) to 2.4 by [EMAIL PROTECTED])
++ according to video mode. (2.5.66 patch/idea by Antonino Daplas
++ reversed to give override possibility (allocate more fb memory
++ than the kernel would) to 2.4 by [EMAIL PROTECTED])
+
+ vtotal:n
+ if the video BIOS of your card incorrectly determines the total
+ amount of video RAM, use this option to override the BIOS (in MiB).
+
+-Have fun!
++Options accepted only by vesafb-tng:
+
+- Gerd
++<mode> The mode you want to set, in the standard modedb format. Refer to
++ modedb.txt for detailed description. If you specify a mode that is
++ not supported by your board's BIOS, vesafb will attempt to set a
++ similar mode. The list of supported modes can be found in
++ /proc/fbx/modes, where x is the framebuffer number (usually 0).
++ When vesafb is compiled as a module, the mode string should be
++ provided as a value of the parameter 'mode'.
++
++vbemode:x
++ Force the use of VBE mode x. The mode will only be set if it's
++ found in VBE-provided list of supported modes.
++ NOTE: The mode number 'x' should be specified in VESA mode number
++ notation, not the Linux kernel one (ie. 257 instead of 769).
++ HINT: If you use this option because normal <mode> parameter does
++ not work for you and you use a X server, you'll probably want to
++ set the 'nocrtc' option to ensure that the video mode is properly
++ restored after console <-> X switches.
++
++nocrtc Do not use CRTC timings while setting the graphic mode. This option
++ makes sence only with VBE3.0 compliant systems. Use it if you have
++ problems with the modes set in the standard way. Note that specifying
++ this option means the refresh rate will be ignored and will stay at
++ your BIOS default (60 Hz).
++
++noedid Do not try to fetch and use EDID-provided modes.
++
++noblank Disable hardware blanking.
++
++gtf Force the use of VESA's GTF (Generalized Timing Formula). Specifying
++ this will cause vesafb to skip it's internal modedb and EDID-modedb
++ and jump straight to the GTF part of the code (normally used only if
++ everything else failed). This can be useful if you want to get as
++ much as possible from you graphics board but your BIOS doesn't
++ support modes with refresh rates you require. Note that you may need
++ to specify the maxhf, maxvf and maxclk parameters if they are not
++ provided by EDID.
++
++Additionally, the following parameters may be provided. They all override the
++EDID-provided values and BIOS defaults. Refer to your monitor's specs to get
++the correct values for maxhf, maxvf and maxclk for your hardware.
++
++maxhf:n Maximum horizontal frequency (in kHz).
++maxvf:n Maximum vertical frequency (in Hz).
++maxclk:n Maximum pixel clock (in MHz).
++
++Have fun!
+
+ --
++Original document for the vesafb driver by
+ Gerd Knorr <[EMAIL PROTECTED]>
+
+-Minor (mostly typo) changes
+-by Nico Schmoigl <[EMAIL PROTECTED]>
++Minor (mostly typo) changes by
++Nico Schmoigl <[EMAIL PROTECTED]>
++
++Extended documentation for vm86, VBE3.0 and vesafb-tng by
++Michal Januszewski <[EMAIL PROTECTED]>
++
+diff --git a/arch/i386/boot/video.S b/arch/i386/boot/video.S
+index 2ac40c8..335d401 100644
+--- a/arch/i386/boot/video.S
++++ b/arch/i386/boot/video.S
+@@ -164,10 +164,12 @@ basret: ret
+ # parameters in the default 80x25 mode -- these are set directly,
+ # because some very obscure BIOSes supply insane values.
+ mode_params:
++#ifdef CONFIG_FB_VESA_STD
+ #ifdef CONFIG_VIDEO_SELECT
+ cmpb $0, graphic_mode
+ jnz mopar_gr
+ #endif
++#endif
+ movb $0x03, %ah # Read cursor position
+ xorb %bh, %bh
+ int $0x10
+@@ -200,6 +202,7 @@ mopar2: movb %al, %fs:(PARAM_VIDEO_LINES
+ ret
+
+ #ifdef CONFIG_VIDEO_SELECT
++#ifdef CONFIG_FB_VESA_STD
+ # Fetching of VESA frame buffer parameters
+ mopar_gr:
+ leaw modelist+1024, %di
+@@ -278,6 +281,7 @@ dac_done:
+ movw %es, %fs:(PARAM_VESAPM_SEG)
+ movw %di, %fs:(PARAM_VESAPM_OFF)
+ no_pm: ret
++#endif
+
+ # The video mode menu
+ mode_menu:
+@@ -492,10 +496,12 @@ mode_set:
+
+ cmpb $VIDEO_FIRST_V7>>8, %ah
+ jz setv7
+-
++
++#ifdef CONFIG_FB_VESA_STD
+ cmpb $VIDEO_FIRST_VESA>>8, %ah
+ jnc check_vesa
+-
++#endif
++
+ orb %ah, %ah
+ jz setmenu
+
+@@ -567,6 +573,7 @@ setr1: lodsw
+ movw -4(%si), %ax # Fetch mode ID
+ jmp _m_s
+
++#ifdef CONFIG_FB_VESA_STD
+ check_vesa:
+ leaw modelist+1024, %di
+ subb $VIDEO_FIRST_VESA>>8, %bh
+@@ -600,6 +607,7 @@ check_vesa:
+ ret
+
+ _setbad: jmp setbad # Ugly...
++#endif
+
+ # Recalculate vertical display end registers -- this fixes various
+ # inconsistencies of extended modes on many adapters. Called when
+diff --git a/drivers/video/Kconfig b/drivers/video/Kconfig
+index fdebd60..a1d6253 100644
+--- a/drivers/video/Kconfig
++++ b/drivers/video/Kconfig
+@@ -454,8 +454,8 @@ config FB_TGA
+ cards. Say Y if you have one of those.
+
+ config FB_VESA
+- bool "VESA VGA graphics support"
+- depends on (FB = y) && X86
++ tristate "VESA VGA graphics support"
++ depends on (FB = y) && (X86 || X86_64)
+ select FB_CFB_FILLRECT
+ select FB_CFB_COPYAREA
+ select FB_CFB_IMAGEBLIT
+@@ -465,6 +465,48 @@ config FB_VESA
+ You will get a boot time penguin logo at no additional cost. Please
+ read <file:Documentation/fb/vesafb.txt>. If unsure, say Y.
+
++choice
++ prompt "VESA driver type"
++ depends on FB_VESA
++ default FB_VESA_STD if X86_64
++ default FB_VESA_TNG if X86
++
++config FB_VESA_STD
++ bool "vesafb"
++ help
++ This is the frame buffer device driver for generic VESA 2.0
++ compliant graphic cards. The older VESA 1.2 cards are not supported.
++ You will get a boot time penguin logo at no additional cost. Please
++ read <file:Documentation/fb/vesafb.txt>. Choose this driver if you
++ are experiencing problems with vesafb-tng or if you own a 64-bit
system.
++
++ Note that this driver cannot be compiled as a module.
++
++config FB_VESA_TNG
++ bool "vesafb-tng"
++ depends on !X86_64
++ select FB_MODE_HELPERS
++ help
++ This is the frame buffer device driver for generic VESA 2.0
++ compliant graphic cards. It is capable of taking advantage of
++ VBE 3.0 features. With this driver you will be able to adjust
++ the refresh rate (VBE 3.0 compliant boards only) and change
++ the graphic mode on-the-fly.
++
++ You will also get a boot time penguin logo at no additional cost.
Please
++ read <file:Documentation/fb/vesafb.txt>.
++
++endchoice
++
++config FB_VESA_DEFAULT_MODE
++ string "VESA default mode"
++ depends on FB_VESA_TNG
++ default "[EMAIL PROTECTED]"
++ help
++ This option is used to determine the default mode vesafb is
++ supposed to switch to in case no mode is provided as a kernel
++ command line parameter.
++
+ config VIDEO_SELECT
+ bool
+ depends on FB_VESA
+diff --git a/drivers/video/Makefile b/drivers/video/Makefile
+index aa434e7..247cfb3 100644
+--- a/drivers/video/Makefile
++++ b/drivers/video/Makefile
+@@ -96,7 +96,11 @@ obj-$(CONFIG_FB_IMX) += imx
+ obj-$(CONFIG_FB_S3C2410) += s3c2410fb.o
+
+ # Platform or fallback drivers go here
+-obj-$(CONFIG_FB_VESA) += vesafb.o
++ifeq ($(CONFIG_FB_VESA_STD),y)
++ obj-y += vesafb.o
++else
++ obj-$(CONFIG_FB_VESA) += vesafb-thread.o vesafb-tng.o
++endif
+ obj-$(CONFIG_FB_VGA16) += vga16fb.o vgastate.o
+ obj-$(CONFIG_FB_OF) += offb.o
+
+diff --git a/drivers/video/fbmem.c b/drivers/video/fbmem.c
+index 996c7b5..f48eee0 100644
+--- a/drivers/video/fbmem.c
++++ b/drivers/video/fbmem.c
+@@ -1430,6 +1430,7 @@ fbmem_init(void)
+ printk(KERN_WARNING "Unable to create fb class; errno = %ld\n",
PTR_ERR(fb_class));
+ fb_class = NULL;
+ }
++
+ return 0;
+ }
+
+diff --git a/drivers/video/modedb.c b/drivers/video/modedb.c
+index 1da2f84..6755cb2 100644
+--- a/drivers/video/modedb.c
++++ b/drivers/video/modedb.c
+@@ -667,6 +667,7 @@ void fb_var_to_videomode(struct fb_video
+ {
+ u32 pixclock, hfreq, htotal, vtotal;
+
++ mode->refresh = 0;
+ mode->name = NULL;
+ mode->xres = var->xres;
+ mode->yres = var->yres;
+diff --git a/drivers/video/vesafb-thread.c b/drivers/video/vesafb-thread.c
+new file mode 100644
+index 0000000..249d794
+--- /dev/null
++++ b/drivers/video/vesafb-thread.c
+@@ -0,0 +1,722 @@
++/*
++ * Framebuffer driver for VBE 2.0+ compliant graphic boards.
++ * Kernel thread and vm86 routines.
++ *
++ * (c) 2004-2005 Michal Januszewski <[EMAIL PROTECTED]>
++ *
++ */
++
++#include <linux/config.h>
++#include <linux/slab.h>
++#include <linux/workqueue.h>
++#include <linux/completion.h>
++#include <linux/module.h>
++#include <linux/kernel.h>
++#include <linux/errno.h>
++#include <linux/mm.h>
++#include <linux/delay.h>
++#include <linux/signal.h>
++#include <linux/suspend.h>
++#include <linux/unistd.h>
++#include <video/vesa.h>
++#include <video/edid.h>
++#include <asm/mman.h>
++#include <asm/page.h>
++#include <asm/vm86.h>
++#include <asm/thread_info.h>
++#include <asm/uaccess.h>
++#include <asm/mmu_context.h>
++#include "edid.h"
++
++#ifdef MODULE
++int errno;
++#endif
++
++static DECLARE_COMPLETION(vesafb_th_completion);
++static DECLARE_MUTEX(vesafb_task_list_sem);
++static LIST_HEAD(vesafb_task_list);
++static DECLARE_WAIT_QUEUE_HEAD(vesafb_wait);
++
++static struct vm86_struct vm86;
++static int vesafb_pid = 0;
++
++_syscall3(int,ioperm,unsigned long, a, unsigned long, b, unsigned long, c);
++_syscall1(int,vm86old,struct vm86_struct __user*, v86);
++
++#define DEFAULT_VM86_FLAGS (IF_MASK | IOPL_MASK)
++#define VM86_PUSHW(x) \
++do { \
++ vm86.regs.esp -= 2; \
++ *(u16*)(STACK_ADDR + vm86.regs.esp) = x; \
++} while(0);
++
++/* Stack, the return code and buffers will be put into
++ * one contiguous memory chunk:
++ *
++ * [ STACK | RET_CODE | BUFFER ]
++ *
++ * We only need a buffer that is ca. 0x2000 bytes in size.
++ * Some video BIOSes (sis6326) try to store data somewhere
++ * in 0x7000-0x7fff, so we zeromap more memory to be safe.
++ */
++#define IVTBDA_SIZE PAGE_SIZE
++#define RET_CODE_SIZE 0x0010
++#define STACK_SIZE 0x0500
++#define BUFFER_SIZE 0x10000
<<Diff was trimmed, longer than 597 lines>>
_______________________________________________
pld-cvs-commit mailing list
[email protected]
http://lists.pld-linux.org/mailman/listinfo/pld-cvs-commit
