| Age | Commit message (Collapse) | Author |
|---|
|
From: Mauro Carvalho Chehab <mchehab@redhat.com>
On Mon, 02 Feb 2009, Hartmut wrote:
This change set is wrong. The affected functions cannot be called from
an interrupt context, because they may process large buffers. In this
case, interrupts are disabled for a long time. Functions, like
dvb_dmx_swfilter_packets(), could be called only from a tasklet.
This change set does hide some strong design bugs in dm1105.c and
au0828-dvb.c.
Please revert this change set and do fix the bugs in dm1105.c and
au0828-dvb.c (and other files).
On Sun, 15 Feb 2009, Oliver Endriss wrote:
This changeset _must_ be reverted! It breaks all kernels since 2.6.27
for applications which use DVB and require a low interrupt latency.
It is a very bad idea to call the demuxer to process data buffers with
interrupts disabled!
On Mon, 16 Feb 2009, Trent Piepho wrote:
I agree, this is bad. The demuxer is far too much work to be done with
IRQs off. IMHO, even doing it under a spin-lock is excessive. It should
be a mutex. Drivers should use a work-queue to feed the demuxer.
Thank you for testing this changeset and discovering the issues on it.
Priority: high
CC: Trent Piepho <xyzzy@speakeasy.org>
CC: Hartmut <e9hack@googlemail.com>
CC: Oliver Endriss <o.endriss@gmx.de>
CC: Andreas Oberritter <obi@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Hans Verkuil <hverkuil@xs4all.nl>
Priority: normal
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Jan Engelhardt <jengelh@medozas.de>
dvb: constify VFTs
Signed-off-by: Jan Engelhardt <jengelh@medozas.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Mauro Carvalho Chehab <mchehab@redhat.com>
Author: Laurent Pinchart <laurent.pinchart@skynet.be>
Check fops_get() return value
Several subsystem open handlers dereference the fops_get() return value
without checking it for nullness. This opens a race condition between the
open handler and module unloading.
A module can be marked as being unloaded (MODULE_STATE_GOING) before its
exit function is called and gets the chance to unregister the driver.
During that window open handlers can still be called, and fops_get() will
fail in try_module_get() and return a NULL pointer.
This change checks the fops_get() return value and returns -ENODEV if NULL.
kernel-sync:
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Mauro Carvalho Chehab <mchehab@redhat.com>
/home/v4l/master/v4l/dvb_frontend.c:838:19: warning: symbol 'dtv_cmds' was not declared. Should it be static?
/home/v4l/master/v4l/dvb_frontend.c:1035:6: warning: symbol 'dtv_property_dump' was not declared. Should it be static?
/home/v4l/master/v4l/dvb_frontend.c:1066:5: warning: symbol 'is_legacy_delivery_system' was not declared. Should it be static?
/home/v4l/master/v4l/dvb_frontend.c:1080:6: warning: symbol 'dtv_property_cache_sync' was not declared. Should it be static?
/home/v4l/master/v4l/dvb_frontend.c:1132:6: warning: symbol 'dtv_property_legacy_params_sync' was not declared. Should it be static?
/home/v4l/master/v4l/dvb_frontend.c:1187:6: warning: symbol 'dtv_property_adv_params_sync' was not declared. Should it be static?
/home/v4l/master/v4l/dvb_frontend.c:1222:6: warning: symbol 'dtv_property_cache_submit' was not declared. Should it be static?
/home/v4l/master/v4l/dvb_frontend.c:1253:5: warning: symbol 'dtv_property_process_get' was not declared. Should it be static?
/home/v4l/master/v4l/dvb_frontend.c:1362:5: warning: symbol 'dtv_property_process_set' was not declared. Should it be static?
Priority: normal
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Mauro Carvalho Chehab <mchehab@redhat.com>
Author: Wang Chen <wangchen@cn.fujitsu.com>
dvb: Kill directly reference of netdev->priv
Simply replace netdev->priv with netdev_priv().
kernel-sync:
Priority: normal
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Darron Broad <darron@kewl.org>
It has been noticed that in HW tuning mode in cx24123.c that
a check is made to not alter the status var when in one shot
mode, a simpler solution is to not update here for all cards.
Priority: normal
Signed-off-by: Darron Broad <darron@kewl.org>
CC: Steven Toth <stoth@hauppauge.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Devin Heitmueller <devin.heitmueller@gmail.com>
It is not safe to assume that the i2c gate will be open before issuing the
command to power down the tuner. In fact, many demods only open the gate
long enough to issue the tuning command.
This fix allows power management to work properly for those tuners behind an
i2c gate (in my case the problem was with the HVR-950Q)
Signed-off-by: Devin Heitmueller <devin.heitmueller@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Devin Heitmueller <devin.heitmueller@gmail.com>
ATSC should be considered a legacy delivery system, or else fields such as
p->u.vsb.modulation do not get populated (resulting in set_frontend failures)
CC: Steven Toth <stoth@linuxtv.org>
Signed-off-by: Devin Heitmueller <devin.heitmueller@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Hans Verkuil <hverkuil@xs4all.nl>
Priority: normal
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
|
|
From: Stephen Rothwell <sfr@canb.auug.org.au>
Today's linux-next build (x86_64 allmodconfig) failed like this:
drivers/media/dvb/dvb-core/dvbdev.c: In function 'dvb_register_device':
drivers/media/dvb/dvb-core/dvbdev.c:254: error: implicit declaration of function 'nums2minor'
Caused by commit f3d9182030d9655d9ffff33dfceb45b4a127b6ec ("V4L/DVB
(9361): Dynamic DVB minor allocation") which clearly has not even been
built with CONFIG_DVB_DYNAMIC_MINORS=y.
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Mauro Carvalho Chehab <mchehab@redhat.com>
Priority: normal
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
Issue: Currently, the DVBFE_ALGO_CUSTOM is called only at the first tuning.
After the first call, the status is never set to force a new tuning.
From: Marco Schluessler <marco@lordzodiac.de>
Signed-off-by: Marco Schluessler <marco@lordzodiac.de>
Signed-off-by: Manu Abraham <manu@linuxtv.org>
|
|
From: Manu Abraham <abraham.manu@gmail.com>
Signed-off-by: Manu Abraham <manu@linuxtv.org>
|
|
From: Matthias Schwarzott <zzam@gentoo.org>
Signed-off-by: Matthias Schwarzott <zzam@gentoo.org>
Signed-off-by: Manu Abraham <manu@linuxtv.org>
|
|
From: Kay Sievers <kay.sievers@vrfy.org>
This allows udev to create proper device nodes without any
hacky shell scripts/programs to call, which guess these numbers
from the kernel device names.
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Andreas Oberritter <obi@linuxtv.org>
This is a multi-part message in MIME format.
Implement dynamic minor allocation for DVB, to allow more than four
devices of the same type per adapter, based on drivers/usb/core/file.c.
Add a new config option, DVB_DYNAMIC_MINORS, to make use of this
feature, which defaults to no for backwards compatibility.
Signed-off-by: Andreas Oberritter <obi@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Mauro Carvalho Chehab <mchehab@redhat.com>
Priority: normal
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
rather than implementing different ways leading to confusion.
This allows multiple gate_enable/disable's in the tuner_read/write
functions, thereby lesser number of I/O operations throughout,
eventually leading to better results. As a side effect demods that
detect the STOP bit for auto closing of the gate can be avoided, thereby
a very minimal gain in disabling the auto detect feature as well.
Improves readability on the device control.
From: Manu Abraham <abraham.manu@gmail.com>
Signed-off-by: Manu Abraham <manu@linuxtv.org>
|
|
From: Arvo Jarve <arvo@softshark.ee>
Signed-off-by: Arvo Jarve <arvo@softshark.ee>
Signed-off-by: Manu Abraham <manu@linuxtv.org>
|
|
From: Manu Abraham <abraham.manu@gmail.com>
Signed-off-by: Manu Abraham <manu@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
Remove spurious messages and turn into debug.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Darron Broad <darron@kewl.org>
When moving from one frontend to another
an application could spawn multiple threads opening
the same new frontend and in some circumstances all of
these could become delayed waiting for the previous
frontend readers or previous frontend writer thread to
complete.
In this scenario the first thread will succeed on open
to bring the new frontend online but any others will return
EBUSY. This is a fault. If the first succeeds and all others
are on the same frontend then they should succeed also.
Priority: normal
Signed-off-by: Darron Broad <darron@kewl.org>
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Mauro Carvalho Chehab <mchehab@redhat.com>
Priority: normal
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Manu Abraham <abraham.manu@gmail.com>
Signed-off-by: Manu Abraham <manu@linuxtv.org>
|
|
From: Darron Broad <darron@kewl.org>
This add frontend R/W mutual exclusion.
Prior to this point in time it was possible to open both
frontends simultaneously which an MFE card cannot support.
In order to stop this, a delayed open is performed which
has the following function:
Return EBUSY after a configurable amount of time
if a frontend is unavailable due to the other being
in use.
Only allow opening of a frontend if the kernel thread
of the other has stopped.
This solution was chosen to allow switching between
frontends to work as seamlessly as possible. When both
frontends are actually opened simultaneously then one
will only open, but if quick switching is performed
between one of many then the new open will succeed in
a clean fashion rather than interrupting a kernel
thread.
Priority: normal
Signed-off-by: Darron Broad <darron@kewl.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
A detailed description from the original patches 2 years ago:
"The WinTV-HVR3000 has a single transport bus which is shared between
a DVB-T and DVB-S modulator. These patches build on the bus acquisition
cx88 work from a few weeks ago to add support for this.
So to applications the HVR3000 looks like this:
/dev/dvb/adapter0/fe0 (cx24123 DVB-S demod)
/dev/dvb/adapter0/fe1 (cx22702 DVB-T demod)
Additional boards continue as before, eg:
/dev/dvb/adapter1/fe0 (lgdt3302 ATSC demod)
The basic change is removing the single instance of the videobuf_dvb in
cx8802_dev and saa7134_dev(?) and replacing it with a list and some
supporting functions.
*NOTE* This branch was taken before v4l-dvb was closed for 2.6.19 so
two or three current cx88 patches appear to be reversed by this tree,
this will be cleaned up in the near future. The patches missing change
the mutex handing to core->lock, fix an enumeration problem."
It should be recognised that a number of people have been maintaining
this patchset. Significant levels of Kudos to everyone one involved,
including but not limited to:
Darron Broad
Fabio M. Di Nitto
Carlo Scarfoglio
Hans Werner
Without the work of these people, and countless others, my two year old
patches would of died on the Mercurial linuxtv.org vine a long time
ago.
TODO: Revise these patches a little further so that the need for
demux1 and dvr0 is optional, not mandatory on the HVR3000.
HISTORY (darron):
This is the last update to MFE prepared by Hans which is based
upon the `scratchpad' diff created by Carlo.
All MFE work prior to that point must be attributed to Fabio
who ported and maintained Steve's original patch up to that
time.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
Signed-off-by: Darron Broad <darron@kewl.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
S2API: Return error of the caller provides 0 commands.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
A user tuning DVB-T via the S2API reports that this was not implemented,
and his tuning was failing.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
Tuning DVB-T via the S2API was failing, missing some essential items.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
Reports from users that using the new API for tuning DTV was failing,
and the cache was missing some essential items.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
... and cleanup any drivers using them.
I've also removed NBC_QPSK and modified the cx24116 driver to check
the delivery_type also, removing some excess namespace baggage.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
This means that when developers add new commands then they'll be see
the DTV_MAX_COMMAND define and will be more likely to modify it, without
having to modify the command validation code.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Darron Broad <darron@kewl.org>
Fixes switching from S2API to legacy where legacy params given to S2API
aware demods are cached ones and not those implied in the legacy API.
It was found (on an S2API aware demod at least) that after using
an S2API aware application and then switching to a legacy application
that prior S2API params were in effect were legacy has only
implied values. This fixes this.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
Signed-off-by: Darron Broad <darron@kewl.org>
|
|
From: Darron Broad <darron@kewl.org>
Quick fix to stop an OOPS if illegal commands are dumped in S2API.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
Signed-off-by: Darron Broad <darron@kewl.org>
|
|
From: Mauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Steven Toth <stoth@linuxtv.org>
This allows application developers to query the dvb-core API version
dynamically, helping developers understand whether certain features
will be available.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
This allows application developers to determine which particular
command in a sequence is invalid, or failing with error.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Andreas Oberritter <obi@linuxtv.org>
The functions dvb_dmxdev_section_callback, dvb_dmxdev_ts_callback,
dvb_dmx_swfilter_packet, dvb_dmx_swfilter_packets, dvb_dmx_swfilter and
dvb_dmx_swfilter_204 may be called from both interrupt and process
context. Therefore they need to be protected by spin_lock_irqsave()
instead of spin_lock().
This fixes a deadlock discovered by lockdep.
Signed-off-by: Andreas Oberritter <obi@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
From: Steven Toth <stoth@linuxtv.org>
We don't want to push the ISDB-T definitions into the kernel until
we have a high level of confidence in the ISDB-T API. More testing
is required before this code is released.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
Many thanks to Darron Broad for pointing out the obvious.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
Also added some compat code for the older API.
Added more ISDB message/command suggestions, current not connected in dvb-core.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Igor M. Liplianin <liplianin@me.by>
Bug fix: ioctl FE_SET_PROPERTY/FE_GET_PROPERTY always return error
Signed-off-by: Igor M. Liplianin <liplianin@me.by>
|
|
From: Steven Toth <stoth@linuxtv.org>
The frontends will be notified (if they chose) of all _get and _set commands
so they can help determine result or action. Results are now returned
to userspace correctly.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
The API now assumes that ioctl calls for FE_SET_PROPERTY and all
set commands, and FE_GET_PROPERTY are get commands.
Simplification.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
This helps complex demods which support different modulation types
be better informed.
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
|
From: Steven Toth <stoth@linuxtv.org>
Priority: normal
Signed-off-by: Steven Toth <stoth@linuxtv.org>
|
