This patchset fixes a couple of bugs in the arm ptw code that were
inadvertently introduced by the recent changes to support FEAT_RME.
One is with the handling of debug accesses in Secure state with
EL2 disabled, reported by Jean-Philippe. The other is one I noticed
while reading the code where if VSTCR.SA/VTCR.NSA require the
output of a two-stage translation to be squashed to NS then we
don't also set the f.attrs.space accordingly.
I also think we need to refactor the ptw code to stop passing bool
in_secure around (both directly as function arguments and in the
S1Translate struct) and instead always use an ARMSecuritySpace.
However I don't think that will affect any use cases that don't
turn on the (experimental) FEAT_RME support, so it's not so
urgent as this bugfix.
Patch 1 is just documenting the meaning of some of the S1Translate
struct fields, since I've now had to think through this twice and
hope to avoid doing so a third time :-)
thanks
-- PMM
Peter Maydell (3):
target/arm/ptw.c: Add comments to S1Translate struct fields
target/arm: Fix S1_ptw_translate() debug path
target/arm/ptw.c: Account for FEAT_RME when applying {N}SW,SA bits
target/arm/ptw.c | 90 ++++++++++++++++++++++++++++++++++++++++++------
1 file changed, 80 insertions(+), 10 deletions(-)
--
2.34.1
