search

uwcms / cms-calo-layer1

Xilinx Microblaze Projects for RCT Upgrade
2 stars 8 forks source link

WIP get SPI echo test code working #29

Closed ekfriis closed 11 years ago

ekfriis commented 11 years ago

Hi @dabelknap

I'm just collecting all the fixes I do to the SPI echo tests in this PR before merging, as there are some common issues you should be aware of. Most of this is copying over bugs we killed on the CTP6 standalone code from the August integration test that I never ported to the other program.s

Infos practiques:

If suddenly the programs won't fit in the bram, you can set GCC to optimize for size instead of speed, see (da656b7).

Note that when setting up the interrupts, you have to also enable the Xilinx exception table (0b4e89e). I don't know why this is necessary, but it is.

ekfriis commented 11 years ago

You need to lookup the correct device from the xparameters.h file in the corresponding BSP (497327d)

dabelknap commented 11 years ago

Some address strangeness in FE SPI echo test:

XMD% stp
    653C:   B0000000  imm      0
    6540:   30E03970  addik    r7 , r0 , 14704

XMD% stp
    6544:   B0000000  imm      0
    6548:   31007164  addik    r8 , r0 , 29028

XMD% stp
    654C:   B000FFFF  imm      -1
    6550:   B9F4A320  brlid    r15, -23776
    6554:   10D30000  addk     r6 , r19, r0

XMD% stp
     870:   DEADBEEF

870 is supposed to read 870: 3021ffe0 addik r1, r1, -32 in <XIntc_Connect>.

ekfriis commented 11 years ago

How do you know it's supposed to be that? From the .S?

On Fri, Oct 18, 2013 at 2:34 PM, Austin Belknap notifications@github.comwrote:

Some address strangeness in FE SPI echo test:

XMD% stp 653C: B0000000 imm 0 6540: 30E03970 addik r7 , r0 , 14704

XMD% stp 6544: B0000000 imm 0 6548: 31007164 addik r8 , r0 , 29028

XMD% stp 654C: B000FFFF imm -1 6550: B9F4A320 brlid r15, -23776 6554: 10D30000 addk r6 , r19, r0

XMD% stp 870: DEADBEEF

870 is supposed to read 870: 3021ffe0 addik r1, r1, -32 in .

— Reply to this email directly or view it on GitHubhttps://github.com/uwcms/cms-calo-layer1/pull/29#issuecomment-26592002 .

dabelknap commented 11 years ago

For posterity: yes, it is from payload.S