Re: kern/50613: unpausing audio(4) causes already written data to be dropped

[Kamil Rytarowski <n54%gmx.com@localhost>]

The following reply was made to PR kern/50613; it has been noted by GNATS.

From: Kamil Rytarowski <n54%gmx.com@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: kern/50613: unpausing audio(4) causes already written data to be
 dropped
Date: Mon, 4 Jan 2016 22:57:33 +0100

 -----BEGIN PGP SIGNED MESSAGE-----
 Hash: SHA256
 
 On 04.01.2016 22:50, Onno van der Linden wrote:
 > The following reply was made to PR kern/50613; it has been noted
 > by GNATS.
 > 
 > From: Onno van der Linden <o.vd.linden%quicknet.nl@localhost> To: 
 > gnats-bugs%NetBSD.org@localhost Cc: Subject: Re: kern/50613: unpausing 
 > audio(4) causes already written data to be dropped Date: Mon, 4
 > Jan 2016 21:27:26 +0100
 > 
 > And a test case
 > 
 > #include <sys/audioio.h> #include <stdio.h> #include <fcntl.h> 
 > #include <string.h>
 > 
 > main() { int     fd; char    buf[1024]; unsigned long	wseek; 
 > audio_offset_t  ao; struct audio_info       ai;
 > 
 > memset(buf,'\0',sizeof buf); AUDIO_INITINFO(&ai); 
 > fd=open("/dev/audio",O_WRONLY,0); ai.play.pause=1; 
 > ai.blocksize=1024; ioctl(fd, AUDIO_SETINFO, &ai); ioctl(fd, 
 > AUDIO_GETOOFFS, &ao); printf("After open/pause samples=%u 
 > deltablks=%u offset=%u\n",ao.samples,ao.deltablks, ao.offset); 
 > write(fd,buf,sizeof buf); ioctl(fd, AUDIO_GETOOFFS, &ao); 
 > printf("After write samples=%u deltablks=%u 
 > offset=%u\n",ao.samples,ao.deltablks, ao.offset); 
 > AUDIO_INITINFO(&ai); ai.play.pause=0; ioctl(fd, AUDIO_SETINFO, 
 > &ai); ioctl(fd, AUDIO_GETOOFFS, &ao); printf("After unpause 
 > samples=%u deltablks=%u offset=%u\n",ao.samples,ao.deltablks, 
 > ao.offset); write(fd,buf,sizeof buf); ioctl(fd, AUDIO_DRAIN); 
 > ioctl(fd, AUDIO_GETOOFFS, &ao); printf("After write samples=%u 
 > deltablks=%u offset=%u\n",ao.samples,ao.deltablks, ao.offset); 
 > close(fd); }
 > 
 > Run straight after booting an unpatched kernel:
 > 
 > After open/pause samples=0 deltablks=0 offset=1024 After write 
 > samples=1024 deltablks=1 offset=2048 After unpause samples=0 
 > deltablks=0 offset=2048 After write samples=1024 deltablks=1 
 > offset=3072
 > 
 > 
 > Run straight after booting a kernel with the patch: After 
 > open/pause samples=0 deltablks=0 offset=1024 After write 
 > samples=1024 deltablks=1 offset=2048 After unpause samples=1024 
 > deltablks=0 offset=2048 After write samples=2048 deltablks=1 
 > offset=3072
 > 
 > samples=0 after unpause vs. samples=1024 after unpause
 > 
 > 
 
 
 Thank you. If that's not too much. Can you try to write a new test or
 set of tests to ensure it works and will work in the future for everybod
 y?
 
 This patch makes sense.
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2
 
 iQIcBAEBCAAGBQJWiurLAAoJEEuzCOmwLnZsnn8QAMM5REaue6Mu0brWL8j9Yk51
 K5aPxO3eyFKtYI54YgX2KOowG0PPdrjAMsR8HGrQlhD2ego50k7m9E8rxxQ0I3fc
 36V8HIcouG6U4MgK4JN3e5FqlHH9fniEwVkkZ+EW09zVXkB+FEy/2/JyZh7V9fwo
 VSsN7yK71q8zev7mVjkoxnd6jNW9MoP/VKMyxg/ffaNkeqZqmuJ4q8A2JiGoiR81
 3kxrodD5JkBeSqjD1N52iptf0aFviCsGq8cfxIG+xtrfPUP6L9z/y+iDoy3a8ReD
 DhcRrYVsUozHFBlaIhTGjh2UVOX0zWNl0ReXdJGeyyo+c/44xkZIpsN+5dGFPiFP
 YIcY7BdMca989Mva0EI7qIE8kfueP2ItiMQa1TknVyD1QnDruJ+xPgP1emkO8CGi
 RGkNfWeM5ng2/YhxDKRUYlVSsbDTGoTxeCrwbLMgcnNDGtSOq4IiyelV8LC5b4kG
 I1jTb8fkOJd0JqnwMDEi3NVBZ8LtMhadMejiMwUWV94AHSFoRZuJ7gQ4JyBp5noh
 14CiQRDbbYuybAAMWQUNbGwt8x3Hd5KY/FYWqLbMI2lEERLiTy8Xg6LWzqQsey7i
 BJTRBoL/xiIEE6ZF+rUNsBGvJYVnEH1luHjcpC7kbmPCDzOD+uNRiVxH2erfv/JL
 SKLXKgZTYSnLReur41Qq
 =fZra
 -----END PGP SIGNATURE-----