target/riscv: Fix PMP range boundary address bug
authorDayeol Lee <dayeol@berkeley.edu>
Fri, 17 May 2019 22:10:58 +0000 (15:10 -0700)
committerPalmer Dabbelt <palmer@sifive.com>
Mon, 24 Jun 2019 06:44:41 +0000 (23:44 -0700)
commit49db9fa1fd7c252596b53cf80876e06f407d09ed
tree69e2619c1889d2aa3baeeca54faa1ba1e7aab3d7
parentb9d1848ebe8f083482307be725cca7883fca7566
target/riscv: Fix PMP range boundary address bug

A wrong address is passed to `pmp_is_in_range` while checking if a
memory access is within a PMP range.
Since the ending address of the pmp range (i.e., pmp_state.addr[i].ea)
is set to the last address in the range (i.e., pmp base + pmp size - 1),
memory accesses containg the last address in the range will always fail.

For example, assume that a PMP range is 4KB from 0x87654000 such that
the last address within the range is 0x87654fff.
1-byte access to 0x87654fff should be considered to be fully inside the
PMP range.
However the access now fails and complains partial inclusion because
pmp_is_in_range(env, i, addr + size) returns 0 whereas
pmp_is_in_range(env, i, addr) returns 1.

Signed-off-by: Dayeol Lee <dayeol@berkeley.edu>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Michael Clark <mjc@sifive.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
target/riscv/pmp.c