This is the mail archive of the
mailing list for the eCos project.
Re: Adjustable timeout for Intel flash operations
- From: Jonathan Larmour <jifl at jifvik dot org>
- To: Anthony Tonizzo <atonizzo at gmail dot com>
- Cc: ecos-patches at sourceware dot org
- Date: Fri, 17 Nov 2006 18:11:16 +0000
- Subject: Re: Adjustable timeout for Intel flash operations
- References: <email@example.com>
Anthony Tonizzo wrote:
The patch is entirely straightforward so I've checked it in.
The default value for CYGNUM_DEVS_FLASH_INTEL_28FXXX_TIMEOUT
is set to 50000000, per Alexander's CDL modifications. The previous
hard-coded value was 5000000. Wouldn't it be sensible to make the
two the same, so that user not aware of this change will not see a
tenfold increase in the time when they switch to a new version?
Well spotted - I indeed did not notice that, but I think the change is good
anyway. Firstly this only triggers in the unlikely event of things going
wrong; and secondly, from experience with these chips on some (fast)
platforms using admittedly the flash v2 versions of these drivers, the
timeouts do need to be of the magnitude that Alexander proposed.
--["No sense being pessimistic, it wouldn't work anyway"]-- Opinions==mine