Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

classic Classic list List threaded Threaded
17 messages Options
Reply | Threaded
Open this post in threaded view
|

Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Basil Chupin-2
Driver 384.98 does not compile under kernel 4.14.9-1. The
nvidia-installer log shows this error (for those who understand what it
means):

-> Error.
ERROR: An error occurred while performing the step: "Checking to see
whether the nvidia-uvm kernel module was successfully built". See
/var/log/nvidia-installer.log for details.
-> The command `cd ./kernel; /usr/bin/make -k -j8
NV_KERNEL_MODULES="nvidia-uvm" NV_EXCLUDE_KERNEL_MODULES=""
SYSSRC="/lib/modules/4.14.9-1.g9423ca2-default/source"
SYSOUT="/lib/modules/4.14.9-1.g9423ca2-default/build"` failed with the
following output:

/usr/bin/make "CC=cc"
KBUILD_OUTPUT=/lib/modules/4.14.9-1.g9423ca2-default/build
KBUILD_VERBOSE= -C /lib/modules/4.14.9-1.g9423ca2-default/source
M=/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel ARCH=x86_64
NV_KERNEL_SOURCES=/lib/modules/4.14.9-1.g9423ca2-default/source
NV_KERNEL_OUTPUT=/lib/modules/4.14.9-1.g9423ca2-default/build
NV_KERNEL_MODULES="nvidia-uvm" INSTALL_MOD_DIR=kernel/drivers/video modules
make[1]: Entering directory '/usr/src/linux-4.14.9-1.g9423ca2'
make[2]: Entering directory
'/usr/src/linux-4.14.9-1.g9423ca2-obj/x86_64/default'
  CC [M] 
/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm8_va_block.o
In file included from
/usr/src/linux-4.14.9-1.g9423ca2/arch/x86/include/asm/cpufeature.h:5:0,
                 from
/usr/src/linux-4.14.9-1.g9423ca2/arch/x86/include/asm/thread_info.h:53,
                 from
/usr/src/linux-4.14.9-1.g9423ca2/include/linux/thread_info.h:38,
                 from
/usr/src/linux-4.14.9-1.g9423ca2/arch/x86/include/asm/preempt.h:7,
                 from
/usr/src/linux-4.14.9-1.g9423ca2/include/linux/preempt.h:81,
                 from
/usr/src/linux-4.14.9-1.g9423ca2/include/linux/spinlock.h:51,
                 from
/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/common/inc/nv-lock.h:16,
                 from
/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/common/inc/nv-linux.h:19,
                 from
/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm_linux.h:39,
                 from
/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm8_va_block.c:24:
/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm8_va_block.c:
In function ‘block_cpu_fault_locked’:
/usr/src/linux-4.14.9-1.g9423ca2/arch/x86/include/asm/processor.h:826:39:
error: implicit declaration of function ‘task_stack_page’; did you mean
‘task_stack_vm_area’? [-Werror=implicit-function-declaration]
  unsigned long __ptr = (unsigned long)task_stack_page(task); \
                                       ^
/usr/src/linux-4.14.9-1.g9423ca2/arch/x86/include/asm/processor.h:900:26:
note: in expansion of macro ‘task_pt_regs’
 #define KSTK_EIP(task)  (task_pt_regs(task)->ip)
                          ^~~~~~~~~~~~
/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm8_va_block.c:8771:41:
note: in expansion of macro ‘KSTK_EIP’
                                         KSTK_EIP(current));
                                         ^~~~~~~~
cc1: some warnings being treated as errors
make[3]: ***
[/usr/src/linux-4.14.9-1.g9423ca2/scripts/Makefile.build:315:
/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm8_va_block.o]
Error 1
make[3]: Target '__build' not remade because of errors.
make[2]: *** [/usr/src/linux-4.14.9-1.g9423ca2/Makefile:1509:
_module_/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel] Error 2
make[2]: Target 'modules' not remade because of errors.
make[2]: Leaving directory
'/usr/src/linux-4.14.9-1.g9423ca2-obj/x86_64/default'
make[1]: *** [Makefile:146: sub-make] Error 2
make[1]: Target 'modules' not remade because of errors.
make[1]: Leaving directory '/usr/src/linux-4.14.9-1.g9423ca2'
make: *** [Makefile:81: modules] Error 2
ERROR: The nvidia-uvm kernel module was not created.
ERROR: The nvidia-uvm kernel module failed to build. This kernel module
is required for the proper operation of CUDA. If you do not need to use
CUDA, you can try to install this driver package again with the
'--no-unified-memory' option.


Anyone know of a work-around (in lieu of the suggested option given in
the last para.?

BC

--
Always be nice to people on your way up -- you'll see
the same people on your way down.

N�����r��y隊Z)z{.���r�+�맲��r��z�^�ˬz��N�(�֜��^� ޭ隊Z)z{.���r�+��0�����Ǩ�
Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Christophe Giboudeaux-2
On mercredi 27 décembre 2017 07:18:48 CET Basil Chupin wrote:

> Driver 384.98 does not compile under kernel 4.14.9-1. The
> nvidia-installer log shows this error (for those who understand what it
> means):
>
> -> Error.
> ERROR: An error occurred while performing the step: "Checking to see
> whether the nvidia-uvm kernel module was successfully built". See
> /var/log/nvidia-installer.log for details.
> -> The command `cd ./kernel; /usr/bin/make -k -j8
> NV_KERNEL_MODULES="nvidia-uvm" NV_EXCLUDE_KERNEL_MODULES=""
> SYSSRC="/lib/modules/4.14.9-1.g9423ca2-default/source"
> SYSOUT="/lib/modules/4.14.9-1.g9423ca2-default/build"` failed with the
> following output:
>
> /usr/bin/make "CC=cc"
> KBUILD_OUTPUT=/lib/modules/4.14.9-1.g9423ca2-default/build
> KBUILD_VERBOSE= -C /lib/modules/4.14.9-1.g9423ca2-default/source
> M=/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel ARCH=x86_64
> NV_KERNEL_SOURCES=/lib/modules/4.14.9-1.g9423ca2-default/source
> NV_KERNEL_OUTPUT=/lib/modules/4.14.9-1.g9423ca2-default/build
> NV_KERNEL_MODULES="nvidia-uvm" INSTALL_MOD_DIR=kernel/drivers/video modules
> make[1]: Entering directory '/usr/src/linux-4.14.9-1.g9423ca2'
> make[2]: Entering directory
> '/usr/src/linux-4.14.9-1.g9423ca2-obj/x86_64/default'
>   CC [M]
> /tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm8_va_block.o
> In file included from
> /usr/src/linux-4.14.9-1.g9423ca2/arch/x86/include/asm/cpufeature.h:5:0,
>                  from
> /usr/src/linux-4.14.9-1.g9423ca2/arch/x86/include/asm/thread_info.h:53,
>                  from
> /usr/src/linux-4.14.9-1.g9423ca2/include/linux/thread_info.h:38,
>                  from
> /usr/src/linux-4.14.9-1.g9423ca2/arch/x86/include/asm/preempt.h:7,
>                  from
> /usr/src/linux-4.14.9-1.g9423ca2/include/linux/preempt.h:81,
>                  from
> /usr/src/linux-4.14.9-1.g9423ca2/include/linux/spinlock.h:51,
>                  from
> /tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/common/inc/nv-lock.h:16,
>                  from
> /tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/common/inc/nv-linux.h:19,
>                  from
> /tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm_linux.h:39,
> from
> /tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm8_va_block.c
> :24:
> /tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm8_va_block.
> c: In function ‘block_cpu_fault_locked’:
> /usr/src/linux-4.14.9-1.g9423ca2/arch/x86/include/asm/processor.h:826:39:
> error: implicit declaration of function ‘task_stack_page’; did you mean
> ‘task_stack_vm_area’? [-Werror=implicit-function-declaration]
>   unsigned long __ptr = (unsigned long)task_stack_page(task); \
>                                        ^
> /usr/src/linux-4.14.9-1.g9423ca2/arch/x86/include/asm/processor.h:900:26:
> note: in expansion of macro ‘task_pt_regs’
>  #define KSTK_EIP(task)  (task_pt_regs(task)->ip)
>                           ^~~~~~~~~~~~
> /tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm8_va_block.c
> :8771:41: note: in expansion of macro ‘KSTK_EIP’
>                                          KSTK_EIP(current));
>                                          ^~~~~~~~
> cc1: some warnings being treated as errors
> make[3]: ***
> [/usr/src/linux-4.14.9-1.g9423ca2/scripts/Makefile.build:315:
> /tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel/nvidia-uvm/uvm8_va_block.o
> ] Error 1
> make[3]: Target '__build' not remade because of errors.
> make[2]: *** [/usr/src/linux-4.14.9-1.g9423ca2/Makefile:1509:
> _module_/tmp/selfgz2375/NVIDIA-Linux-x86_64-384.98/kernel] Error 2
> make[2]: Target 'modules' not remade because of errors.
> make[2]: Leaving directory
> '/usr/src/linux-4.14.9-1.g9423ca2-obj/x86_64/default'
> make[1]: *** [Makefile:146: sub-make] Error 2
> make[1]: Target 'modules' not remade because of errors.
> make[1]: Leaving directory '/usr/src/linux-4.14.9-1.g9423ca2'
> make: *** [Makefile:81: modules] Error 2
> ERROR: The nvidia-uvm kernel module was not created.
> ERROR: The nvidia-uvm kernel module failed to build. This kernel module
> is required for the proper operation of CUDA. If you do not need to use
> CUDA, you can try to install this driver package again with the
> '--no-unified-memory' option.
>
>
> Anyone know of a work-around (in lieu of the suggested option given in
> the last para.?
>

Yes, read https://devtalk.nvidia.com/default/topic/1028016/b/t/post/5228999/




--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Basil Chupin-2
On 27/12/17 19:58, Christophe Giboudeaux wrote:
> On mercredi 27 décembre 2017 07:18:48 CET Basil Chupin wrote:
>> Driver 384.98 does not compile under kernel 4.14.9-1. The
>> nvidia-installer log shows this error (for those who understand what it
>> means):
>>
>> -> Error.
>> ERROR: An error occurred while performing the step: "Checking to see
>> whether the nvidia-uvm kernel module was successfully built". See
>> /var/log/nvidia-installer.log for details.
[pruned]


>> Anyone know of a work-around (in lieu of the suggested option given in
>> the last para.?
>>
> Yes, read https://devtalk.nvidia.com/default/topic/1028016/b/t/post/5228999/

Thank you, Christophe. Will apply the patch later today.

BC

--
Always be nice to people on your way up -- you'll see
the same people on your way down.


--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Basil Chupin-2
On 28/12/17 15:53, Basil Chupin wrote:

> On 27/12/17 19:58, Christophe Giboudeaux wrote:
>> On mercredi 27 décembre 2017 07:18:48 CET Basil Chupin wrote:
>>> Driver 384.98 does not compile under kernel 4.14.9-1. The
>>> nvidia-installer log shows this error (for those who understand what it
>>> means):
>>>
>>> -> Error.
>>> ERROR: An error occurred while performing the step: "Checking to see
>>> whether the nvidia-uvm kernel module was successfully built". See
>>> /var/log/nvidia-installer.log for details.
> [pruned]
>
>
>>> Anyone know of a work-around (in lieu of the suggested option given in
>>> the last para.?
>>>
>> Yes, read https://devtalk.nvidia.com/default/topic/1028016/b/t/post/5228999/
> Thank you, Christophe. Will apply the patch later today.
>
> BC

OK, I give up! :-)

Here is the patch mentioned in the URL above (I named the patch as
Testing.txt):

diff -durN nvidia-384-orig/nvidia-uvm/uvm8_va_block.c
nvidia-384/nvidia-uvm/uvm8_va_block.c
--- nvidia-384-orig/nvidia-uvm/uvm8_va_block.c    2017-12-26
11:20:17.097715622 +0100
+++ nvidia-384/nvidia-uvm/uvm8_va_block.c    2017-12-26
11:20:29.674381760 +0100
@@ -36,6 +36,10 @@
#include "uvm8_perf_prefetch.h"
#include "uvm8_mem.h"
    
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(4,14,9)
+#include <linux/sched/task_stack.h>
+#endif
+
typedef enum
{
   BLOCK_PTE_OP_MAP,


 I apply it and get the following result:


linux-8qdh:/data/Nvidia-384.98 # sh ./NVIDIA*.run --apply-patch Testing.txt
Verifying archive integrity... OK
Uncompressing NVIDIA Accelerated Graphics Driver for Linux-x86_64
384.98..........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
can't find file to patch at input line 4
Perhaps you used the wrong -p or --strip option?
The text leading up to this was:
--------------------------
|diff -durN nvidia-384-orig/nvidia-uvm/uvm8_va_block.c
nvidia-384/nvidia-uvm/uvm8_va_block.c
|--- nvidia-384-orig/nvidia-uvm/uvm8_va_block.c 2017-12-26
11:20:17.097715622 +0100
|+++ nvidia-384/nvidia-uvm/uvm8_va_block.c      2017-12-26
11:20:29.674381760 +0100
--------------------------
File to patch:
Skip this patch? [y]
Skipping patch.
/usr/bin/patch: **** malformed patch at line 12: typedef enum

Failed to apply patch file "/data/Nvidia-384.98/Testing.txt".


Now, I have only applied 3 or 4 patches in my life so I know bugger all
about patches and how to apply them. The cli command above is what I
found after searching around using a search engine. Obviously the wrong
command :-(, but I know not any better.

Could someone please "tell" me how to correctly apply the patch
mentioned in Christophe's post?

(Yesterday the latest version of the kernel -- 4.1.10-1.1 -- was
installed but this did not have the "fix" necessary to handle the Nvidia
driver 384.98.)

BC

--
Always be nice to people on your way up -- you'll see
the same people on your way down.


--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Dave Plater lst


On 31/12/2017 06:13, Basil Chupin wrote:

> On 28/12/17 15:53, Basil Chupin wrote:
>> On 27/12/17 19:58, Christophe Giboudeaux wrote:
>>> On mercredi 27 décembre 2017 07:18:48 CET Basil Chupin wrote:
>>>> Driver 384.98 does not compile under kernel 4.14.9-1. The
>>>> nvidia-installer log shows this error (for those who understand what it
>>>> means):
>>>>
>>>> -> Error.
>>>> ERROR: An error occurred while performing the step: "Checking to see
>>>> whether the nvidia-uvm kernel module was successfully built". See
>>>> /var/log/nvidia-installer.log for details.
>> [pruned]
>>
>>
>>>> Anyone know of a work-around (in lieu of the suggested option given in
>>>> the last para.?
>>>>
>>> Yes, read https://devtalk.nvidia.com/default/topic/1028016/b/t/post/5228999/
>> Thank you, Christophe. Will apply the patch later today.
>>
>> BC
>
> OK, I give up! :-)
>
> Here is the patch mentioned in the URL above (I named the patch as
> Testing.txt):
>
> diff -durN nvidia-384-orig/nvidia-uvm/uvm8_va_block.c
> nvidia-384/nvidia-uvm/uvm8_va_block.c
> --- nvidia-384-orig/nvidia-uvm/uvm8_va_block.c    2017-12-26
> 11:20:17.097715622 +0100
> +++ nvidia-384/nvidia-uvm/uvm8_va_block.c    2017-12-26

The above "---" and "---" lines show patch where to find the file to
patch, it seems your working directory is /data/Nvidia-384.98 so if you
delete the "nvidia-384-orig/" and "nvidia-384/" parts from the:
"--- nvidia-384-orig/nvidia-uvm/uvm8_va_block.c" and
"+++ nvidia-384/nvidia-uvm/uvm8_va_block.c" parts in patch then it
should apply, with a bit of luck.

> 11:20:29.674381760 +0100
> @@ -36,6 +36,10 @@
> #include "uvm8_perf_prefetch.h"
> #include "uvm8_mem.h"
>      
> +#if LINUX_VERSION_CODE >= KERNEL_VERSION(4,14,9)
> +#include <linux/sched/task_stack.h>
> +#endif
> +
> typedef enum
> {
>     BLOCK_PTE_OP_MAP,
>
>
>   I apply it and get the following result:
>
>
> linux-8qdh:/data/Nvidia-384.98 # sh ./NVIDIA*.run --apply-patch Testing.txt
> Verifying archive integrity... OK
> Uncompressing NVIDIA Accelerated Graphics Driver for Linux-x86_64
> 384.98..........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
> can't find file to patch at input line 4
> Perhaps you used the wrong -p or --strip option?

If you could have specified patch -p1 the patch would have applied but
it's nvidia's script which is running patch so hopefully the advice I
gave above will work.

> The text leading up to this was:
> --------------------------
> |diff -durN nvidia-384-orig/nvidia-uvm/uvm8_va_block.c
> nvidia-384/nvidia-uvm/uvm8_va_block.c
> |--- nvidia-384-orig/nvidia-uvm/uvm8_va_block.c 2017-12-26
> 11:20:17.097715622 +0100
> |+++ nvidia-384/nvidia-uvm/uvm8_va_block.c      2017-12-26
> 11:20:29.674381760 +0100
> --------------------------
> File to patch:
> Skip this patch? [y]
> Skipping patch.
> /usr/bin/patch: **** malformed patch at line 12: typedef enum
>
> Failed to apply patch file "/data/Nvidia-384.98/Testing.txt".
>
>
> Now, I have only applied 3 or 4 patches in my life so I know bugger all
> about patches and how to apply them. The cli command above is what I
> found after searching around using a search engine. Obviously the wrong
> command :-(, but I know not any better.
>
> Could someone please "tell" me how to correctly apply the patch
> mentioned in Christophe's post?
>
> (Yesterday the latest version of the kernel -- 4.1.10-1.1 -- was
> installed but this did not have the "fix" necessary to handle the Nvidia
> driver 384.98.)
>
> BC
>
Hope this works
Dave P
--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Markus Koßmann-3
Am Sonntag, 31. Dezember 2017, 08:14:51 schrieb Dave Plater:

> The above "---" and "---" lines show patch where to find the file to
> patch, it seems your working directory is /data/Nvidia-384.98 so if you
> delete the "nvidia-384-orig/" and "nvidia-384/" parts from the:
> "--- nvidia-384-orig/nvidia-uvm/uvm8_va_block.c" and
> "+++ nvidia-384/nvidia-uvm/uvm8_va_block.c" parts in patch then it
> should apply, with a bit of luck.
>
Another option would be to unpack the NVidia blob first with
NVIDIA-Linux-x86_64-384.98.run --extract-only
Then apply the patch and run nvidia-installer ( in the directory with the
unpacked files)  
--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Basil Chupin-2
On 31/12/17 17:40, Markus Koßmann wrote:

> Am Sonntag, 31. Dezember 2017, 08:14:51 schrieb Dave Plater:
>
>> The above "---" and "---" lines show patch where to find the file to
>> patch, it seems your working directory is /data/Nvidia-384.98 so if you
>> delete the "nvidia-384-orig/" and "nvidia-384/" parts from the:
>> "--- nvidia-384-orig/nvidia-uvm/uvm8_va_block.c" and
>> "+++ nvidia-384/nvidia-uvm/uvm8_va_block.c" parts in patch then it
>> should apply, with a bit of luck.
>>
> Another option would be to unpack the NVidia blob first with
> NVIDIA-Linux-x86_64-384.98.run --extract-only
> Then apply the patch and run nvidia-installer ( in the directory with the
> unpacked files)  

This was the first patch version I ran before posting my "I give up!"
message. Got the same error msg as I mentioned earlier.

@ Dave: doing what you suggested didn't work. Perhaps I misread your
comment about the "---" and the "---"? I don't suppose you would be so
kind as to alter that patch to what it should read in full? Anyway, will
try doing it all again tomorrow before I totally "lose it" tonight :-).

@ you and Markus: thank you for your help. Greatly appreciated. And a
Happy New Year to you both.

BC

--
Always be nice to people on your way up -- you'll see
the same people on your way down.


--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Dave Plater lst


On 31/12/2017 11:10, Basil Chupin wrote:
> @ Dave: doing what you suggested didn't work. Perhaps I misread your
> comment about the "---" and the "---"? I don't suppose you would be so
> kind as to alter that patch to what it should read in full? Anyway, will
> try doing it all again tomorrow before I totally "lose it" tonight:-).

I'll try attaching it first, then you can just apply it.
If you use Markus's method then you use "patch -p 1 Testing.txt" or
"patch -p 0 Testing_Basil_nVidia.patch" in the same directory.
The -p option strips x directory levels from the original patch.
 From the working directory use "find . -name uvm8_va_block.c" to find
what should actually be in the "---" and "+++" headers in the patch.
Using the "linux-8qdh:/data/Nvidia-384.98 # sh ./NVIDIA*.run
--apply-patch Testing.txt" method you are dependent on the -p level that
nvidia use in their build script. Haven't used nvidia run files for a
while but AFAIR you can use --help and it tells you an advanced help
option and maybe there's a way of specifying the -p option that you can
play around with.
Dave P

Testing_Basil_nVidia.patch (456 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Markus Koßmann-3
In reply to this post by Basil Chupin-2
Am Sonntag, 31. Dezember 2017, 20:10:07 schrieb Basil Chupin:

> On 31/12/17 17:40, Markus Koßmann wrote:
> > Am Sonntag, 31. Dezember 2017, 08:14:51 schrieb Dave Plater:
> >> The above "---" and "---" lines show patch where to find the file to
> >> patch, it seems your working directory is /data/Nvidia-384.98 so if you
> >> delete the "nvidia-384-orig/" and "nvidia-384/" parts from the:
> >> "--- nvidia-384-orig/nvidia-uvm/uvm8_va_block.c" and
> >> "+++ nvidia-384/nvidia-uvm/uvm8_va_block.c" parts in patch then it
> >> should apply, with a bit of luck.
> >
> > Another option would be to unpack the NVidia blob first with
> > NVIDIA-Linux-x86_64-384.98.run --extract-only
> > Then apply the patch and run nvidia-installer ( in the directory with the
> > unpacked files)
>
> This was the first patch version I ran before posting my "I give up!"
> message. Got the same error msg as I mentioned earlier.
>
As the patch just adds three three lines to uvm8_va_block.c you can also
"apply" the patch by editing that file manually.  Just add the three lines
behind the plus signs ( without the plus signs ) to  uvm8_va_block.c after the
block with the other #includes.  

--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Brüns, Stefan
In reply to this post by Dave Plater lst
On Sunday, December 31, 2017 11:54:35 AM CET Dave Plater wrote:

> On 31/12/2017 11:10, Basil Chupin wrote:
> > @ Dave: doing what you suggested didn't work. Perhaps I misread your
> > comment about the "---" and the "---"? I don't suppose you would be so
> > kind as to alter that patch to what it should read in full? Anyway, will
> > try doing it all again tomorrow before I totally "lose it" tonight:-).
>
> I'll try attaching it first, then you can just apply it.
> If you use Markus's method then you use "patch -p 1 Testing.txt" or
> "patch -p 0 Testing_Basil_nVidia.patch" in the same directory.
> The -p option strips x directory levels from the original patch.
>  From the working directory use "find . -name uvm8_va_block.c" to find
> what should actually be in the "---" and "+++" headers in the patch.
> Using the "linux-8qdh:/data/Nvidia-384.98 # sh ./NVIDIA*.run
> --apply-patch Testing.txt" method you are dependent on the -p level that
> nvidia use in their build script. Haven't used nvidia run files for a
> while but AFAIR you can use --help and it tells you an advanced help
> option and maybe there's a way of specifying the -p option that you can
> play around with.
> Dave P
The patch is malformed (hint: every line in the diff part has to start with
either '+', '-' or ' '). Thanks go to the person who posted in the forum
without providing an actual patch file ...

Regards,

Stefan

--
Stefan Brüns  /  Bergstraße 21  /  52062 Aachen
home: +49 241 53809034     mobile: +49 151 50412019

signature.asc (201 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Dave Plater lst


On 31/12/2017 21:43, Stefan Brüns wrote:

> On Sunday, December 31, 2017 11:54:35 AM CET Dave Plater wrote:
>> On 31/12/2017 11:10, Basil Chupin wrote:
>>> @ Dave: doing what you suggested didn't work. Perhaps I misread your
>>> comment about the "---" and the "---"? I don't suppose you would be so
>>> kind as to alter that patch to what it should read in full? Anyway, will
>>> try doing it all again tomorrow before I totally "lose it" tonight:-).
>>
>> I'll try attaching it first, then you can just apply it.
>> If you use Markus's method then you use "patch -p 1 Testing.txt" or
>> "patch -p 0 Testing_Basil_nVidia.patch" in the same directory.
>> The -p option strips x directory levels from the original patch.
>>   From the working directory use "find . -name uvm8_va_block.c" to find
>> what should actually be in the "---" and "+++" headers in the patch.
>> Using the "linux-8qdh:/data/Nvidia-384.98 # sh ./NVIDIA*.run
>> --apply-patch Testing.txt" method you are dependent on the -p level that
>> nvidia use in their build script. Haven't used nvidia run files for a
>> while but AFAIR you can use --help and it tells you an advanced help
>> option and maybe there's a way of specifying the -p option that you can
>> play around with.
>> Dave P
>
> The patch is malformed (hint: every line in the diff part has to start with
> either '+', '-' or ' '). Thanks go to the person who posted in the forum
> without providing an actual patch file ...
>
> Regards,
>
> Stefan
>
I never noticed that before so:
@@ -36,6 +36,10 @@
#include "uvm8_perf_prefetch.h"
#include "uvm8_mem.h"

+#if LINUX_VERSION_CODE >= KERNEL_VERSION(4,14,9)
+#include <linux/sched/task_stack.h>
+#endif
+
typedef enum
{
    BLOCK_PTE_OP_MAP,
should be:
@@ -36,6 +36,10 @@
  #include "uvm8_perf_prefetch.h"
  #include "uvm8_mem.h"

+#if LINUX_VERSION_CODE >= KERNEL_VERSION(4,14,9)
+#include <linux/sched/task_stack.h>
+#endif
+
  typedef enum
  {
     BLOCK_PTE_OP_MAP,
?

Dave P
--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Dave Plater lst
In reply to this post by Brüns, Stefan


On 31/12/2017 21:43, Stefan Brüns wrote:
> The patch is malformed (hint: every line in the diff part has to start with
> either '+', '-' or ' '). Thanks go to the person who posted in the forum
> without providing an actual patch file ...
>
> Regards,
>
> Stefan
This attached patch should apply, it has spaces in the right places.
Unfortunately I haven't a run package to actually check the patch.

Dave P

Testing_Basil_nVidia1.patch (461 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Basil Chupin-2
In reply to this post by Markus Koßmann-3
On 31/12/17 23:00, Markus Koßmann wrote:

> Am Sonntag, 31. Dezember 2017, 20:10:07 schrieb Basil Chupin:
>> On 31/12/17 17:40, Markus Ko�mann wrote:
>>> Am Sonntag, 31. Dezember 2017, 08:14:51 schrieb Dave Plater:
>>>> The above "---" and "---" lines show patch where to find the file to
>>>> patch, it seems your working directory is /data/Nvidia-384.98 so if you
>>>> delete the "nvidia-384-orig/" and "nvidia-384/" parts from the:
>>>> "--- nvidia-384-orig/nvidia-uvm/uvm8_va_block.c" and
>>>> "+++ nvidia-384/nvidia-uvm/uvm8_va_block.c" parts in patch then it
>>>> should apply, with a bit of luck.
>>> Another option would be to unpack the NVidia blob first with
>>> NVIDIA-Linux-x86_64-384.98.run --extract-only
>>> Then apply the patch and run nvidia-installer ( in the directory with the
>>> unpacked files)
>> This was the first patch version I ran before posting my "I give up!"
>> message. Got the same error msg as I mentioned earlier.
>>
> As the patch just adds three three lines to uvm8_va_block.c you can also
> "apply" the patch by editing that file manually.  Just add the three lines
> behind the plus signs ( without the plus signs ) to  uvm8_va_block.c after the
> block with the other #includes.

Good advice -- did that and the Nvidia 384.98 then compiled :-). Thank you.

@ Dave: sorry, Dave, but that rewrite of the patch still does not
'work', but your help is greatly appreciated.


BC

--
Always be nice to people on your way up -- you'll see
the same people on your way down.


--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Dave Plater lst


On 02/01/2018 10:35, Basil Chupin wrote:

> On 31/12/17 23:00, Markus Koßmann wrote:
>> Am Sonntag, 31. Dezember 2017, 20:10:07 schrieb Basil Chupin:
>>> On 31/12/17 17:40, Markus Ko�mann wrote:
>>>> Am Sonntag, 31. Dezember 2017, 08:14:51 schrieb Dave Plater:
>>>>> The above "---" and "---" lines show patch where to find the file to
>>>>> patch, it seems your working directory is /data/Nvidia-384.98 so if you
>>>>> delete the "nvidia-384-orig/" and "nvidia-384/" parts from the:
>>>>> "--- nvidia-384-orig/nvidia-uvm/uvm8_va_block.c" and
>>>>> "+++ nvidia-384/nvidia-uvm/uvm8_va_block.c" parts in patch then it
>>>>> should apply, with a bit of luck.
>>>> Another option would be to unpack the NVidia blob first with
>>>> NVIDIA-Linux-x86_64-384.98.run --extract-only
>>>> Then apply the patch and run nvidia-installer ( in the directory with the
>>>> unpacked files)
>>> This was the first patch version I ran before posting my "I give up!"
>>> message. Got the same error msg as I mentioned earlier.
>>>
>> As the patch just adds three three lines to uvm8_va_block.c you can also
>> "apply" the patch by editing that file manually.  Just add the three lines
>> behind the plus signs ( without the plus signs ) to  uvm8_va_block.c after the
>> block with the other #includes.
>
> Good advice -- did that and the Nvidia 384.98 then compiled :-). Thank you.
>
> @ Dave: sorry, Dave, but that rewrite of the patch still does not
> 'work', but your help is greatly appreciated.
>
>
> BC
>
You could have created the patch by using "diff -u
unedited-uvm8_va_block.c uvm8_va_block.c > nvidiapatch.patch", just
trying to explain what a patch is.
Dave P
--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Werner Flamme
Dave Plater [03.01.2018 09:42]:

> You could have created the patch by using "diff -u
> unedited-uvm8_va_block.c uvm8_va_block.c > nvidiapatch.patch", just
> trying to explain what a patch is.
> Dave P
>

The next problem is already lurking around the corner, I guess. I just
found <http://rglinuxtech.com/?p=2180>, titled "Kernel – 4.14.11 ‘Final’
Breaks NVIDIA – GPL Error.."

I love proprietary drivers...

Werner
--



smime.p7s (7K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Basil Chupin-2
On 04/01/18 03:35, Werner Flamme wrote:

> Dave Plater [03.01.2018 09:42]:
>
>> You could have created the patch by using "diff -u
>> unedited-uvm8_va_block.c uvm8_va_block.c > nvidiapatch.patch", just
>> trying to explain what a patch is.
>> Dave P
>>
> The next problem is already lurking around the corner, I guess. I just
> found <http://rglinuxtech.com/?p=2180>, titled "Kernel – 4.14.11 ‘Final’
> Breaks NVIDIA – GPL Error.."
>
> I love proprietary drivers...
>
> Werner

Yes, I just discovered this :-(.

Even trying to compile the driver with the additional command
'--no-unified-memory' now fails :-(.

Thankfully the 'clean old kernels' feature left the previous kernel
(4.14.10.3) in place and I was therefore able to boot up with the
'--no-unified-memory' option in place.

BC

--
Always be nice to people on your way up -- you'll see
the same people on your way down.


--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Nvidia driver 384.98 + latest kernel 4.14.9-1 in TW

Basil Chupin-2
In reply to this post by Dave Plater lst
On 03/01/18 19:42, Dave Plater wrote:

>
>
> On 02/01/2018 10:35, Basil Chupin wrote:
>> On 31/12/17 23:00, Markus Koßmann wrote:
>>> Am Sonntag, 31. Dezember 2017, 20:10:07 schrieb Basil Chupin:
>>>> On 31/12/17 17:40, Markus Ko�mann wrote:
>>>>> Am Sonntag, 31. Dezember 2017, 08:14:51 schrieb Dave Plater:
>>>>>> The above "---" and "---" lines show patch where to find the file to
>>>>>> patch, it seems your working directory is /data/Nvidia-384.98 so
>>>>>> if you
>>>>>> delete the "nvidia-384-orig/" and "nvidia-384/" parts from the:
>>>>>> "--- nvidia-384-orig/nvidia-uvm/uvm8_va_block.c" and
>>>>>> "+++ nvidia-384/nvidia-uvm/uvm8_va_block.c" parts in patch then it
>>>>>> should apply, with a bit of luck.
>>>>> Another option would be to unpack the NVidia blob first with
>>>>> NVIDIA-Linux-x86_64-384.98.run --extract-only
>>>>> Then apply the patch and run nvidia-installer ( in the directory
>>>>> with the
>>>>> unpacked files)
>>>> This was the first patch version I ran before posting my "I give up!"
>>>> message. Got the same error msg as I mentioned earlier.
>>>>
>>> As the patch just adds three three lines to uvm8_va_block.c you can
>>> also
>>> "apply" the patch by editing that file manually.  Just add the three
>>> lines
>>> behind the plus signs ( without the plus signs ) to  uvm8_va_block.c
>>> after the
>>> block with the other #includes.
>>
>> Good advice -- did that and the Nvidia 384.98 then compiled :-).
>> Thank you.
>>
>> @ Dave: sorry, Dave, but that rewrite of the patch still does not
>> 'work', but your help is greatly appreciated.
>>
>>
>> BC
>>
> You could have created the patch by using "diff -u
> unedited-uvm8_va_block.c uvm8_va_block.c > nvidiapatch.patch", just
> trying to explain what a patch is.
> Dave P

Thank you, Dave, I'll put this into my 'little black book' and apply it
(as part of my learning curve re patches; my latest quest is to get the
wording right to get 'makeself.sh' to compress the files into a *.run
file after a patch is applied).

BC

--
Always be nice to people on your way up -- you'll see
the same people on your way down.


--
To unsubscribe, e-mail: [hidden email]
To contact the owner, e-mail: [hidden email]