Skip to contentSkip to conversation

vt: add ioctl commands to /dev/vcsaN to get/put the current palette of the given tty

PATCH In review Cedric Roux 2 participants 5 messages Mar 4, 2009 – May 14, 2009

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 vt: add ioctl commands to /dev/vcsaN to get/put the current palette of the given tty

From: Cedric Roux <[email protected]>

A ioctl interface and two ioctl commands added to /dev/vcsaN to get/put the current palette of the given tty.

---
This patch exists because there is no way to get the current installed palette of a given tty. The PIO_CMAP and GIO_CMAP in vt_ioctl.c:vt_ioctl play with the global default colormap. And /dev/vcsaN don't dump the palette through their read interface. And since a user may change colors of a given tty by escape sequences, one should be able to retrieve the palette through /dev/vcsaN, leading to this patch.

Based on remarks from Andrew Morton:
- the return value in case of error should now be OK
+ ENOTTY for /dev/vcs since the ioctl is not defined for /dev/vcs but only for /dev/vcsa
+ ENXIO if the console is not allocated as in other code that deal with consoles
+ EFAULT if userland buffers are not OK
+ ENOIOCTLCMD if we are given a wrong ioctl command
- registering as .unlocked_ioctl instead of .ioctl

Some points I am not fully confident with:
- vt.c:set_colormap is now used in vc_screen.c so cannot be static anymore.
The goal of this patch is to allow read access to the palette, I can remove the PIO_CMAP case and make set_colormap static again. This case is there because /dev/vcsaN is read/write so the access to the palette should also be read/write. (Or I can do it differently, no problem, just tell me how.)
- the palette is returned as it is stored in the kernel, without taking into account color_table. So if a userland program gets an attribute for a given character, it won't match with the palette information. Say the user reads 0xXY through /dev/vcsa and takes Y as an index in the palette, it will return a bad color. One should do palette[color_table[Y]] to get the real color.
Two "solutions" to this "problem" (it might be an non-issue):
+ export the palette by indexing with color_table:
I don't like it.
+ let the userland program deal with that:
the program may open a tty, send escape sequences to set the current color, write a character at a given position (through the tty, not /dev/vcs) and read back the attribute through /dev/vcsa this time. She will see what attribute corresponds to what color she sets by using escape sequences.
If the color_table is non-mutable, all this is a non-issue. Userland programs just store the color_table and live with that. I've seen no code that modifies the color_table, so I bet it's non-mutable. (May it change in the future?)
- lock_kernel() is called. Maybe it's not necessary? I'm not confident enough. You decide.
- release_console_sem() is called before calling copy_from/to_user as is done in vcs_write/vcs_read, once again I am not confident with that.

drivers/char/vc_screen.c +66 -0
diff -uprN -X linux-2.6.28-vanilla/Documentation/dontdiff linux-2.6.28-vanilla/drivers/char/vc_screen.c linux-2.6.28/drivers/char/vc_screen.c
@@ -19,6 +19,8 @@
  * [email protected] - modified not to send characters to wrong console
  *	 - fixed some fatal off-by-one bugs (0-- no longer == -1 -> looping and looping and looping...)
  *	 - making it shorter - scr_readw are macros which expand in PRETTY long code
+ *
+ * Colormap put/get for /dev/vcsaN, Cedric Roux <[email protected]>, March 2009.
  */
 
 #include <linux/kernel.h>
@@ -470,11 +472,75 @@ vcs_open(struct inode *inode, struct fil
 	return ret;
 }
 
+static long
+vcs_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
+{
+	unsigned int currcons = iminor(file->f_path.dentry->d_inode);
+	struct vc_data *vc;
+	unsigned char pal[16*3];
+	long ret = 0;
+
+	/* access only defined for /dev/vcsaN, not /dev/vcsN */
+	if (currcons < 128)
+		return -ENOTTY;
+
+	lock_kernel();
+
+	acquire_console_sem();
+
+	currcons &= 127;
+	if (currcons == 0)
+		currcons = fg_console;
+	else
+		currcons--;
+	if (!vc_cons_allocated(currcons)) {
+		ret = -ENXIO;
+		goto unlock_out;
+	}
+	vc = vc_cons[currcons].d;
+
+	switch (cmd) {
+	case PIO_CMAP:
+		release_console_sem();
+		ret = copy_from_user(pal, (void __user *)arg, 16*3);
+		acquire_console_sem();
+
+		if (ret) {
+			ret = -EFAULT;
+			goto unlock_out;
+		}
+		memcpy(vc->vc_palette, pal, 16*3);
+		set_palette(vc);
+		break;
+	case GIO_CMAP:
+		release_console_sem();
+		ret = copy_to_user((void __user *)arg, vc->vc_palette, 16*3);
+		acquire_console_sem();
+
+		if (ret) {
+			ret = -EFAULT;
+			goto unlock_out;
+		}
+		break;
+	default:
+		ret = -ENOIOCTLCMD;
+		break;
+	}
+
+unlock_out:
+	release_console_sem();
+
+	unlock_kernel();
+
+	return ret;
+}
+
 static const struct file_operations vcs_fops = {
 	.llseek		= vcs_lseek,
 	.read		= vcs_read,
 	.write		= vcs_write,
 	.open		= vcs_open,
+	.unlocked_ioctl	= vcs_ioctl,
 };
 
 static struct class *vc_class;
drivers/char/vt.c +1 -2
diff -uprN -X linux-2.6.28-vanilla/Documentation/dontdiff linux-2.6.28-vanilla/drivers/char/vt.c linux-2.6.28/drivers/char/vt.c
@@ -156,7 +156,6 @@ static void set_cursor(struct vc_data *v
 static void hide_cursor(struct vc_data *vc);
 static void console_callback(struct work_struct *ignored);
 static void blank_screen_t(unsigned long dummy);
-static void set_palette(struct vc_data *vc);
 
 static int printable;		/* Is console ready for printing? */
 int default_utf8 = true;
@@ -3756,7 +3755,7 @@ void poke_blanked_console(void)
  *	Palettes
  */
 
-static void set_palette(struct vc_data *vc)
+void set_palette(struct vc_data *vc)
 {
 	WARN_CONSOLE_UNLOCKED();
 
include/linux/vt_kern.h +1 -0
diff -uprN -X linux-2.6.28-vanilla/Documentation/dontdiff linux-2.6.28-vanilla/include/linux/vt_kern.h linux-2.6.28/include/linux/vt_kern.h
@@ -97,6 +97,7 @@ void reset_vc(struct vc_data *vc);
 extern int unbind_con_driver(const struct consw *csw, int first, int last,
 			     int deflt);
 int vty_init(const struct file_operations *console_fops);
+void set_palette(struct vc_data *vc);
 
 /*
  * vc_screen.c shares this temporary buffer with the console write code so that

--
To unsubscribe from this list: send the line "unsubscribe linux-api" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#1 lore raw
Quoted 9 lines · On Wed, 4 Mar 2009 09:58:48 +0100 (CET) Cedric Roux…

From: Cedric Roux <[email protected]>

A ioctl interface and two ioctl commands added to /dev/vcsaN to get/put the current palette of the given tty.

...

- lock_kernel() is called. Maybe it's not necessary? I'm not confident enough. You decide.

I wouldn't have a clue what it's protecting in there. But it seems sane to put it in - once someone gets down and de-lock_kernel()s that code, they can work out what it should be replaced by.

--
To unsubscribe from this list: send the line "unsubscribe linux-api" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#2 lore raw

PATCH 1/1 vt: add ioctl commands to /dev/vcsaN to get/put the current palette of the given tty

In reply to Cedric Roux

From: Cedric Roux <[email protected]>

A ioctl interface and two ioctl commands added to /dev/vcsaN to get/put the current palette of the given tty.

---
This patch exists because there is no way to get the current installed palette of a given tty. The PIO_CMAP and GIO_CMAP in vt_ioctl.c:vt_ioctl play with the global default colormap. And /dev/vcsaN don't dump the palette through their read interface. And since a user may change colors of a given tty by escape sequences, one should be able to retrieve the palette through /dev/vcsaN, leading to this patch.

Some points I am not fully confident with:
- vt.c:set_colormap is now used in vc_screen.c so cannot be static anymore.
The goal of this patch is to allow read access to the palette, I can remove the PIO_CMAP case and make set_colormap static again. This case is there because /dev/vcsaN is read/write so the access to the palette should also be read/write. (Or I can do it differently, no problem, just tell me how.)
- the palette is returned as it is stored in the kernel, without taking into account color_table. So if a userland program gets an attribute for a given character, it won't match with the palette information. Say the user reads 0xXY through /dev/vcsa and takes Y as an index in the palette, it will return a bad color. One should do palette[color_table[Y]] to get the real color.
Two "solutions" to this "problem" (it might be an non-issue):
+ export the palette by indexing with color_table:
I don't like it.
+ let the userland program deal with that:
the program may open a tty, send escape sequences to set the current color, write a character at a given position (through the tty, not /dev/vcs) and read back the attribute through /dev/vcsa this time. She will see what attribute corresponds to what color she sets by using escape sequences.
If the color_table is non-mutable, all this is a non-issue. Userland programs just store the color_table and live with that. I've seen no code that modifies the color_table, so I bet it's non-mutable. (May it change in the future?)
- lock_kernel() is called. Maybe it's not necessary? I'm not confident enough. You decide.

drivers/char/vc_screen.c +58 -0
diff -uprN -X linux-2.6.29.1/Documentation/dontdiff linux-2.6.29.1-vanilla/drivers/char/vc_screen.c linux-2.6.29.1/drivers/char/vc_screen.c
@@ -19,6 +19,8 @@
  * [email protected] - modified not to send characters to wrong console
  *	 - fixed some fatal off-by-one bugs (0-- no longer == -1 -> looping and looping and looping...)
  *	 - making it shorter - scr_readw are macros which expand in PRETTY long code
+ *
+ * Colormap put/get for /dev/vcsaN, Cedric Roux <[email protected]>, March 2009.
  */
 
 #include <linux/kernel.h>
@@ -470,11 +472,67 @@ vcs_open(struct inode *inode, struct fil
 	return ret;
 }
 
+static long
+vcs_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
+{
+	unsigned int currcons = iminor(file->f_path.dentry->d_inode);
+	struct vc_data *vc;
+	unsigned char pal[16*3];
+	long ret = 0;
+
+	/* access only defined for /dev/vcsaN, not /dev/vcsN */
+	if (currcons < 128)
+		return -ENOTTY;
+
+	lock_kernel();
+
+	acquire_console_sem();
+
+	currcons &= 127;
+	if (currcons == 0)
+		currcons = fg_console;
+	else
+		currcons--;
+	if (!vc_cons_allocated(currcons)) {
+		ret = -ENXIO;
+		goto unlock_out;
+	}
+	vc = vc_cons[currcons].d;
+
+	switch (cmd) {
+	case PIO_CMAP:
+		if (copy_from_user(pal, (void __user *)arg, 16*3)) {
+			ret = -EFAULT;
+			goto unlock_out;
+		}
+		memcpy(vc->vc_palette, pal, 16*3);
+		set_palette(vc);
+		break;
+	case GIO_CMAP:
+		if (copy_to_user((void __user *)arg, vc->vc_palette, 16*3)) {
+			ret = -EFAULT;
+			goto unlock_out;
+		}
+		break;
+	default:
+		ret = -ENOIOCTLCMD;
+		break;
+	}
+
+unlock_out:
+	release_console_sem();
+
+	unlock_kernel();
+
+	return ret;
+}
+
 static const struct file_operations vcs_fops = {
 	.llseek		= vcs_lseek,
 	.read		= vcs_read,
 	.write		= vcs_write,
 	.open		= vcs_open,
+	.unlocked_ioctl	= vcs_ioctl,
 };
 
 static struct class *vc_class;
drivers/char/vt.c +1 -2
diff -uprN -X linux-2.6.29.1/Documentation/dontdiff linux-2.6.29.1-vanilla/drivers/char/vt.c linux-2.6.29.1/drivers/char/vt.c
@@ -156,7 +156,6 @@ static void set_cursor(struct vc_data *v
 static void hide_cursor(struct vc_data *vc);
 static void console_callback(struct work_struct *ignored);
 static void blank_screen_t(unsigned long dummy);
-static void set_palette(struct vc_data *vc);
 
 static int printable;		/* Is console ready for printing? */
 int default_utf8 = true;
@@ -3753,7 +3752,7 @@ void poke_blanked_console(void)
  *	Palettes
  */
 
-static void set_palette(struct vc_data *vc)
+void set_palette(struct vc_data *vc)
 {
 	WARN_CONSOLE_UNLOCKED();
 
include/linux/vt_kern.h +1 -0
diff -uprN -X linux-2.6.29.1/Documentation/dontdiff linux-2.6.29.1-vanilla/include/linux/vt_kern.h linux-2.6.29.1/include/linux/vt_kern.h
@@ -97,6 +97,7 @@ void reset_vc(struct vc_data *vc);
 extern int unbind_con_driver(const struct consw *csw, int first, int last,
 			     int deflt);
 int vty_init(const struct file_operations *console_fops);
+void set_palette(struct vc_data *vc);
 
 /*
  * vc_screen.c shares this temporary buffer with the console write code so that

--
To unsubscribe from this list: send the line "unsubscribe linux-api" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#3 lore raw

Hi,

I sent this patch a month ago and didn't get any reply. Should I consider you dropped it and should I forget about it or what?

I know you are busy, sorry to bother with this mini-stuff, I just would like to know.

Best regards,
Cedric.

Quoted 173 lines · ----- "Cedric Roux

From: Cedric Roux <[email protected]>

A ioctl interface and two ioctl commands added to /dev/vcsaN to get/put the current palette of the given tty.

---
This patch exists because there is no way to get the current installed palette of a given tty. The PIO_CMAP and GIO_CMAP in vt_ioctl.c:vt_ioctl play with the global default colormap. And /dev/vcsaN don't dump the palette through their read interface. And since a user may change colors of a given tty by escape sequences, one should be able to retrieve the palette through /dev/vcsaN, leading to this patch.

Some points I am not fully confident with:
- vt.c:set_colormap is now used in vc_screen.c so cannot be static anymore.
The goal of this patch is to allow read access to the palette, I can remove the PIO_CMAP case and make set_colormap static again. This case is there because /dev/vcsaN is read/write so the access to the palette should also be read/write. (Or I can do it differently, no problem, just tell me how.)
- the palette is returned as it is stored in the kernel, without taking into account color_table. So if a userland program gets an attribute for a given character, it won't match with the palette information. Say the user reads 0xXY through /dev/vcsa and takes Y as an index in the palette, it will return a bad color. One should do palette[color_table[Y]] to get the real color.
Two "solutions" to this "problem" (it might be an non-issue):
+ export the palette by indexing with color_table:
I don't like it.
+ let the userland program deal with that:
the program may open a tty, send escape sequences to set the current color, write a character at a given position (through the tty, not /dev/vcs) and read back the attribute through /dev/vcsa this time. She will see what attribute corresponds to what color she sets by using escape sequences.
If the color_table is non-mutable, all this is a non-issue. Userland programs just store the color_table and live with that. I've seen no code that modifies the color_table, so I bet it's non-mutable. (May it change in the future?)
- lock_kernel() is called. Maybe it's not necessary? I'm not confident enough. You decide.

diff -uprN -X linux-2.6.29.1/Documentation/dontdiff linux-2.6.29.1-vanilla/drivers/char/vc_screen.c linux-2.6.29.1/drivers/char/vc_screen.c
--- linux-2.6.29.1-vanilla/drivers/char/vc_screen.c 2009-04-02
22:55:27.000000000 +0200
+++ linux-2.6.29.1/drivers/char/vc_screen.c 2009-04-14
10:16:33.000000000 +0200

(quoted diff) +0 -0
@@ -19,6 +19,8 @@
  * [email protected] - modified not to send characters to

wrong console
* - fixed some fatal off-by-one bugs (0-- no longer == -1 ->
looping and looping and looping...)
* - making it shorter - scr_readw are macros which expand in PRETTY
long code
+ *
+ * Colormap put/get for /dev/vcsaN, Cedric Roux <[email protected]>, March 2009.
*/

#include <linux/kernel.h>

drivers/char/vc_screen.c +56 -0
@@ -470,11 +472,67 @@ vcs_open(struct inode *inode, struct fil
 	return ret;
 }
 
+static long
+vcs_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
+{
+	unsigned int currcons = iminor(file->f_path.dentry->d_inode);
+	struct vc_data *vc;
+	unsigned char pal[16*3];
+	long ret = 0;
+
+	/* access only defined for /dev/vcsaN, not /dev/vcsN */
+	if (currcons < 128)
+		return -ENOTTY;
+
+	lock_kernel();
+
+	acquire_console_sem();
+
+	currcons &= 127;
+	if (currcons == 0)
+		currcons = fg_console;
+	else
+		currcons--;
+	if (!vc_cons_allocated(currcons)) {
+		ret = -ENXIO;
+		goto unlock_out;
+	}
+	vc = vc_cons[currcons].d;
+
+	switch (cmd) {
+	case PIO_CMAP:
+		if (copy_from_user(pal, (void __user *)arg, 16*3)) {
+			ret = -EFAULT;
+			goto unlock_out;
+		}
+		memcpy(vc->vc_palette, pal, 16*3);
+		set_palette(vc);
+		break;
+	case GIO_CMAP:
+		if (copy_to_user((void __user *)arg, vc->vc_palette, 16*3)) {
+			ret = -EFAULT;
+			goto unlock_out;
+		}
+		break;
+	default:
+		ret = -ENOIOCTLCMD;
+		break;
+	}
+
+unlock_out:
+	release_console_sem();
+
+	unlock_kernel();
+
+	return ret;
+}
+
 static const struct file_operations vcs_fops = {
 	.llseek		= vcs_lseek,
 	.read		= vcs_read,
 	.write		= vcs_write,
 	.open		= vcs_open,
+	.unlocked_ioctl	= vcs_ioctl,
 };
 
 static struct class *vc_class;

diff -uprN -X linux-2.6.29.1/Documentation/dontdiff linux-2.6.29.1-vanilla/drivers/char/vt.c linux-2.6.29.1/drivers/char/vt.c
--- linux-2.6.29.1-vanilla/drivers/char/vt.c 2009-04-02
22:55:27.000000000 +0200
+++ linux-2.6.29.1/drivers/char/vt.c 2009-04-14 10:17:20.000000000
+0200

drivers/char/vt.c +1 -2
@@ -156,7 +156,6 @@ static void set_cursor(struct vc_data *v
 static void hide_cursor(struct vc_data *vc);
 static void console_callback(struct work_struct *ignored);
 static void blank_screen_t(unsigned long dummy);
-static void set_palette(struct vc_data *vc);
 
 static int printable;		/* Is console ready for printing? */
 int default_utf8 = true;
@@ -3753,7 +3752,7 @@ void poke_blanked_console(void)
  *	Palettes
  */
 
-static void set_palette(struct vc_data *vc)
+void set_palette(struct vc_data *vc)
 {
 	WARN_CONSOLE_UNLOCKED();
 

diff -uprN -X linux-2.6.29.1/Documentation/dontdiff linux-2.6.29.1-vanilla/include/linux/vt_kern.h linux-2.6.29.1/include/linux/vt_kern.h
--- linux-2.6.29.1-vanilla/include/linux/vt_kern.h 2009-04-02
22:55:27.000000000 +0200
+++ linux-2.6.29.1/include/linux/vt_kern.h 2009-04-14
10:17:59.000000000 +0200

include/linux/vt_kern.h +1 -0
@@ -97,6 +97,7 @@ void reset_vc(struct vc_data *vc);
 extern int unbind_con_driver(const struct consw *csw, int first, int
 last,
 			     int deflt);
 int vty_init(const struct file_operations *console_fops);
+void set_palette(struct vc_data *vc);
 
 /*
Quoted 2 lines

* vc_screen.c shares this temporary buffer with the console write
code so that

--
To unsubscribe from this list: send the line "unsubscribe linux-api" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#4 lore raw
Quoted 3 lines · On Thu, 14 May 2009 09:20:17 +0200 (CEST) Cedric Roux…

I sent this patch a month ago and didn't get any reply. Should I consider you dropped it and should I forget about it or what?

I dropped it in response to Alan's off-list comments and then lost track of it.

--
To unsubscribe from this list: send the line "unsubscribe linux-api" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html