From: Jasmin Jessich
On older Kernels (prior to 4.15) irqreturn_t and devm_request_threaded_irq
is not defined when compiling adv7604.c. It seems more recent Kernels
include it via another header which is included by adv7604.c.
Now we include linux/interrupt.h explicitly to get the type also
Hi!
Temporary add this patch to media-build until it is integrated in mainline.
This should fix the build errors for Kernels older than 4.12.
BR,
Jasmin
From: Jasmin Jessich
The portable type for a wait queue is wait_queue_head_t.
Signed-off-by: Jasmin Jessich
---
include/media/media-request.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/media/media-request.h b/include/media/media-request.h
index
Hello Hans!
> I'm inclined to drop support for 2.6 altogether.
RHEL 6 is on kernel 2.6.32, which we do not support since long time.
Most other distributions switched to 3.x also years in the past.
So lets drop 2.6.x then!
As you know I maintain a media-tree DKMS package and this should work for
Hi!
This was really hard to fix and I am still not ready.
But it should now compile for 4.19, 4.18 and 4.17.
Compiling is one thing, but functional is another. I added some inline
functions to compat.h (ida_alloc_min, ...) in 7283154cd56d to support them for
older Kernels. If someone is using an
From: Jasmin Jessich
This fixes the "'idx' may be used uninitialized in this function"
warning.
Cc: Jacopo Mondi
Signed-off-by: Jasmin Jessich
---
drivers/media/i2c/mt9v111.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/media/i2c/mt9v111.c
Hello Dan!
I checked this and this is in since ages. Good catch!
I did a quick look to the other places where this is checked also
and they seem all ok.
Acked-by: Jasmin Jessich
BR,
Jasmin
Hello Mauro!
> Patch looks ok, although it would be great to get some acks there.
I have no time to test this, but I know Daniel does all this tests and in fact
he is currently adapting changes done by DD in their repo, modifying them to
work in the current media-tree version and submits them to
Hello Mauro!
I got a request from a user, that media_build does not compile for older
Kernels. I checked that and found that the downloaded linux-media.tar.bz2 does
not contain "include/linux/overflow.h". Please can you add this file, because
media-build needs that file now.
BR,
Jasmin
You can add my Acked-by: Jasmin Jessich
On 06/19/2018 08:51 PM, Daniel Scheller wrote:
> From: Daniel Scheller
>
> There's a superfluous whitespace in the boilerplate license text in both
> .c and .h files. Fix this.
>
> Cc: Ralph Metzler
> Cc: Manfred Voelkel
> Signed-off-by: Daniel
You can add my Acked-by: Jasmin Jessich
On 06/19/2018 08:51 PM, Daniel Scheller wrote:
> From: Daniel Scheller
>
> In commit 3db30defab4b ("use correct MODULE_LINCESE for GPL v2 only
> according to notice in header") in the upstream repository for the
> mentioned driver at
You can add my Acked-by: Jasmin Jessich
On 06/19/2018 08:51 PM, Daniel Scheller wrote:
> From: Daniel Scheller
>
> As both the MODULE_LICENSE and the boilerplates are now in sync and clear
> that the driver is licensed under the terms of the GPLv2-only, add a
> matching SPDX license identifier
Hello Hans!
> linux-4.9.91-x86_64: ERRORS
> /home/hans/work/build/media_build/v4l/dw9807.c:321:3: error
The build works in my tree ... ?!?
I can't find the file in media_tree:
find . -name "*9807*"
Can you please check if there is something wrong with your
media_tree version which is used for
Hi!
THX for the patch.
@Hans:
I applied the patch already, but I can't delete it from Patchwork.
Please remove it there.
BR,
Jasmin
Hi!
I can't test it currently, but I reviewed it and it looks good.
So you can add my Acked-by .
BR,
Jasmin
Hi!
I can't test it currently, but I reviewed it and it looks good.
So you can add my Acked-by .
BR,
Jasmin
Hello Brad!
> Are the build logs public?
Look for the eMail
cron job: media_tree daily build: ?
The ? is OK, WARNINGS or ERRORS
In this eMail there are links to the short and long logfile.
Hans uses a complete build system which can download the Kernel sources, build
media-build
Hello Brad!
This patch added the inclusion of "linux/of_i2c.h".
This gave the warnings in the last nightly build for Kernel
3.6 - 3.9.
I just pushed a fix for that, so we should have an OK build this
night.
BR,
Jasmin
Hello Brad!
THX for the quick fix!
Hans did already apply it and Kernel 3.16 builds now.
I hope we will see a complete build tomorrow.
BR,
Jasmin
Hello Brad!
Tonight build broke due to patch 95ee4c285022!
You enabled VIDEO_OV2685 for 3.13., which doesn't
compile for Kernels older than 3.17. When you look
to the Kernel 3.17 section a lot of the drivers you
enabled for 3.13 with your patch should be enabled
for 3.17 only.
So please test
Hello Brad!
THX for this clarification!
So you tried already to fix the config_compat script and I agree with you that
it is difficult for you because of the various Kernels and distributions you
have to maintain.
Then your workaround is indeed a possibility to use media-build to build your
Hello Brad!
> and which the media_build system does not pick up on for whatever
> reason.
Maybe it would be better to analyse why "make_config_compat.pl" selects
wrongly the compatibility code.
> It seems there is quite often at least one backport I must disable,
> and some target kernels
From: Jasmin Jessich
In commit 20a63349b142 a new variable ktime_t delay has been added.
We decided to use the API functions to initialize ktime_t variables
within media-tree. Thus variable delay needs to be initialized with
ktime_set() instead of setting it directly.
Hello!
I just pushed a fix to media-build.
But I had to disable the driver for Kernels older than 4.16.
-> VIDEO_I2C requires device_get_match_data which requires
the function pointer device_get_match_data in fwnode_operations.
The latter has been added first in Kernel 4.16.
If someone wants
From: Jasmin Jessich
This reverts 8b0e195314fa, because media-tree drivers should use the
API functions to initialize variables of type ktime_t.
Cc: Thomas Gleixner
Signed-off-by: Jasmin Jessich
---
drivers/media/dvb-core/dmxdev.c | 2 +-
Hello Mauro!
> This patch looks fine, but not for the above-mentioned.
So I shall reword the commit message?
> The thing is that it is not consistent to have some places with
> things like:
> timeout = ktime_set(1, ir->polling * 100);
>
> and others with:
> timeout = ir->polling
Hello Mauro/Hans!
> Then you're probably not using the right version
Might be ...
The build script from Hans uses the Versions from here:
git://repo.or.cz/smatch.git
git://git.kernel.org/pub/scm/devel/sparse/sparse.git
> Yesterday, I added both trees I'm using here at:
>
Hello Mauro/Hans!
> There is already an upstream patch for hidding it:
The patch from https://patchwork.kernel.org/patch/10334353 will not
apply at the smatch tree.
Attached is an updated version for smatch.
Even with the patched tools, sparse still complains:
CC
From: Jasmin Jessich
This reverts 8b0e195314fa, because this will not compile for Kernels
older than 4.10.
Cc: Thomas Gleixner
Signed-off-by: Jasmin Jessich
---
drivers/media/dvb-core/dmxdev.c | 2 +-
drivers/media/pci/cx88/cx88-input.c |
From: Jasmin Jessich
Older Kernels use a struct for ktime_t, which requires the conversion
function ktime_to_ns to be used on some places. With this patch it will
compile now also for older Kernel versions.
Signed-off-by: Jasmin Jessich
---
THX for the patch!
Acked-by: Jasmin Jessich
Hi!
Sorry, my fault!
Will fix that soon.
BR,
Jasmin
On 03/11/2018 05:36 AM, Hans Verkuil wrote:
> This message is generated daily by a cron job that builds media_tree for
> the kernels and architectures in the list below.
>
> Results of the daily build of media_tree:
>
> date:
Hi!
Is OK for me, so:
Acked-by: Jasmin Jessich
BR,
Jasmin
Hallo Ralph!
> 1. The SW bit is cleared too early during the whole buffer size negotiation.
> This should be fixed.
I will look into this when I have time again. Probably end of next week.
> 2. IRQEN = CMDREG_DAIE = 0x80 is always set in the command register.
> So, they should probably only be
Hi!
Please hold on in merging this series, because I have to investigate a hint
I got related to the buffer size handshake of the protocol driver:
https://www.linuxtv.org/pipermail/linux-dvb/2007-July/019116.html
BR,
Jasmin
On 12/21/2017 02:22 PM, Jasmin J. wrote:
> From: Jasmin Jess
Hi!
Please hold on in merging this series, because I have to investigate a hint
I got related to the buffer size handshake of the protocol driver:
https://www.linuxtv.org/pipermail/linux-dvb/2007-July/019116.html
BR,
Jasmin
On 12/21/2017 02:22 PM, Jasmin J. wrote:
> From: Jasmin Jess
Hi!
Please hold on in merging this series, because I have to investigate a hint
I got related to the buffer size handshake of the protocol driver:
https://www.linuxtv.org/pipermail/linux-dvb/2007-July/019116.html
BR,
Jasmin
On 12/21/2017 02:22 PM, Jasmin J. wrote:
> From: Jasmin Jess
On 12/21/2017 02:22 PM, Jasmin J. wrote:
> From: Jasmin Jessich <jas...@anw.at>
>
> Some (older) CAMs are really slow in accepting data. I got sometimes the
> already known error "CAM tried to send a buffer larger than the ecount
> si
Will look into this during the next days.
BR,
Jasmin
Hi!
> Strictly speaking, building the media subsystem on older kernels is a job of
> the media build system.
Yes I agree.
> In general I would thus ask for the fix to be merged in media-build.git.
Which I do mostly, but in this case it is a coding error in mainline.
> In this specific case,
Hi!
> We're waiting for the end of the 4.16 merge window.
Ahh, this is the reason for the delay.
> For me that meant that I simply did (and still don't) have time to look at
> fixing the media_build.
No problem, I do it when I have time ;)
> The media_build is also broken worse than usual
Hi!
I tested to build with Kernel 4.14.17 and 4.4 and this worked.
Kernel 3.17 is currently not working, but I would like to wait with fixes
for Kernel 3.x after this patch has been applied to media_build.
BR,
Jasmin
From: Jasmin Jessich
Due the delay with merging patches to media_tree, add the mentioned
patch temporarily to backports. Once this patch is merged to media_tree,
revert this patch.
Signed-off-by: Jasmin Jessich
---
backports/backports.txt|
; Hi Jasmin,
>
> Thank you for the patch.
>
> On Sunday, 14 January 2018 12:21:43 EET Jasmin J. wrote:
>> From: Jasmin Jessich <jas...@anw.at>
>>
>> Commit 828ee8c71950 ("media: uvcvideo: Use ktime_t for timestamps")
>> changed to use ktime_t for
Hello Mauro!
> Ping!
> It is required to compile for Kernels older 4.10.
It would be nice to get this merged, so that we can see if older Kernels
will compile again.
BR,
Jasmin
Ping!
It is required to compile for Kernels older 4.10.
Hi!
With that it compiles again for Kernel 3.13.
Please note, that
https://patchwork.linuxtv.org/patch/46464/
needs to be applied to compile for Kernels older 4.10.
BR,
Jasmin
From: Jasmin Jessich
The driver requires linux/property.h, which is available since Kernel
3.18.17 only.
Signed-off-by: Jasmin Jessich
---
v4l/versions.txt | 4
1 file changed, 4 insertions(+)
diff --git a/v4l/versions.txt b/v4l/versions.txt
index
From: Jasmin Jessich
Signed-off-by: Jasmin Jessich
---
v4l/versions.txt | 2 ++
1 file changed, 2 insertions(+)
diff --git a/v4l/versions.txt b/v4l/versions.txt
index ed1e056..945e1c3 100644
--- a/v4l/versions.txt
+++ b/v4l/versions.txt
@@ -35,6 +35,8 @@
Hi!
I tested this patch (compile only) on Kernel 4.4.
For 3.13 there is something else not working, so I (or Hans) needs to have a
look on that.
BR,
Jasmin
From: Jasmin Jessich
Commit 828ee8c71950 ("media: uvcvideo: Use ktime_t for timestamps")
changed to use ktime_t for timestamps. Older Kernels use a struct for
ktime_t, which requires the conversion function ktime_to_ns to be used on
some places. With this patch it will compile now
Hi!
Maybe someone can apply:
https://patchwork.linuxtv.org/patch/46252/
BR,
Jasmin
Hi!
THX Mauro for only moving and not changing the videobuf2 files, so it was easy
to fix.
I tested the build with Kernel 4.4 and 3.13.
BR,
Jasmin
From: Jasmin Jessich
Commit 03fbdb2f moved videobuf2 to drivers/media/common. Modified the
backport patches to patch the files on the new location.
Signed-off-by: Jasmin Jessich
---
backports/v4.0_dma_buf_export.patch | 18 +-
Ping!
Missing to compile Kernel 4.2:
https://patchwork.linuxtv.org/patch/46153/
BR,
Jasmin
Hi!
I tested this with Kernel 4.4 and Kernel 3.13. It should fix the build for
newer Kernels also (requires my other build patch "Disabled
MEDIA_TUNER_TDA18250 for Kernels older than 4.3" (newer version) to be
applied).
Merry Christmas,
Jasmin
From: Jasmin Jessich
Also fixed v4.8_user_pages_flag.patch.
Signed-off-by: Jasmin Jessich
---
backports/v4.8_user_pages_flag.patch | 5 +++--
v4l/compat.h | 13 +
v4l/scripts/make_config_compat.pl| 1 +
3 files changed,
From: Jasmin Jessich
Added the documentation for the new ca? sysfs nodes in
/sys/class/dvb/dvb?/ca?/tim_wr_.
Signed-off-by: Jasmin Jessich
---
Documentation/ABI/testing/sysfs-class-ca| 63 ++
From: Jasmin Jessich
The device created by device_create in dvb_register_device was not
available for DVB device drivers.
Added "struct device *dev" to "struct dvb_device" and store the created
device.
Signed-off-by: Jasmin Jessich
Acked-by: Ralph Metzler
From: Jasmin Jessich
Some (older) CAMs are really slow in accepting data. The CI interface
specification doesn't define a handshake for accepted data. Thus, the
en50221 protocol driver can't control if a data byte has been correctly
written to the CAM.
The current implementation
From: Jasmin Jessich
Some (older) CAMs are really slow in accepting data. I got sometimes the
already known error "CAM tried to send a buffer larger than the ecount
size". I could track it down to the dvb_ca_en50221_write_data function not
waiting between sending the data
From: Jasmin Jessich
The previouse patch did enable the TDA18250 driver for Kernel 4.2.
Corrected this now.
Signed-off-by: Jasmin Jessich
---
v4l/versions.txt | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/v4l/versions.txt
Hi!
I could test only the compilation for Kernel 3.13 due to my laptop I am using
on vacation. It should compile now for all Kernels from 4.2 .. 3.13 also.
BR,
Jasmin
From: Jasmin Jessich
Due to recent changes the TDA18250 driver requires now regmap_write_bits,
which is available since Kernel 4.3 only.
Signed-off-by: Jasmin Jessich
---
v4l/versions.txt | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git
Hi!
> I don't know if you are also waiting for a comment from me.
All positive ones are welcome ;)
> I am fine with those changes. They are even desperately needed
> to get some "problem CAMs" working.
THX, so we can interpret this as an "Acked-by"?
BR,
Jasmin
Ping 2!
On 10/15/2017 03:59 PM, Jasmin J. wrote:
> Ping!
>
Hi!
With this patch it compiles back to kernel 2.6.36.
I tested it with 3.13, 3.4 and 2.6.36.
Don't forget to merge also my "READ_ONCE" patch.
BR,
Jasmin
From: Jasmin Jessich
I tried to to define "ktime_get_real_seconds" in "v4l/compat.h", but it is
using static variables from "kernel/time/timekeeping.c". So there is no other
way than reverting 41e33085284dd2bc6b6180d8381ff8a509b9d8ba for Kernels
older than 3.19.
Signed-off-by:
Hello Hans!
In one of my eMails I added a debug patch for linux/Makefile.
Patchwork interpreted this as a new version of the timer_setup_on_stack
patch. Please remove that from Patchwork.
BR,
Jasmin
Hi!
> The real problem is another, which I will fix soon.
I mixed up some branches, so there is no problem and there is nothing to fix.
Please delete this patch from Patchwork!
BR,
Jasmin
Hi!
Please ignore this patch!
The real problem is another, which I will fix soon.
BR,
Jasmin
Hi!
The code of this patch is already upstream, so it does no longer apply when
compiling for Kernel 3.13.
BR,
Jasmin
From: Jasmin Jessich
Signed-off-by: Jasmin Jessich
---
backports/backports.txt | 3 ---
backports/v3.13_ddbridge_pcimsi.patch | 29 -
2 files changed, 32 deletions(-)
delete mode 100644
From: Jasmin Jessich
Signed-off-by: Jasmin Jessich
---
v4l/compat.h | 5 +
v4l/scripts/make_config_compat.pl | 1 +
2 files changed, 6 insertions(+)
diff --git a/v4l/compat.h b/v4l/compat.h
index 89deae1..fba7a99 100644
--- a/v4l/compat.h
Hello Hans!
I try to fix the compilation for Kernel 3.13 (the kernel I use on my VDR).
In commit 5aeaa3e668de0782d1502f3d5751e2266a251d7c the timer handling in
the driver has been changed and now it uses "input_enable_softrepeat". This
function has been added with Kernel 4.4.
I tried to define
Hello Vincent!
> Not quite there yet however.
I always test with several Kernel headers ...
> patch -s -f -N -p1 -i ../backports/api_version.patch
> patch -s -f -N -p1 -i ../backports/pr_fmt.patch
> 1 out of 1 hunk FAILED
I have tested the patches on Linux Mint 18.1 (Kernel 4.4.0-62-generic):
Hello Hans!
With this patch it compiles for Kernel 4.4, but not on 3.13. I will work on
that soon.
I am not sure if this patch keeps pvrusb2 working, but it compiles. I tried
first a solution by reverting 8da0edf2f90b6c74b69ad420fdd230c9bd2bd1ed. If you
prefer this, I have it on a branch and can
From: Jasmin Jessich
Signed-off-by: Jasmin Jessich
---
v4l/compat.h | 5 +
v4l/scripts/make_config_compat.pl | 1 +
2 files changed, 6 insertions(+)
diff --git a/v4l/compat.h b/v4l/compat.h
index 4af407e..89deae1 100644
--- a/v4l/compat.h
Hello Hans!
I am not sure why compiler-gcc.h is used from the kernel tree and not from
the installed kernel headers. So a better fix for this problem might be to
remove compiler-gcc.h from media_build/linux/Makefile. This is something
you will know better.
Please note that this doesn't fix all
From: Jasmin Jessich
The type definitions have been splitting from compiler.h into
compiler_types.h. This requires to check for the new and the old
include guard in compiler-gcc.h.
Signed-off-by: Jasmin Jessich
---
backports/backports.txt | 3 +++
From: Jasmin Jessich
Signed-off-by: Jasmin Jessich
---
backports/pr_fmt.patch | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/backports/pr_fmt.patch b/backports/pr_fmt.patch
index 53f0273..d36e071 100644
--- a/backports/pr_fmt.patch
+++
Hello Hans!
> Jasmin, you've done some work on this in the past. Let me know if you
> have time to look at this, I would really appreciate that.
THX!
> No, it really is broken. I don't quite understand why it is failing ...
I noticed this already, but I am currently very busy to finish al my
From: Jasmin Jessich
Signed-off-by: Jasmin Jessich
---
v4l/versions.txt | 1 +
1 file changed, 1 insertion(+)
diff --git a/v4l/versions.txt b/v4l/versions.txt
index dab91c7..4ee0840 100644
--- a/v4l/versions.txt
+++ b/v4l/versions.txt
@@ -69,6 +69,7 @@
Hello Vincent!
> can someone take a look please?
Well, Matthias may have fixed that (I didn't try).
See:
https://www.mail-archive.com/linux-media@vger.kernel.org/msg121610.html
Maybe you need that also:
https://www.mail-archive.com/linux-media@vger.kernel.org/msg121625.html
BR,
Jasmin
Ping!
Good catch!
> Seems that this bug has been in the driver forever.
Indeed ... and I overlooked it during my recent changes to that module.
Reviewed-by: Jasmin Jessich
BR,
Jasmin
Hello Hans!
> Do you need build support for kernels < 2.6.36?
I don't, but I have the headers laying around on my disk and I simply tried to
fix the build ;)
> I gave up supporting such old kernels 2 or 3 years ago.
It seems so. I just learned, that RHEL 6.x does still use 2.6.32.
But there is
Hello Hans!
I provided some patches to solve all of this errors.
Please can apply them.
BR,
Jasmin
Hi!
I have fixed the patches for 2.6.32 partly.
File rc-ir-raw.c is still missing, but I can't fix that.
Moreover, when compiling for 2.6.32, I get errors from compat.h:
implicit declaration of function 'ktime_to_ms'
So it seems this needs a fix in compat.h also.
I am sending this patch
From: Jasmin Jessich
Signed-off-by: Jasmin Jessich
---
backports/v2.6.32_dvb_net.patch | 16 +--
backports/v2.6.32_kfifo.patch | 209
2 files changed, 112 insertions(+), 113 deletions(-)
diff --git
Hi!
I haven't tried to compile for 2.6.35, but the patches will now apply back
till version 2.6.33.
BR,
Jasmin
From: Jasmin Jessich
Signed-off-by: Jasmin Jessich
---
backports/v2.6.35_kfifo.patch| 11 ++-
backports/v2.6.35_work_handler.patch | 18 +-
2 files changed, 15 insertions(+), 14 deletions(-)
diff --git
Hi!
With this patch, the media-tree can be compiled back to Kernel 2.6.37.
For 2.6.32 some patches do not apply. I will fix that later.
@David:
Please can you review my changes if I reverted the patch correctly.
"git revert" didn't work, because of the changes in lirc_dev.c.
BR,
Jasmin
From: Jasmin Jessich
Compiling for Kernel 2.6.x failed in "lirc_dev.c" with
implicit declaration of function 'ida_simple_get'
This will in principle revert commit
46c8f4771154eb0dc21f5f2bc2640a33e8fe1d02 .
Signed-off-by: Jasmin Jessich
---
Hello Hans!
Will there be some "powerpoint" (or what ever) file of the presentations
available afterwards?
Or any other media to listen/see it later?
BR,
Jasmin
Hi!
Even with this patch Kernel 2.6.x still can't be compiled due to
the next error "ida_simple_get" missing.
I will try to fix the "ida_simple_get" error tomorrow.
BR,
Jasmin
From: Jasmin Jessich
Compiling for Kernel 2.6.x failed in "rc-main.c" with
linux/bsearch.h: No such file or directory
Beside adding the missing function, also the include of "linux/bsearch.h"
has been removed by new patch "v2.6_rc_main_bsearch_h.patch".
Signed-off-by: Jasmin
Hi!
Even with this patch Kernel 2.6.x can't be compiled due to "linux/bsearch.h"
missing. But this patch is the first step in the right direction.
BR,
Jasmin
From: Jasmin Jessich
Signed-off-by: Jasmin Jessich
---
backports/v2.6.37_dont_use_alloc_ordered_workqueue.patch | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/backports/v2.6.37_dont_use_alloc_ordered_workqueue.patch
From: Jasmin Jessich
Compiling for Kernel 3.4 failed in "ir-lirc-codec.c" with
error: 'U32_MAX' undeclared
Signed-off-by: Jasmin Jessich
---
v4l/compat.h | 4
v4l/scripts/make_config_compat.pl | 1 +
2 files changed, 5 insertions(+)
1 - 100 of 214 matches
Mail list logo