summary refs log tree commit diff stats
path: root/results/scraper/fex/3713
diff options
context:
space:
mode:
authorChristian Krinitsin <mail@krinitsin.com>2025-07-17 09:10:43 +0200
committerChristian Krinitsin <mail@krinitsin.com>2025-07-17 09:10:43 +0200
commitf2ec263023649e596c5076df32c2d328bc9393d2 (patch)
tree5dd86caab46e552bd2e62bf9c4fb1a7504a44db4 /results/scraper/fex/3713
parent63d2e9d409831aa8582787234cae4741847504b7 (diff)
downloadqemu-analysis-main.tar.gz
qemu-analysis-main.zip
add downloaded fex bug-reports HEAD main
Diffstat (limited to 'results/scraper/fex/3713')
-rw-r--r--results/scraper/fex/371313
1 files changed, 13 insertions, 0 deletions
diff --git a/results/scraper/fex/3713 b/results/scraper/fex/3713
new file mode 100644
index 000000000..b55936abc
--- /dev/null
+++ b/results/scraper/fex/3713
@@ -0,0 +1,13 @@
+Eliminate adr+str InlineJITBlockHeader
+What is the inlinejitblockheader for exactly?

+

+> It's what we use for RIP reconstruction

+

+Why can't we keep a table of host pc --> guest rip and just look at the regular backtrace?

+

+> Once we support deferring async signals inside the JIT, we can

+> I also did some benchmarking to see the overhead for the adr+str and it fell within noise. It's single cycle + single cycle so it is hard to beat

+

+But on my end multiblock sped up bytemark a lot and I don't think DSE is the reason 😉

+

+Tracker for later.
\ No newline at end of file