Bug 58193

Summary:	getc operations via RedBoot lock the system
Product:	[Retired] eCos	Reporter:	Jonathan Larmour <jlarmour>
Component:	HAL	Assignee:	eCos bugs internal list <es-ecos-bugs-int>
Status:	CLOSED WONTFIX	QA Contact:	eCos bugs internal list <es-ecos-bugs-int>
Severity:	medium	Docs Contact:
Priority:	low
Version:	CVS
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2003-06-20 16:11:02 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 Jonathan Larmour 2002-01-10 21:33:21 UTC

Description of Problem:

If someone does a getc operation that uses the virtual vectors to call into
redboot, then the system locks up and no other threads run because interrupts
are disabled and the scheduler is locked.

Instead hal_if_diag_read_char() should call a new CYGACC_COMM_IF_GETC_NONBLOCK
operation so that the looping to detect for a character is done with interrupts
enabled at least some of the time. Slower, but at least not stopped.

Comment 1 Jesper Skov 2002-01-11 06:57:41 UTC

I disagree. If you use the polled routines provided by the VV, well, you'll
be polling. If you need an interrupt safe environment, use an IO driver.

Comment 2 Jonathan Larmour 2002-01-11 13:10:22 UTC

I can't disagree with "If you need an interrupt safe environment, use an IO
driver.", but I don't see any harm in changing this behaviour (except arguably
bloat because we can't get rid of the non-block version now it's there).

Comment 3 Alex Schuilenburg 2003-06-20 16:11:02 UTC

This bug has moved to http://bugs.ecos.sourceware.org/show_bug.cgi?id=58193