linux.git/drivers/tty/vt/vt.c, branch v4.7

tty: vt: Fix soft lockup in fbcon cursor blink timer.

2016-06-25T17:41:13+00:00

We are getting somewhat random soft lockups with this signature:

[   86.992215] [] el1_irq+0xa0/0x10c
[   86.997082] [] cursor_timer_handler+0x30/0x54
[   87.002991] [] call_timer_fn+0x54/0x1a8
[   87.008378] [] run_timer_softirq+0x1c4/0x2bc
[   87.014200] [] __do_softirq+0x114/0x344
[   87.019590] [] irq_exit+0x74/0x98
[   87.024458] [] __handle_domain_irq+0x98/0xfc
[   87.030278] [] gic_handle_irq+0x94/0x190

This is caused by the vt visual_init() function calling into
fbcon_init() with a vc_cur_blink_ms value of zero.  This is a
transient condition, as it is later set to a non-zero value.  But, if
the timer happens to expire while the blink rate is zero, it goes into
an endless loop, and we get soft lockup.

The fix is to initialize vc_cur_blink_ms before calling the con_init()
function.

Signed-off-by: David Daney 
Cc: stable@vger.kernel.org
Acked-by: Pavel Machek 
Tested-by: Ming Lei 
Acked-by: Scot Doyle 
Tested-by: Henrique de Moraes Holschuh 
Signed-off-by: Greg Kroah-Hartman

tty: vt, finish looping on duplicate

2016-05-03T22:59:15+00:00

When the console is already registered, stop crawling the
registered_con_driver array and return an error immediatelly.

This makes the code more obvious. And we do not need to initialize
retval anymore.

Signed-off-by: Jiri Slaby 
Signed-off-by: Greg Kroah-Hartman

tty: vt, return error when con_startup fails

2016-05-03T22:59:15+00:00

When csw->con_startup() fails in do_register_con_driver, we return no
error (i.e. 0). This was changed back in 2006 by commit 3e795de763.
Before that we used to return -ENODEV.

So fix the return value to be -ENODEV in that case again.

Fixes: 3e795de763 ("VT binding: Add binding/unbinding support for the VT console")
Signed-off-by: Jiri Slaby 
Reported-by: "Dan Carpenter" 
Signed-off-by: Greg Kroah-Hartman

tty: vt, use proper type for default colors

2016-04-30T16:26:55+00:00

Every user of default_red, default_grn, and default_blu treats them as
unsigned char. So make it really unsigned char.

And indent the initializers and module_param properly.

This saves ~ 100 bytes of data.

Signed-off-by: Jiri Slaby 
Signed-off-by: Greg Kroah-Hartman

tty: vt, make color_table const

2016-04-30T16:26:55+00:00

This means all ->con_set_palette have to have the second parameter
const too now.

Signed-off-by: Jiri Slaby 
Signed-off-by: Greg Kroah-Hartman

tty: vt, get rid of weird source code flow

2016-04-30T16:26:55+00:00

Some code in vc_allocate is indented by 4 spaces. It is inside a
condition. Invert the condition and move the code to the first
indentation level (using \tab). And insert some empty lines to have
logical code blocks separated.

Then, instead of freeing in an 'if' false branch, use goto-error
label as fail path.

Maybe better to look at this patch with diff -w -b.

Signed-off-by: Jiri Slaby 
Signed-off-by: Greg Kroah-Hartman

tty: vt, remove reduntant check

2016-04-30T16:26:55+00:00

MAX_NR_CONSOLES and MAX_NR_USER_CONSOLES are both 63 since they were
introduced in 1.1.54. And since vc_allocate does:

if (currcons >= MAX_NR_CONSOLES)
	return -ENXIO;

if (!vc_cons[currcons].d) {
	if (currcons >= MAX_NR_USER_CONSOLES && !capable(CAP_SYS_RESOURCE))
		return -EPERM;
}

the second check is pointless. Remove both the check and the macro
MAX_NR_USER_CONSOLES.

Signed-off-by: Jiri Slaby 
Reported-by: Fugang Duan 
Signed-off-by: Greg Kroah-Hartman

tty: vt: initialize softcursor_original correctly

2016-02-07T07:42:25+00:00

add_softcursor() stores the contents of the text buffer position in this
variable before drawing the softcursor, whereas hide_softcursor() writes
the value back. A value of -1 means that no cursor has been drawn and
therefore no character is to be restored. softcursor_original, however,
is only implicitly initialized with 0. Therefore, when hide_softcursor
is called for the first time (console_init -> con_init -> redraw_screen
-> hide_cursor), it wrongly writes 0x0000 in the top left corner of
the text buffer. Normally, this is just as black as the rest of the
screen (vc_video_erase_char) and can't be seen, but it appears as a
black cursor rectangle on non-black backgrounds e.g. with boot option
"vt.global_cursor_default=0 vt.color=0xf0". softcursor_original needs
to be initialized with -1.

Signed-off-by: Melchior FRANZ 
Signed-off-by: Greg Kroah-Hartman

Merge tag 'staging-4.5-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging

2016-02-01T01:00:27+00:00

Pull staging fixes from Greg KH:
 "Here are some small staging driver fixes for 4.5-rc2.

  One of them predated 4.4-final, but I missed that merge window due to
  the holliday.  The others fix reported issues that have come up
  recently.  The tty change is needed for the speakup driver fix and has
  the ack of the tty driver maintainer as well, i.e.  myself :)

  All have been in linux-next with no reported issues"

* tag 'staging-4.5-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  Staging: speakup: fix read scrolled-back VT
  Staging: speakup: Fix getting port information
  Revert "Staging: panel: usleep_range is preferred over udelay"
  iio: adis_buffer: Fix out-of-bounds memory access

Staging: speakup: fix read scrolled-back VT

2016-01-29T06:55:40+00:00

Previously, speakup would always read the non-scrolled part of the VT,
even when the VT is scrolled back with shift-page.  This patch makes
vt.c export screen_pos so that speakup can use it to properly access
the content of the scrolled-back VT.

This was tested with both vgacon and fbcon.

Signed-off-by: Samuel Thibault 
Reviewed-by: Alan Cox 
Signed-off-by: Greg Kroah-Hartman