aboutsummaryrefslogtreecommitdiff
path: root/docs
diff options
context:
space:
mode:
authorPavel Pisa <pisa@cmp.felk.cvut.cz>2019-03-16 21:18:48 +0100
committerPavel Pisa <pisa@cmp.felk.cvut.cz>2019-03-16 21:18:48 +0100
commit7a40807e86aa0d16714a8bf854075c739d52ca40 (patch)
tree310831ebc665e6951b7bd352d52f3126b454ea17 /docs
parent371e31948fb3bc01359b24a7f325d3e1d149f98b (diff)
downloadqtmips-7a40807e86aa0d16714a8bf854075c739d52ca40.tar.gz
qtmips-7a40807e86aa0d16714a8bf854075c739d52ca40.tar.bz2
qtmips-7a40807e86aa0d16714a8bf854075c739d52ca40.zip
Update readme - include how CACHE and SYNCI instructions are implemented.
Signed-off-by: Pavel Pisa <pisa@cmp.felk.cvut.cz>
Diffstat (limited to 'docs')
-rw-r--r--docs/exec-formats-and-tools.md4
1 files changed, 2 insertions, 2 deletions
diff --git a/docs/exec-formats-and-tools.md b/docs/exec-formats-and-tools.md
index 85367c5..e2f29f3 100644
--- a/docs/exec-formats-and-tools.md
+++ b/docs/exec-formats-and-tools.md
@@ -83,8 +83,8 @@ instructions in the pipeline are discarded and execution stops.
PC is set to the address of instruction causing the exception
or to the branch instruction address if the exception occurs in
delay slot. When the single step or continuous execution
-is requested again then the "hardware" breakpoint exception
-in the fetch stage is masked for the first executed instruction which.
+is requested again then the "hardware" breakpoint exception
+in the fetch stage is masked for the first executed instruction.
But then CPU accepts breakpoint exceptions again. This is why it
is not a good idea to set up breakpoint to address of an instruction
in the delay slot.