whitequark
dd829afebd
firmware: fix another TOCTTOU race in sync/async RPC code.
2018-11-12 15:42:07 +00:00
whitequark
583bba8777
Revert "firmware: workaround for RPC failures"
...
This reverts commit 59033d2588
2018-11-12 15:36:36 +00:00
whitequark
0edae64afb
firmware: fix TOCTTOU race in sync/async RPC code.
...
Before this commit, the main loop in session code was laid like:
1. process_kern_queued_rpc
2. process_host_message
3. process_kern_message
If a host message (such as an RPC reply) caused the kernel to exit,
then any async RPCs would not complete, since RunFinished immediately
shuts down the kernel.
Fix this by reordering 1 and 2.
2018-11-12 15:30:59 +00:00
59033d2588
firmware: workaround for RPC failures
2018-11-12 19:51:54 +08:00
84a6b3d09b
runtime: fix DMA recording after now-pinning
2018-11-10 14:14:55 +08:00
1f7858b80b
test/dsp: fix rtio_output
2018-11-09 22:11:44 +08:00
e509ab8553
test/dsp: use absolute import path
...
Avoids "ImportError: attempted relative import with no known parent package"
when doing a simple "python -m unittest test_XXX.py".
2018-11-09 22:10:46 +08:00
c990b5e4f1
Merge remote-tracking branch 'origin/master' into new
2018-11-08 20:21:56 +08:00
a0cc7311ad
test: tighten test_pulse_rate
2018-11-08 20:17:55 +08:00
0bee43aa58
sawg: use new rtio_output() API
2018-11-08 20:16:30 +08:00
bec25cbaa0
suservo: use new rtio_output() API
2018-11-08 20:13:14 +08:00
e8d58b35b4
spi2: use new rtio_output() API
2018-11-08 20:12:30 +08:00
d18546550e
grabber: use new rtio_output() API
2018-11-08 19:15:50 +08:00
2549e623c1
ad9914: use new rtio_output() API
2018-11-08 19:15:44 +08:00
David Nadlinger
9740032a94
firmware: Fix dma_record_output_wide
2018-11-08 11:06:43 +00:00
f74dda639f
drtio: 8-bit address
2018-11-08 18:36:20 +08:00
8caea0e6d3
gateware,runtime: optimize RTIO kernel interface further
...
* now pinning (TODO: atomicity)
* for inputs, merge request and timeout registers
2018-11-08 18:29:24 +08:00
fcb611d1d2
test_ad9910: don't expect large SYNC_IN delay margins
...
sinara-hw/Urukul#16
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-07 18:18:35 +01:00
aadf5112b7
rtio: remove incorrect comment
2018-11-08 00:02:44 +08:00
fae95e73ad
ttl: use optimized rtio_output API
2018-11-07 23:41:43 +08:00
3d0c3cc1cf
gateware,runtime: optimize RTIO output interface
...
* reduce address to 8 bits
* merge core, channel and address into 32-bit pre-computable "target"
* merge we register into data register
2018-11-07 23:39:58 +08:00
e6efe830c4
ad9910: rewire sync delay tuning
...
* search from wide window end
* decouple margins and minimum window size
* add note about kasli jitter
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-07 14:52:03 +00:00
ad0254c17b
Merge branch 'switching125' into new
2018-11-07 22:03:18 +08:00
efd735a6ab
Revert "drtio: monitor RTIOClockMultiplier PLL ( #1155 )"
...
This reverts commit 469a66db61
2018-11-07 22:01:03 +08:00
6c00ab57c0
test_ad9910: relax SYNC window
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-06 19:31:38 +01:00
172633c7da
test_ad9910: default to a useful seed
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-06 17:35:57 +01:00
0b2661a34d
ad9910: robustify SYNC window finding
...
don't integrate too long, find the window tip fast and early
a couple 100 SYNC pulses are sufficient
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-06 12:41:21 +00:00
ba4bf6e59b
kasli: don't pass rtio pll feedback through bufg
...
UG472: "The MMCM performance increases because the
feedback clock is not subjected to noise on the core supply since it
never passes through a block powered by this supply."
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-06 11:58:55 +00:00
b6e4961b0f
kasli: lower RTIO clock jitter
...
* high bandwidth since the si5324 is good
* no low power ibufgds
* drop bufg between ibufgds and pll
* increase pll vco frequency to 1.5 GHz
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-06 11:43:19 +00:00
e17e458c58
ptb2: add sync to urukul0 for ad9910 usage
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-06 10:06:51 +00:00
73b7124091
test_ad9910: print sync scan for debugging
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-06 10:04:21 +01:00
9a3d81ffee
kasli: fix tester clk_sel
2018-11-06 14:49:21 +08:00
fb12df7e01
Revert "kasli_tester: urukul0 mmcx clock defunct"
...
This reverts commit 68220c316d
2018-11-06 14:33:21 +08:00
31f68ddf6c
Merge branch 'urukul-sync'
...
* urukul-sync: (29 commits)
urukul: flake8 [nfc]
ad9910: flake8 [nfc]
urukul/ad9910 test: remove unused import
test_urukul: relax speed
urukul,ad9910: print speed metrics
kasli: add PTB2 (external clock and SYNC)
kasli: add sync to LUH, HUB, Opticlock
kasli_tester: urukul0 mmcx clock defunct
test_ad9910: relax ifc mode read
tests: add Urukul-AD9910 HITL unittests including SYNC
ad9910: add init bit explanation
test: add Urukul CPLD HITL tests
ad9910: fiducial timestamp for tracking phase mode
ad9910: add phase modes
ad9910: fix pll timeout loop
tester: add urukul sync
ptb: back out urukul-sync
ad9910: add IO_UPDATE alignment and tuning
urukul: set up sync_in generator
ad9910: add io_update alignment measurement
...
close #1143
2018-11-05 19:54:30 +01:00
6fb18270a2
urukul: flake8 [nfc]
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:45:24 +01:00
832690af9a
ad9910: flake8 [nfc]
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:44:51 +01:00
6d525e2f9a
urukul/ad9910 test: remove unused import
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:40:57 +01:00
36c5a7cd04
test_urukul: relax speed
...
works fine at < 3µs here but needs <5 µs on buildbot-kasli-tester
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:37:20 +01:00
89fecfab50
urukul,ad9910: print speed metrics
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:37:18 +01:00
32d538f72b
kasli: add PTB2 (external clock and SYNC)
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:37:16 +01:00
d8a5951a13
kasli: add sync to LUH, HUB, Opticlock
...
for #1143 , also add missing LUH device db
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:37:14 +01:00
68220c316d
kasli_tester: urukul0 mmcx clock defunct
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:37:11 +01:00
89fadab63d
test_ad9910: relax ifc mode read
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:37:08 +01:00
f522e211ba
tests: add Urukul-AD9910 HITL unittests including SYNC
...
for #1143
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:37:06 +01:00
9fb850ae75
ad9910: add init bit explanation
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:37:02 +01:00
bc04da15c5
test: add Urukul CPLD HITL tests
...
for #1143
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:37:00 +01:00
141cc7d99f
ad9910: fiducial timestamp for tracking phase mode
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:36:58 +01:00
2f6d3f79ff
ad9910: add phase modes
...
* simplified and cross-referenced the explanation of the different
phase modes.
* semantically and functionally merged absolute and tracking/coherent
phase modes.
* simplified numerics to calculate phase correction
* added warning about possible inconsistency with DMA and default
phase mode
* restricted __all__ imports
* moved continuous/relative phase offset tracking from an instance
variable to a "handle" returned by set()/set_mu() in order to avoid
state inconsistency with DMA (#1113 #1115 )
for #1143
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:36:55 +01:00
d3ad2b7633
ad9910: fix pll timeout loop
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:36:53 +01:00
4269d5ad5c
tester: add urukul sync
...
Signed-off-by: Robert Jördens <rj@quartiq.de>
2018-11-05 19:36:52 +01:00
