| 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 <2009> 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 | Index | 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 <2009> 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 |
| <== Date ==> | <== Thread ==> |
|---|
| Subject: | Re: RTEMS-mvme5500 BSP update |
| From: | Kate Feng <[email protected]> |
| To: | Joel Sherrill <[email protected]> |
| Cc: | "[email protected]" <[email protected]>, "[email protected]" <[email protected]> |
| Date: | Thu, 19 Mar 2009 11:23:01 -0400 |
Joel Sherrill wrote:
Any chance of getting a patch to apply to the RTEMS source base?
Yes, I can produce a patch for it and file a PR. My questions are: 1) Do you want the patch for RTEMS 4.9.1 only or for both 4.9.1 and 4.8.0 ? 2) There were lots of change, which were not related to any bug. It was just an attempt to be consistent with the shared code in the PowerPC, which was originally written by Eric Valette. Thus, I thought it would be easier if RTEMS takes the entire tar file. If you do not think so, then I will produce a patch. The patch would appear overwhelmed, but the major updates are really only the three things (e.g. bug fixes and PCI performance), that were mentioned in my previous E-mail. Please let me know what you think. -- Kate
--joel sherrill RTEMS Kate Feng wrote:Hi, The stable version of RTEMS-mvme5500 BSP is posted at http://www.nsls.bnl.gov/facility/expsys/software/EPICS/ The "RTEMS-mvme5500 BSP v1.4.9.1", was tested under RTEMS 4.9.1, and the "RTEMS-mvme5500 BSP v1.4.8.0" was tested under RTEMS 4.8.0.. The highlights of the change are : 1) Use the GT_GPP_Value register (0xf104) instead of the GT_GPP_Interrupt_Cause register (0xf108) to monitor the cause of the level sensitive interrupts. See irq/BSP_irq.c. This is an unique solution which solves part of the bugs in the BSP, which especially affects the 1GHz network driver. 2) There is no change in the 100 MHz network driver. Thus, its performance stays the same. Some bugs were fixed in the 1GHz network drivers (e.g. if_wm.c).3)Enable "PCI Read Agressive Prefetch", "PCI Read Line Agressive Prefetch",and "PCI Read Multiple Agressive Prefetch" to improve the performance ofthe PCI based applications (e.g. 1GHz network driver). See pci/pci_interface.c.However, performance is driven by applications. There is no known bug at this point. Any feedback which is related to the RTEMS-mvme5500 BSP would be highly appreciated. Hope this helps, Kate _______________________________________________ rtems-users mailing list [email protected] http://rtems.rtems.org/mailman/listinfo/rtems-users