From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:55729) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1eu9QX-0004AP-9b for qemu-devel@nongnu.org; Thu, 08 Mar 2018 23:14:44 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1eu9QW-0006xq-Gf for qemu-devel@nongnu.org; Thu, 08 Mar 2018 23:14:41 -0500 Received: from mail-pg0-x241.google.com ([2607:f8b0:400e:c05::241]:46603) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1eu9QW-0006xZ-9g for qemu-devel@nongnu.org; Thu, 08 Mar 2018 23:14:40 -0500 Received: by mail-pg0-x241.google.com with SMTP id r26so3113105pgv.13 for ; Thu, 08 Mar 2018 20:14:40 -0800 (PST) From: Michael Clark Date: Fri, 9 Mar 2018 17:12:32 +1300 Message-Id: <1520568765-58189-11-git-send-email-mjc@sifive.com> In-Reply-To: <1520568765-58189-1-git-send-email-mjc@sifive.com> References: <1520568765-58189-1-git-send-email-mjc@sifive.com> Subject: [Qemu-devel] [PATCH v2 10/23] RISC-V: Hold rcu_read_lock when accessing memory List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: Michael Clark , Sagar Karandikar , Bastian Koppelmann , Stefan O'Rear , Palmer Dabbelt >>From reading other code that accesses memory regions directly, it appears that the rcu_read_lock needs to be held. Note: the original code for accessing RAM directly was added because there is no other way to use atomic_cmpxchg on guest physical address space. Cc: Sagar Karandikar Cc: Bastian Koppelmann CC: Stefan O'Rear Signed-off-by: Michael Clark Signed-off-by: Palmer Dabbelt --- target/riscv/helper.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/target/riscv/helper.c b/target/riscv/helper.c index 02cbcea..228933c 100644 --- a/target/riscv/helper.c +++ b/target/riscv/helper.c @@ -209,6 +209,7 @@ restart: as the PTE is no longer valid */ MemoryRegion *mr; hwaddr l = sizeof(target_ulong), addr1; + rcu_read_lock(); mr = address_space_translate(cs->as, pte_addr, &addr1, &l, false); if (memory_access_is_direct(mr, true)) { @@ -222,16 +223,19 @@ restart: target_ulong old_pte = atomic_cmpxchg(pte_pa, pte, updated_pte); if (old_pte != pte) { + rcu_read_unlock(); goto restart; } else { pte = updated_pte; } #endif } else { + rcu_read_unlock(); /* misconfigured PTE in ROM (AD bits are not preset) or * PTE is in IO space and can't be updated atomically */ return TRANSLATE_FAIL; } + rcu_read_unlock(); } /* for superpage mappings, make a fake leaf PTE for the TLB's -- 2.7.0