Sysrq operations do not accept tty argument anymore so no need to pass
it to us.
[Stephen Rothwell <sfr@canb.auug.org.au>: fix build breakage in drm code
caused by sysrq using bool but not including linux/types.h]
[Sachin Sant <sachinp@in.ibm.com>: fix build breakage in s390 keyboadr
driver]
Acked-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Acked-by: Jason Wessel <jason.wessel@windriver.com>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
The use of "hvc_con_driver" as the name for a file-static "struct
console" with a ".setup" field pointing to an __init function causes
a modpost warning, since a non-initdata structure points to init code.
Using "hvc_console" as the name triggers the hacky "*_console"
workaround in modpost to silence the warning, and is the same thing
that most of the other console drivers already do.
I made the same change in hvsi.c since I happened to notice it was
likely to suffer from the same problem.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
I don't claim to understand the tty layer, but it seems like hvc_open and
hvc_close should be balanced in their kref reference counting.
Right now we get a kref every call to hvc_open:
if (hp->count++ > 0) {
tty_kref_get(tty); <----- here
spin_unlock_irqrestore(&hp->lock, flags);
hvc_kick();
return 0;
} /* else count == 0 */
tty->driver_data = hp;
hp->tty = tty_kref_get(tty); <------ or here if hp->count was 0
But hvc_close has:
tty_kref_get(tty);
if (--hp->count == 0) {
...
/* Put the ref obtained in hvc_open() */
tty_kref_put(tty);
...
}
tty_kref_put(tty);
Since the outside kref get/put balance we only do a single kref_put when
count reaches 0.
The patch below changes things to call tty_kref_put once for every
hvc_close call, and with that my machine boots fine.
Signed-off-by: Anton Blanchard <anton@samba.org>
Acked-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
Alan pointed out a race in the code where hvc_remove is invoked. The
recent virtio_console work is the first user of hvc_remove().
Alan describes it thus:
The hvc_console assumes that a close and remove call can't occur at the
same time.
In addition tty_hangup(tty) is problematic as tty_hangup is asynchronous
itself....
So this can happen
hvc_close hvc_remove
hung up ? - no
lock
tty = hp->tty
unlock
lock
hp->tty = NULL
unlock
notify del
kref_put the hvc struct
close completes
tty is destroyed
tty_hangup dead tty
tty->ops will be NULL
NULL->...
This patch adds some tty krefs and also converts to using tty_vhangup().
Reported-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Amit Shah <amit.shah@redhat.com>
CC: Alan Cox <alan@lxorguk.ukuu.org.uk>
CC: linuxppc-dev@ozlabs.org
CC: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Virtio consoles can be hotplugged, so hvc_alloc gets called from
multiple sites: from the initial probe() routine as well as later on
from workqueue handlers which aren't __devinit code.
So, drop the __devinit annotation for hvc_alloc.
Signed-off-by: Amit Shah <amit.shah@redhat.com>
Cc: linuxppc-dev@ozlabs.org
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The virtio console, which uses hvc, will get the ability to hot-unplug
ports. Export hvc_remove so that virtio_console can disassociate with
hvc.
Signed-off-by: Amit Shah <amit.shah@redhat.com>
Cc: linuxppc-dev@ozlabs.org
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Rename the locking free hvc_resize() function to __hvc_resize() and
provide an inline function that locks the hvc_struct and calls
__hvc_resize().
The rationale for this patch is that virtio_console calls the hvc_resize()
function without locking the hvc_struct. So it needs to call the lock
itself.
According to naming rules, the unlocked version is renamed and
prefixed with "__".
References to unlocked function calls in hvc back-ends has been updated.
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The result of container_of should not be NULL. In particular, in this case
the argument to the enclosing function has passed though INIT_WORK, which
dereferences it, implying that its container cannot be NULL.
A simplified version of the semantic patch that makes this change is as
follows:
(http://www.emn.fr/x-info/coccinelle/)
// <smpl>
@@
identifier fn,work,x,fld;
type T;
expression E1,E2;
statement S;
@@
static fn(struct work_struct *work) {
... when != work = E1
x = container_of(work,T,fld)
... when != x = E2
- if (x == NULL) S
...
}
// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This function does not have an error return and returning an error is
instead interpreted as having a lot of pending bytes.
Reported by Jeff Harris who provided a list of some of the remaining
offenders.
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch removes the tty->low_latency setting.
For irq based hvc_console backends the tty->low_latency must be set to 0,
because the tty_flip_buffer_push() function must not be called from IRQ context
(see drivers/char/tty_buffer.c).
For polled backends, the low_latency setting causes the bug trace below, because
tty_flip_buffer_push() is called within an atomic context and subsequent calls
might sleep due to mutex_lock.
BUG: sleeping function called from invalid context at /root/cvs/linux-2.6.git/kernel/mutex.c:207
in_atomic(): 1, irqs_disabled(): 0, pid: 748, name: khvcd
1 lock held by khvcd/748:
#0: (hvc_structs_lock){--..}, at: [<00000000002ceb50>] khvcd+0x58/0x12c
CPU: 0 Not tainted 2.6.29-rc1git #29
Process khvcd (pid: 748, task: 000000002fb9a480, ksp: 000000002f66bd78)
070000000000000a 000000002f66ba00 0000000000000002 (null)
000000002f66baa0 000000002f66ba18 000000002f66ba18 0000000000104f08
ffffffffffffc000 000000002f66bd78 (null) (null)
000000002f66ba00 000000000000000c 000000002f66ba00 000000002f66ba70
0000000000466af8 0000000000104f08 000000002f66ba00 000000002f66ba50
Call Trace:
([<0000000000104e7c>] show_trace+0x138/0x158)
[<0000000000104f62>] show_stack+0xc6/0xf8
[<0000000000105740>] dump_stack+0xb0/0xc0
[<000000000013144a>] __might_sleep+0x14e/0x17c
[<000000000045e226>] mutex_lock_nested+0x42/0x3b4
[<00000000002c443e>] echo_char_raw+0x3a/0x9c
[<00000000002c688c>] n_tty_receive_buf+0x1154/0x1208
[<00000000002ca0a2>] flush_to_ldisc+0x152/0x220
[<00000000002ca1da>] tty_flip_buffer_push+0x6a/0x90
[<00000000002cea74>] hvc_poll+0x244/0x2c8
[<00000000002ceb68>] khvcd+0x70/0x12c
[<000000000015bbd0>] kthread+0x68/0xa0
[<0000000000109d5a>] kernel_thread_starter+0x6/0xc
[<0000000000109d54>] kernel_thread_starter+0x0/0xc
1 lock held by khvcd/748:
#0: (hvc_structs_lock){--..}, at: [<00000000002ceb50>] khvcd+0x58/0x12c
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Replace kmalloc() + memset() with kzalloc().
Signed-off-by: Milton Miller <miltonm@bga.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
hvc_console is setting low_latency unconditionally, but some clients are
interrupt driven and will call hvc_poll from irq context. This will cause
tty_flip_buffer_push to be called from irq context, and it very clearly
states it must not be called from IRQ when low_latency is specified.
Looking back through history:
v2.6.16-rc1 via 33f0f88f1c
[PATCH] TTY layer buffering revamp
added this new api.
v2.6.16-rc3 via 8977d929e4
[PATCH] tty buffering stall fix
claims to fix a stall discovered with hvc_console
v2.6.16-rc5 via fb5c594c2a
[PATCH] Fix race condition in hvc console.
said set this flag to avoid a stall problem, and was merged through
the powerpc arch tree.
Without searching for email discussions, it would appear to be an
overlapping "fix", but one that did not consider all users.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
I remember some history on this barrier. There was a race between
open via /dev/console and the tty being fully setup. Its also why
there is a temporary variable and the global is assigned at the end
of the function.
Signed-off-by: Milton Miller <miltonm@bga.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We have special case logic for resizing pty/tty pairs. We also have a per
driver resize method so for the pty case we should use it.
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
this patch uses the new hvc callback hvc_resize to set the window size
which allows to change the tty size of hvc_console via a hvc_resize
function.
I have added a new feature bit VIRTIO_CONSOLE_F_SIZE. The driver will
change the window size on tty open and via the config_changed callback
of the transport. Currently lguest and kvm_s390 have not implemented this
callback, but the callback can be implemented at a later point in time.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The ctrl-o (^O) is a common control key used by several applications,
such as vim, but hvc_console uses ^O as the magic-sysrq key. This
commit allows users to send ^O to applications by pressing ^O twice
in succession.
To implement this, this commit introduces a check if ^O is pressed
again if the sysrq_pressed variable is already set. In this case,
clear sysrq_pressed state and flip the ^O character to the tty. (The
old behavior has always set "sysrq_pressed" if ^O has been entered,
and it has not flipped the ^O character to the tty.)
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The test to check for a new winsize runs out-of-sync with the
underlying tty. After a tty has been released and initialized again,
the winsize might differ between the tty and the hp struct. The
solution is to simply remove the check and always schedule the resize
work.
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Removed __devexit annotation of hvc_remove() to avoid a section mismatch
if the backend initialization fails and hvc_remove() must be used to
clean up allocated hvc structs (called in section __init or __devinit).
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The patch provides the hvc_resize() function to update the terminal
window dimensions (struct winsize) for a specified hvc console.
The function stores the new window size and schedules a function
that finally updates the tty winsize and signals the change to
user space (SIGWINCH).
Because the winsize update must acquire a mutex and might sleep,
the function is scheduled instead of being called from hvc_poll()
or khvcd.
This patch uses the tty_do_resize() routine from the tty layer.
A pending resize work is canceled in hvc_close() and hvc_hangup().
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
If put_char() routine of a hvc console backend returns 0, then the
hvc console starts looping in the following scenarios:
1. hvc_console_print()
If put_char() returns 0 then the while loop may loop forever.
I have added the missing check for 0 to throw away console messages.
2. khvcd may loop:
The thread calls hvc_poll() --> hvc_push()... if there are still
buffered data then the HVC_POLL_WRITE bit is set and causes the
khvcd thread to loop (if yield() returns immediately).
However, instead of looping, the khvcd thread could sleep for
MIN_TIMEOUT (doing the same as for get_chars()).
The MIN_TIMEOUT is set if hvc_push() was not able to write
data to the backend. If data has been written, the timeout is
set to 0 to immediately re-schedule hvc_poll().
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> (virtio_console)
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
After a tty hangup() or close() operation, processes might not reset the
termio settings to a sane state. In order to reset the termios to its
default settings the tty driver flag TTY_DRIVER_RESET_TERMIOS has been added.
TTY driver flag description from include/linux/tty_driver.h:
TTY_DRIVER_RESET_TERMIOS --- requests the tty layer to reset the
termios setting when the last process has closed the device.
Used for PTY's, in particular.
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Andrew Morton
Dmitry Torokhov
Chris Metcalf
Anton Blanchard
Tejun Heo
Amit Shah
Roel Kluin
Rusty Russell
Hendrik Brueckner
Julia Lawall
Alan Cox
Milton Miller
Alan Cox
Christian Borntraeger