On 09.09.2024 12:57, ichthyo wrote:
On 08.09.24 11:17, Kristian Amlie wrote:
.... but I have to narrow it down before I can make anything like a bug
report out of it, which might take me a few days.
On 08.09.24 16:31, ichthyo wrote:
otherwise optimised build). If you're able to produce a core dump from
such a
build, and extract a stack trace, this would get us a big step ahead.
Yesterday I wrote a lot, thus I'd like to emphasise this point:
before you start stripping down your project, which could be a lot
of work, maybe we could focus first on collecting info about the crash
and possibly extract a stack trace.
Experience shows that many "simple" bugs can be guessed, once the
exact location are known or hints are available regarding the context.
Good point, I went for that approach instead. 🙂
This is the backtrace I got:
----------------
#0 0x000000003bbc24be in ()
#1 0x00007fff23c881f3 in Filter::setfreq_and_q(float, float)
(this=0x7fff800478d0, frequency=7393.34619, q_=0.740651846) at
/home/kristian/code/yoshimi/src/DSP/Filter.cpp:107
#2 0x00007fff23decfc0 in ADnote::computeWorkingParameters()
(this=0x7fff800461b0) at
/home/kristian/code/yoshimi/src/Synth/ADnote.cpp:1507
#3 0x00007fff23df0fd5 in ADnote::noteout(float*, float*)
(this=0x7fff800461b0, outl=0x442c8cf0, outr=0x443c0830) at
/home/kristian/code/yoshimi/src/Synth/ADnote.cpp:2359
#4 0x00007fff23cec23e in Part::ComputePartSmps() (this=0x73da8420) at
/home/kristian/code/yoshimi/src/Misc/Part.cpp:1060
#5 0x00007fff23d00b6d in SynthEngine::MasterAudio(float**, float**,
int) (this=0x6209a170, outl=0x7fff1a0bc9f0, outr=0x7fff1a0bcc00,
to_process=265) at /home/kristian/code/yoshimi/src/Misc/SynthEngine.cpp:2187
#6 0x00007fff23c7b075 in YoshimiLV2Plugin::process(unsigned int)
(this=0x61fe7040, sample_count=265) at
/home/kristian/code/yoshimi/src/LV2_Plugin/YoshimiLV2Plugin.cpp:246
#7 0x00007fff23c7c079 in YoshimiLV2Plugin::run(void*, unsigned int)
(h=0x61fe7040, sample_count=265) at
/home/kristian/code/yoshimi/src/LV2_Plugin/YoshimiLV2Plugin.cpp:423
...
----------------
To me, values appear to be correct at this level.
Except for the innermost frame. I'm not sure what's going on there, it
should be in the source tree as well. But maybe that's exactly where it
goes wrong and why the address seems screwed up there.
Since I didn't otherwise see anything wrong with the backtrace memory
itself, I went ahead and cast this->filter to AnalogFilter, which I'm
99% sure that it is (I have no memory of using any of the other filters
anywhere), which yielded.
----------------
(gdb) p *(AnalogFilter *)filter
$5 = {<Filter_> = {_vptr.Filter_ = 0x7fff800a7cc0, outgain =
-2.25171848e-40}, static MAX_TYPES = 9, x = {_M_elems = {{c1 =
4.59163468e-41, c2 = 0.00797357224}, {c1 = 0.0079896003, c2 =
0.00804527663}, {
c1 = 0, c2 = 0}, {c1 = 0, c2 = 0}, {c1 = 0, c2 = 0}, {c1 = 0,
c2 = 0}}}, y = {_M_elems = {{c1 = 0.0079896003, c2 = 0.00804527663}, {c1
= 0.00803724397, c2 = 0.00810743775}, {c1 = 0, c2 = 0}, {c1 = 0,
c2 = 0}, {c1 = 0, c2 = 0}, {c1 = 0, c2 = 0}}}, oldx = {_M_elems
= {{c1 = 0, c2 = 0}, {c1 = 0, c2 = 0}, {c1 = 0, c2 = 0}, {c1 = 0, c2 =
0}, {c1 = 0, c2 = 0}, {c1 = 0, c2 = 0}}}, oldy = {_M_elems = {{
c1 = 0, c2 = 0}, {c1 = 0, c2 = 0}, {c1 = 0, c2 = 0}, {c1 = 0,
c2 = 0}, {c1 = 0, c2 = 0}, {c1 = 0, c2 = 0}}}, type = 2, stages = 1,
freq = 8511.19043, q = 0.740651846, gain = 1.12201846, order = 2, c = {
_M_elems = {0.19894433, 0.39788866, 0.19894433}}, d = {_M_elems =
{0, 0.429530174, -0.22530748}}, oldc = {_M_elems = {0, 0, 0}}, oldd =
{_M_elems = {0, 0, 0}}, needsinterpolation = false,
firsttime = false, abovenq = false, oldabovenq = false, tmpismp =
{<std::unique_ptr<float [], std::default_delete<float []> >> =
std::unique_ptr<float []> = {get() = 0x0}, <No data fields>}, synth =
@0x130}
----------------
Looks mostly fine as well, except maybe for the last value: "synth =
@0x130" looks a bit suspicious, or?
Does it ring any bells?
--
Kristian
_______________________________________________
Yoshimi-devel mailing list
Yoshimi-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/yoshimi-devel
