summary refs log tree commit diff stats
path: root/results/scraper/fex/4150
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/4150
parent63d2e9d409831aa8582787234cae4741847504b7 (diff)
downloadqemu-analysis-main.tar.gz
qemu-analysis-main.zip
add downloaded fex bug-reports HEAD main
Diffstat (limited to 'results/scraper/fex/4150')
-rw-r--r--results/scraper/fex/415012
1 files changed, 12 insertions, 0 deletions
diff --git a/results/scraper/fex/4150 b/results/scraper/fex/4150
new file mode 100644
index 000000000..3afff0262
--- /dev/null
+++ b/results/scraper/fex/4150
@@ -0,0 +1,12 @@
+FEX tests may generate unwanted instruction form
+I am not 100% sure if I am right here, so apologies if I am not.

+I was looking at test `OpSize/66_D6.asm` which has the instruction:

+`movq xmm0, xmm2`

+This on NASM generates the (probably unintended) form:

+`movq xmm0, xmm2  ; +0 = f3 0f 7e c2` (https://godbolt.org/z/ebWchEcbb)

+While I assume the one that the test focuses on testing is the form:

+`movq xmm0, xmm2  ; +0 = 66 0f d6 d0`

+

+I assume this doesn't cause any real problems, but I thought you might want to know that the test doesn't test the reg/reg form of the D6 opcode (if I am right).

+

+This may or may not also happen on other tests that have both a xmm/xmm128 and a xmm128/xmm form where NASM compiles them to the unintended form.
\ No newline at end of file