qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Deepak Gupta <debug@rivosinc.com>
To: qemu-riscv@nongnu.org, qemu-devel@nongnu.org
Cc: palmer@dabbelt.com, Alistair.Francis@wdc.com, bmeng.cn@gmail.com,
	liwei1518@gmail.com, dbarboza@ventanamicro.com,
	zhiwei_liu@linux.alibaba.com, jim.shu@sifive.com,
	andy.chiu@sifive.com, kito.cheng@sifive.com,
	Richard Henderson <richard.henderson@linaro.org>
Subject: Re: [PATCH v5 11/15] target/riscv: mmu changes for zicfiss shadow stack protection
Date: Tue, 20 Aug 2024 00:35:24 -0700	[thread overview]
Message-ID: <ZsRHPDG/GkyhdrQi@debug.ba.rivosinc.com> (raw)
In-Reply-To: <20240820000129.3522346-12-debug@rivosinc.com>

On Mon, Aug 19, 2024 at 05:01:25PM -0700, Deepak Gupta wrote:
>zicfiss protects shadow stack using new page table encodings PTE.W=0,
>PTE.R=0 and PTE.X=0. This encoding is reserved if zicfiss is not
>implemented or if shadow stack are not enabled.
>Loads on shadow stack memory are allowed while stores to shadow stack
>memory leads to access faults. Shadow stack accesses to RO memory
>leads to store page fault.
>
>To implement special nature of shadow stack memory where only selected
>stores (shadow stack stores from sspush) have to be allowed while rest
>of regular stores disallowed, new MMU TLB index is created for shadow
>stack.
>
>Signed-off-by: Deepak Gupta <debug@rivosinc.com>
>Suggested-by: Richard Henderson <richard.henderson@linaro.org>
>---
>@@ -1406,6 +1431,11 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address, int size,
>     qemu_log_mask(CPU_LOG_MMU, "%s ad %" VADDR_PRIx " rw %d mmu_idx %d\n",
>                   __func__, address, access_type, mmu_idx);
>
>+    /* If shadow stack instruction initiated this access, treat it as store */
>+    if (mmu_idx & MMU_IDX_SS_WRITE) {
>+        access_type = MMU_DATA_STORE;
>+    }
>+

I think I forgot to address this. Do you still want me to fix this up like you
had suggested?

IIRC, you mentioned to use TARGET_INSN_START_EXTRA_WORDS=2. Honestly I don't know
what it means and how its used. Based on git grep and some readup, are you expecting something
along the below lines?


"""

diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
index fee31b8037..dfd2efa941 100644
--- a/target/riscv/cpu.h
+++ b/target/riscv/cpu.h
@@ -47,7 +47,7 @@ typedef struct CPUArchState CPURISCVState;
   * RISC-V-specific extra insn start words:
   * 1: Original instruction opcode
   */
-#define TARGET_INSN_START_EXTRA_WORDS 1
+#define TARGET_INSN_START_EXTRA_WORDS 2
  
  #define RV(x) ((target_ulong)1 << (x - 'A'))
  
diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c
index f74a1216b1..b266177e46 100644
--- a/target/riscv/cpu_helper.c
+++ b/target/riscv/cpu_helper.c
@@ -1271,6 +1271,11 @@ static void raise_mmu_exception(CPURISCVState *env, target_ulong address,
  {
      CPUState *cs = env_cpu(env);
  
+     if (!pmp_violation &&
+         tcg_ctx->gen_insn_data[TARGET_INSN_START_EXTRA_WORDS-1] & 1) {
+         tcg_ctx->gen_insn_data[TARGET_INSN_START_EXTRA_WORDS-1] &= ~1;
+         access_type = MMU_DATA_STORE;
+     }
+
      switch (access_type) {
      case MMU_INST_FETCH:
          if (pmp_violation) {
@@ -1433,7 +1438,7 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address, int size,
  
      /* If shadow stack instruction initiated this access, treat it as store */
      if (mmu_idx & MMU_IDX_SS_WRITE) {
-        access_type = MMU_DATA_STORE;
+        tcg_ctx->gen_insn_data[TARGET_INSN_START_EXTRA_WORDS-1] |= 1;
      }
  
      pmu_tlb_fill_incr_ctr(cpu, access_type);
@@ -1529,6 +1534,7 @@ bool riscv_cpu_tlb_fill(CPUState *cs, vaddr address, int size,
      if (ret == TRANSLATE_SUCCESS) {
          tlb_set_page(cs, address & ~(tlb_size - 1), pa & ~(tlb_size - 1),
                       prot, mmu_idx, tlb_size);
+        tcg_ctx->gen_insn_data[TARGET_INSN_START_EXTRA_WORDS-1] &= ~1;
          return true;
      } else if (probe) {


"""

>     pmu_tlb_fill_incr_ctr(cpu, access_type);
>     if (two_stage_lookup) {
>         /* Two stage lookup */
>diff --git a/target/riscv/internals.h b/target/riscv/internals.h
>index 0ac17bc5ad..ddbdee885b 100644
>--- a/target/riscv/internals.h
>+++ b/target/riscv/internals.h
>@@ -30,12 +30,15 @@
>  *  - U+2STAGE          0b100
>  *  - S+2STAGE          0b101
>  *  - S+SUM+2STAGE      0b110
>+ *  - Shadow stack+U   0b1000
>+ *  - Shadow stack+S   0b1001
>  */
> #define MMUIdx_U            0
> #define MMUIdx_S            1
> #define MMUIdx_S_SUM        2
> #define MMUIdx_M            3
> #define MMU_2STAGE_BIT      (1 << 2)
>+#define MMU_IDX_SS_WRITE    (1 << 3)
>
> static inline int mmuidx_priv(int mmu_idx)
> {
>-- 
>2.44.0
>


  reply	other threads:[~2024-08-20  7:36 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-20  0:01 [PATCH v5 00/15] riscv support for control flow integrity extensions Deepak Gupta
2024-08-20  0:01 ` [PATCH v5 01/15] target/riscv: Add zicfilp extension Deepak Gupta
2024-08-20  0:01 ` [PATCH v5 02/15] target/riscv: Introduce elp state and enabling controls for zicfilp Deepak Gupta
2024-08-20  5:17   ` Richard Henderson
2024-08-20  0:01 ` [PATCH v5 03/15] target/riscv: save and restore elp state on priv transitions Deepak Gupta
2024-08-20  0:01 ` [PATCH v5 04/15] target/riscv: additional code information for sw check Deepak Gupta
2024-08-20  0:01 ` [PATCH v5 05/15] target/riscv: tracking indirect branches (fcfi) for zicfilp Deepak Gupta
2024-08-20  5:24   ` Richard Henderson
2024-08-20  0:01 ` [PATCH v5 06/15] target/riscv: zicfilp `lpad` impl and branch tracking Deepak Gupta
2024-08-20  5:29   ` Richard Henderson
2024-08-20  0:01 ` [PATCH v5 07/15] disas/riscv: enable `lpad` disassembly Deepak Gupta
2024-08-20  0:01 ` [PATCH v5 08/15] target/riscv: Add zicfiss extension Deepak Gupta
2024-08-20  0:01 ` [PATCH v5 09/15] target/riscv: introduce ssp and enabling controls for zicfiss Deepak Gupta
2024-08-20  5:34   ` Richard Henderson
2024-08-20  0:01 ` [PATCH v5 10/15] target/riscv: tb flag for shadow stack instructions Deepak Gupta
2024-08-20  0:01 ` [PATCH v5 11/15] target/riscv: mmu changes for zicfiss shadow stack protection Deepak Gupta
2024-08-20  7:35   ` Deepak Gupta [this message]
2024-08-20  9:20     ` Richard Henderson
2024-08-20 18:55       ` Deepak Gupta
2024-08-20 19:45         ` Deepak Gupta
2024-08-20 22:33         ` Richard Henderson
2024-08-20  0:01 ` [PATCH v5 12/15] target/riscv: implement zicfiss instructions Deepak Gupta
2024-08-20  0:01 ` [PATCH v5 13/15] target/riscv: compressed encodings for sspush and sspopchk Deepak Gupta
2024-08-20  0:01 ` [PATCH v5 14/15] disas/riscv: enable disassembly for zicfiss instructions Deepak Gupta
2024-08-20  0:01 ` [PATCH v5 15/15] disas/riscv: enable disassembly for compressed sspush/sspopchk Deepak Gupta

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ZsRHPDG/GkyhdrQi@debug.ba.rivosinc.com \
    --to=debug@rivosinc.com \
    --cc=Alistair.Francis@wdc.com \
    --cc=andy.chiu@sifive.com \
    --cc=bmeng.cn@gmail.com \
    --cc=dbarboza@ventanamicro.com \
    --cc=jim.shu@sifive.com \
    --cc=kito.cheng@sifive.com \
    --cc=liwei1518@gmail.com \
    --cc=palmer@dabbelt.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-riscv@nongnu.org \
    --cc=richard.henderson@linaro.org \
    --cc=zhiwei_liu@linux.alibaba.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).