Problem with SystemACE cmd_ace.c patch of 7 Aug 2005
I have a problem with some changes you (Jeff) did in u-boot related to compact flash. Can you help me understand why you put this:
+ /* Reset the configruation controller */ + val = ace_readw(0x18); + val|=0x0080; + ace_writew(val, 0x18);
+
into the systemace_read function? This is causing me grief, with the linux SystemACE indicating an error:
JTAG controller couldn't read configuration from the CompactFlash
while booting Linux with a ext3 partition on the SystemACE CF. When I comment those lines out of cmd_ace.c, my boot starts to work again.
I suspect this "reset" of the config portion of the systemace is causing the SystemACE to start trying again to configure when the reset is released at the end of a block read. I would so like to remove that little snippet of code, but if you had a positive reason why it should be there, we need to negotiate.
Signature · 4 lines
Steve Williams "The woods are lovely, dark and deep. steve at icarus.com But I have promises to keep, http://www.icarus.com and lines to code before I sleep, http://www.picturel.com And lines to code before I sleep."
Review
Quoted 8 lines · Stephen Williams
I have a problem with some changes you (Jeff) did in u-boot related to compact flash. Can you help me understand why you put this:
(quoted diff) +5
-0+ /* Reset the configruation controller */ + val = ace_readw(0x18); + val|=0x0080; + ace_writew(val, 0x18); +
This is directly from the the Xilinx SystemACE Manual, "Reading Sector Data from Compact Flash Control Flow Porcess". In my manual the flow diagram is Figure 23.
Quoted 9 lines
into the
systemace_readfunction? This is causing me grief, with the linux SystemACE indicating an error:JTAG controller couldn't read configuration from the CompactFlash
while booting Linux with a ext3 partition on the SystemACE CF. When I comment those lines out of
cmd_ace.c, my boot starts to work again.
Quoted 5 lines
I suspect this "reset" of the config portion of the systemace is causing the SystemACE to start trying again to configure when the reset is released at the end of a block read. I would so like to remove that little snippet of code, but if you had a positive reason why it should be there, we need to negotiate.
Like I said, this is taken directly from the Xilinx documentation.
Furthermore, Xilinx tech support also instructed me to include these steps when I was troubleshooting some earlier issues with our board design. Here is a snippet from the conversation with the Xilinx staff:
<begin snippet>
"The ERRLED sometimes sporadically goes active when the MPU interface is used to read/write sectors from/to the CompactFlash device,respectively. The problem is due to the fact that the configuration engine receives spurious interrupts that are not handled correctly. There are couple of workarounds for this problem":
1)The work-around is to hold the configuration engine in reset(CFGRESET=1)while accessing the CF device via the MPU interface. 2)Safely ignore the error led and proceed with the process.
We are aware of the problem and there is nothing much we will be able to do on this issue, the reason for the Error LED going high is because of the fact that the compactflash is an ASIC and there are issues interfacing with the Flash card.
<end snippet>
Jeff Angielski
The PTR Group
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Quoted 10 lines · Jeff Angielski
Review
Quoted 8 lines · Stephen Williams
I have a problem with some changes you (Jeff) did in u-boot related to compact flash. Can you help me understand why you put this:
(quoted diff) +5
-0+ /* Reset the configruation controller */ + val = ace_readw(0x18); + val|=0x0080; + ace_writew(val, 0x18); +
Quoted 3 lines
This is directly from the the Xilinx SystemACE Manual, "Reading Sector Data from Compact Flash Control Flow Porcess". In my manual the flow diagram is Figure 23.
Sure 'nuff, I see it there. Ugh!
So then that begs the question, why am I getting this:
Quoted 1 line
Quoted 1 line
JTAG controller couldn't read configuration from the CompactFlash
interrupting (and killing) my Linux boot? When I take that code back out, the spurious error interrupt does not happen. Do you use the SystemACE in your Linux environment after/during boot?
On my side, I'll press the issue with Xilinx and see what they have to say.
BTW, the Linux SystemACE driver that is apparently written by Xilinx does not go through this reset cycle in its block read.
- --
Steve Williams "The woods are lovely, dark and deep. steve at icarus.com But I have promises to keep, http://www.icarus.com and lines to code before I sleep, http://www.picturel.com And lines to code before I sleep."
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFENVjprPt1Sc2b3ikRAv6oAKC14ZLKpNnkPsoKuHqB5m4e9pKORQCgliBB NVaI9g6d8B1Pc8kDX5k1iF0=
=JcUO
-----END PGP SIGNATURE-----