summary refs log tree commit diff stats
path: root/results/scraper/fex/806
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/806
parent63d2e9d409831aa8582787234cae4741847504b7 (diff)
downloadqemu-analysis-main.tar.gz
qemu-analysis-main.zip
add downloaded fex bug-reports HEAD main
Diffstat (limited to 'results/scraper/fex/806')
-rw-r--r--results/scraper/fex/8066
1 files changed, 6 insertions, 0 deletions
diff --git a/results/scraper/fex/806 b/results/scraper/fex/806
new file mode 100644
index 000000000..3592aa67b
--- /dev/null
+++ b/results/scraper/fex/806
@@ -0,0 +1,6 @@
+Code Flushing: Remove stale code from all theards
+Follow up from #705 

+

+"Derek Bruening's [PhD thesis](http://www.burningcutlery.com/derek/docs/phd.pdf) describes a system for invalidating code blocks (called "fragments") across multiple thread. The relevant part starts on page 156.

+

+However it is quite complex: you need to be able to hotpatch branch instructions and update lookup tables while another thread is executing translated code. At that point it might be worth simply switching to a single shared code cache, which has many other benefits. See [this paper](http://www.burningcutlery.com/derek/docs/threadshared-CGO06.pdf) for details on implementing a thread-shared code cache." (via @Amanieu)
\ No newline at end of file