Skip to contentSkip to conversation

build error in pcm_native.c

DISCUSSION Andrew Gaydenko 3 participants 10 messages Jul 16, 2004 1 message not in the archive

lore
Share & tools
lore
#0 lore raw

I have updated all the alsa CVS tree (. CVS dir). Yesterday there was some updating of the file under problem.

Andrew

======= On Friday 16 July 2004 13:44, Takashi Iwai wrote: ======= At Fri, 16 Jul 2004 13:25:45 +0400,

Quoted 2 lines · Andrew Gaydenko

Takashi, yes I did.

And did you update both alsa-driver and alsa-kernel trees? The configure script was changed to detect the old remap_page_range().

Takashi

Quoted 8 lines

Andrew
======= On Friday 16 July 2004 13:22, you wrote: ======= At Thu, 15 Jul 2004 22:09:21 +0400,

Quoted 3 lines · Andrew Gaydenko

Hi,

Quoted 7 lines
Quoted 2 lines

Have tried to build drivers from current CVS and have got an error shown below. Must I supply some additional info?

Did you run cvscompile?

Takashi

-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click

#1 lore raw

At Fri, 16 Jul 2004 13:57:56 +0400,

Quoted 3 lines · Andrew Gaydenko

I have updated all the alsa CVS tree (. CVS dir). Yesterday there was some updating of the file under problem.

Then check alsa-driver/include/config.h whether
CONFIG_OLD_REMAP_PAGE_RANGE is defined.

Takashi

-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click

#3 lore raw
In reply to a message not in the archive

At Fri, 16 Jul 2004 14:14:54 +0400,

Quoted 4 lines · Andrew Gaydenko

There are no uncommented strings for this macro. The only is:

/* #undef CONFIG_OLD_REMAP_PAGE_RANGE */

Well then configure script couldn't detect it.
Please try the attached patch.

Quoted 4 lines

Andrew

P.S. Must I reply to [email protected] or [email protected]
or both?

Both please. Otherwise the information won't go outside. I don't like to keep such things secret :)

Takashi

Index: alsa-driver/configure.in
===================================================================
RCS file: /suse/tiwai/cvs/alsa/alsa-driver/configure.in,v retrieving revision 1.184

configure.in +3 -1
diff -u -r1.184 configure.in
@@ -1306,8 +1306,10 @@
 #include "$CONFIG_SND_KERNELDIR/include/linux/config.h"
 #include "$CONFIG_SND_KERNELDIR/include/linux/mm.h"
 ],[
-  void * (*func)(struct vm_area_struct *, unsigned long, unsigned long, unsigned long, pgprot_t);
+  int (*func)(struct vm_area_struct *, unsigned long, unsigned long, unsigned long, pgprot_t);
+  pgprot_t p;
   func = remap_page_range;
+  remap_page_range(NULL, 0, 0, 0, p);
 ],
     AC_MSG_RESULT("yes");new_remap="0",
     AC_MSG_RESULT("no");new_remap="1",
#4 lore raw

Have applied the patch. Unfortunately, the same:

gcc -D__KERNEL__ -DMODULE=1 -I/wrk/cvs/alsa/alsa-driver/include -I/lib/modules/2.4.26-lck1/build/include -O2 -mpreferred-stack-boundary=2 -march=i686 -D__SMP__ -DCONFIG_SMP -DLINUX -Wall -Wstrict-prototypes -fomit-frame-pointer -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -pipe -DALSA_BUILD -nostdinc -iwithprefix include -DKBUILD_BASENAME=pcm_native -c -o pcm_native.o pcm_native.c
pcm_native.c: In function `snd_pcm_mmap_data_Rsmp_c77fab7a':
pcm_native.c:3157: warning: passing arg 1 of remap_page_range_Rsmp_69d01e73' makes integer from pointer without a cast pcm_native.c:3157: error: incompatible type for argument 4 of remap_page_range_Rsmp_69d01e73' pcm_native.c:3157: error: too many arguments to function `remap_page_range_Rsmp_69d01e73'
make[1]: *** [pcm_native.o] Error 1

Andrew

======= On Friday 16 July 2004 14:28, you wrote: ======= At Fri, 16 Jul 2004 14:14:54 +0400,

Quoted 4 lines · Andrew Gaydenko

There are no uncommented strings for this macro. The only is:

/* #undef CONFIG_OLD_REMAP_PAGE_RANGE */

Well then configure script couldn't detect it.
Please try the attached patch.

Quoted 4 lines

Andrew

P.S. Must I reply to [email protected] or [email protected]
or both?

Both please. Otherwise the information won't go outside. I don't like to keep such things secret :)

Takashi

-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click

#5 lore raw

At Fri, 16 Jul 2004 14:41:40 +0400,

Quoted 2 lines · Andrew Gaydenko

Have applied the patch. Unfortunately, the same:

Did you run cvscompile again (after make clean)? What about config.h now?
If CONFIG_OLD_REMAP_PAGE_RANGE is still undefined, check config.log file.

Takashi

-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click

#6 lore raw

clean/(re)cvscompile were done.

I'm not bash guru, but from common view some configure fragment seems strange to me. I have got

checking for new remap_page_range... "no"

Here 1 is assigned to new_remap (line 2). But CONFIG_OLD_REMAP_PAGE_RANGE is defined (line 8) only when new_remap equals 0 (line 6). Must new_remap assignement be reverted?

The fragment is:

...
1. echo "$as_me:$LINENO: result: \"no\"" >&5
2. echo "${ECHO_T}\"no\"" >&6;new_remap="1"
3. fi
4. rm -f conftest.err conftest.$ac_objext conftest.$ac_ext
5. CFLAGS=$ac_save_CFLAGS
6. if test "new_remap" = "0"; then
7. cat >>confdefs.h <<\_ACEOF
8. #define CONFIG_OLD_REMAP_PAGE_RANGE 1
9. _ACEOF

fi
...

Andrew
======= On Friday 16 July 2004 16:21, you wrote: ======= At Fri, 16 Jul 2004 14:41:40 +0400,

Quoted 2 lines · Andrew Gaydenko

Have applied the patch. Unfortunately, the same:

Did you run cvscompile again (after make clean)? What about config.h now?
If CONFIG_OLD_REMAP_PAGE_RANGE is still undefined, check config.log file.

Takashi

-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click

#7 lore raw

At Fri, 16 Jul 2004 16:57:23 +0400,

Quoted 7 lines · Andrew Gaydenko

clean/(re)cvscompile were done.

I'm not bash guru, but from common view some configure fragment seems strange to me. I have got

checking for new remap_page_range... "no"

You have an old version of kernel, don't you?

Quoted 3 lines

Here 1 is assigned to new_remap (line 2). But CONFIG_OLD_REMAP_PAGE_RANGE is defined (line 8) only when new_remap equals 0 (line 6). Must new_remap assignement be reverted?

It defines CONFIG_ *OLD* _REMAP_PAGE_RANGE. So it's correct.

I don't understand why config.h wasn't generated properly...

Takashi

-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click

#8 lore raw

At Fri, 16 Jul 2004 15:08:14 +0200,

Quoted 11 lines · I wrote:

At Fri, 16 Jul 2004 16:57:23 +0400,

Quoted 7 lines · Andrew Gaydenko

clean/(re)cvscompile were done.

I'm not bash guru, but from common view some configure fragment seems strange to me. I have got

checking for new remap_page_range... "no"

Quoted 7 lines

You have an old version of kernel, don't you?

Quoted 3 lines

Here 1 is assigned to new_remap (line 2). But CONFIG_OLD_REMAP_PAGE_RANGE is defined (line 8) only when new_remap equals 0 (line 6). Must new_remap assignement be reverted?

It defines CONFIG_ *OLD* _REMAP_PAGE_RANGE. So it's correct.

Doh, I found you're right, the 1 and 0 are reverted.
And another silly typo. The corrected (additional) patch is below.

Takashi

Index: alsa-driver/configure.in
===================================================================
RCS file: /suse/tiwai/cvs/alsa/alsa-driver/configure.in,v retrieving revision 1.185

configure.in +4 -4
diff -u -r1.185 configure.in
@@ -1311,12 +1311,12 @@
   func = remap_page_range;
   remap_page_range(NULL, 0, 0, 0, p);
 ],
-    AC_MSG_RESULT("yes");new_remap="0",
-    AC_MSG_RESULT("no");new_remap="1",
-    AC_MSG_RESULT("unknown");new_remap="1"
+    AC_MSG_RESULT("yes");new_remap="1",
+    AC_MSG_RESULT("no");new_remap="0",
+    AC_MSG_RESULT("unknown");new_remap="0"
 )
 CFLAGS=$ac_save_CFLAGS
-if test "new_remap" = "0"; then
+if test "$new_remap" != "1"; then
   AC_DEFINE(CONFIG_OLD_REMAP_PAGE_RANGE)
 fi
 
#10 lore raw
In reply to Takashi Iwai
Quoted 5 lines · Takashi Iwai
Quoted 1 line · Andrew Gaydenko

/* #undef CONFIG_OLD_REMAP_PAGE_RANGE */

Well then configure script couldn't detect it.
Please try the attached patch.

Without the patch, config.log says:

configure:6138: checking for new remap_page_range
configure:6167: gcc -c -g -O2 -I/lib/modules/2.4.26/build/include conftest.c >&5
configure: In function `main':
configure:6174: warning: assignment from incompatible pointer type configure:6170: $? = 0
configure:6173: test -s conftest.o
configure:6176: $? = 0
configure:6178: result: "yes"

And the other similar checks have the same problem. Perhaps it would be better to use -Werror for these tests?

Clemens

-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click

#9 lore raw
In reply to Takashi Iwai

Takashi,

Drivers were build successfully. Thanks!

Andrew

======= On Friday 16 July 2004 17:15, you wrote: ======= At Fri, 16 Jul 2004 15:08:14 +0200,

Quoted 11 lines · I wrote:

At Fri, 16 Jul 2004 16:57:23 +0400,

Quoted 7 lines · Andrew Gaydenko

clean/(re)cvscompile were done.

I'm not bash guru, but from common view some configure fragment seems strange to me. I have got

checking for new remap_page_range... "no"

Quoted 7 lines

You have an old version of kernel, don't you?

Quoted 3 lines

Here 1 is assigned to new_remap (line 2). But CONFIG_OLD_REMAP_PAGE_RANGE is defined (line 8) only when new_remap equals 0 (line 6). Must new_remap assignement be reverted?

It defines CONFIG_ *OLD* _REMAP_PAGE_RANGE. So it's correct.

Doh, I found you're right, the 1 and 0 are reverted.
And another silly typo. The corrected (additional) patch is below.

Takashi

-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click