Bug 60470

Summary:	(SCSI DC395)playback loops like a stuck record when checking for SCSI scanners
Product:	[Retired] Red Hat Linux	Reporter:	Tethys <sta040>
Component:	kernel	Assignee:	Arjan van de Ven <arjanv>
Status:	CLOSED WORKSFORME	QA Contact:
Severity:	medium	Docs Contact:
Priority:	medium
Version:	7.2	CC:	alan, notting
Target Milestone:	---
Target Release:	---
Hardware:	i386
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2003-06-09 07:27:06 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Tethys 2002-02-28 00:25:09 UTC

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.7) Gecko/20011221

Description of problem:
Bizarre one, this. If you check for scanners with "scanimage -L"
when xmms is playing a track, it'll repeat a short loop like a
stuck record while scanimage checks for SCSI devices. Oddly enough,
all scanimage is doing at this point is parsing /proc/scsi/scsi.
If it was actually scanning the SCSI bus, then I might have expected
a dropout or something, but not a loop like this. The only thing I
can think of is that xmms is reading into a ring buffer which is not
being populated quickly enough to keep up with playback (although
I haven't checked the source, so I could be wildly off base here).

Version-Release number of selected component (if applicable):


How reproducible:
Always

Steps to Reproduce:
1. Play a track
2. Run scanimage -L
	

Actual Results:  Playback loops briefly

Expected Results:  Playback should be unaffected by scanimage

Additional info:

This is a problem when playing both oggs and mp3s. It also only
occurs when there's a valid scanner detected. If no devices are
found, playback in unaffected

Comment 1 Bill Nottingham 2002-03-04 19:28:23 UTC

What sound driver?

Comment 2 Tethys 2002-03-14 19:36:32 UTC

xmms output plugin: OSS Driver 1.2.5
kernel driver: es1371

lspci gives:

00:12.0 Multimedia audio controller: Ensoniq ES1371 [AudioPCI-97] (rev 08)
        Subsystem: Ensoniq Creative Sound Blaster AudioPCI64V, AudioPCI128
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr-
Stepping- SERR+ FastB2B-
        Status: Cap+ 66Mhz- UDF- FastB2B- ParErr- DEVSEL=slow >TAbort- <TAbort+
<MAbort+ >SERR- <PERR-
        Latency: 64 (3000ns min, 32000ns max)
        Interrupt: pin A routed to IRQ 12
        Region 0: I/O ports at da00 [size=64]
        Capabilities: [dc] Power Management version 1
                Flags: PMEClk- DSI+ D1- D2+ AuxCurrent=0mA
PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 PME-Enable- DSel=0 DScale=0 PME-

Comment 3 Alan Cox 2003-06-08 01:47:02 UTC

It is indeed a ring buffer, and some of the old isa scsi cards burn enough cpu
to cause this

Comment 4 Tethys 2003-06-08 16:29:05 UTC

Is it *only* ISA cards that do that? I'm using a PCI card (a Tekram DC395x).
Admittedly, it's getting on a bit now, but still...

More to the point, since it's only parsing /proc/scsi/scsi, which should
already be in memory, why should the card itself have any bearing on it?

Comment 5 Alan Cox 2003-06-08 17:07:11 UTC

Thats very strange then. This happens typically with the cheap 8bit controllers
that come with scanners. The /proc code does talk to the controller and use CPU,
but on the 395 this should not be a factor at all

Comment 6 Tethys 2003-06-09 07:27:06 UTC

Of course, it's now working perfectly for me. D'oh! I see that when
I originally logged the bug, I was running RH7.2. I'm running RH8
now, and can't reproduce the problem. I'm guessing that something
changed in either the kernel or xmms between versions that fixed
this...

Comment 7 Alan Cox 2003-06-09 11:33:11 UTC

That makes both of us happy 8)