diff mbox series

[v4] riscv: hmp: Add a command to show virtual memory mappings

Message ID 1565796812-25870-1-git-send-email-bmeng.cn@gmail.com
State New
Headers show
Series [v4] riscv: hmp: Add a command to show virtual memory mappings | expand

Commit Message

Bin Meng Aug. 14, 2019, 3:33 p.m. UTC
This adds 'info mem' command for RISC-V, to show virtual memory
mappings that aids debugging.

Rather than showing every valid PTE, the command compacts the
output by merging all contiguous physical address mappings into
one block and only shows the merged block mapping details.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>

---

Changes in v4:
- restore to v2, that does not print all harts's PTE, since we
  should switch to a cpu context via the 'cpu' command

Changes in v3:
- print PTEs for all harts instead of just current hart

Changes in v2:
- promote ppn to hwaddr when doing page table address calculation

 hmp-commands-info.hx       |   2 +-
 target/riscv/Makefile.objs |   4 +
 target/riscv/monitor.c     | 229 +++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 234 insertions(+), 1 deletion(-)
 create mode 100644 target/riscv/monitor.c

Comments

Bin Meng Aug. 19, 2019, 5:59 a.m. UTC | #1
On Wed, Aug 14, 2019 at 11:33 PM Bin Meng <bmeng.cn@gmail.com> wrote:
>
> This adds 'info mem' command for RISC-V, to show virtual memory
> mappings that aids debugging.
>
> Rather than showing every valid PTE, the command compacts the
> output by merging all contiguous physical address mappings into
> one block and only shows the merged block mapping details.
>
> Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
> Acked-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
> Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
>
> ---
>
> Changes in v4:
> - restore to v2, that does not print all harts's PTE, since we
>   should switch to a cpu context via the 'cpu' command
>
> Changes in v3:
> - print PTEs for all harts instead of just current hart
>
> Changes in v2:
> - promote ppn to hwaddr when doing page table address calculation
>
>  hmp-commands-info.hx       |   2 +-
>  target/riscv/Makefile.objs |   4 +
>  target/riscv/monitor.c     | 229 +++++++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 234 insertions(+), 1 deletion(-)
>  create mode 100644 target/riscv/monitor.c
>

Ping?

What's the status of this patch?

Regards,
Bin
Palmer Dabbelt Aug. 27, 2019, 11:18 p.m. UTC | #2
On Sun, 18 Aug 2019 22:59:54 PDT (-0700), bmeng.cn@gmail.com wrote:
> On Wed, Aug 14, 2019 at 11:33 PM Bin Meng <bmeng.cn@gmail.com> wrote:
>>
>> This adds 'info mem' command for RISC-V, to show virtual memory
>> mappings that aids debugging.
>>
>> Rather than showing every valid PTE, the command compacts the
>> output by merging all contiguous physical address mappings into
>> one block and only shows the merged block mapping details.
>>
>> Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
>> Acked-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
>> Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
>>
>> ---
>>
>> Changes in v4:
>> - restore to v2, that does not print all harts's PTE, since we
>>   should switch to a cpu context via the 'cpu' command
>>
>> Changes in v3:
>> - print PTEs for all harts instead of just current hart
>>
>> Changes in v2:
>> - promote ppn to hwaddr when doing page table address calculation
>>
>>  hmp-commands-info.hx       |   2 +-
>>  target/riscv/Makefile.objs |   4 +
>>  target/riscv/monitor.c     | 229 +++++++++++++++++++++++++++++++++++++++++++++
>>  3 files changed, 234 insertions(+), 1 deletion(-)
>>  create mode 100644 target/riscv/monitor.c
>>
>
> Ping?
>
> What's the status of this patch?

This is in my patch queue (for-master on github).  I'm still a bit backed up on 
email, but when I get caught back up I'll send a PR.
Bin Meng Aug. 28, 2019, 1:31 a.m. UTC | #3
Hi Palmer,

On Wed, Aug 28, 2019 at 7:18 AM Palmer Dabbelt <palmer@sifive.com> wrote:
>
> On Sun, 18 Aug 2019 22:59:54 PDT (-0700), bmeng.cn@gmail.com wrote:
> > On Wed, Aug 14, 2019 at 11:33 PM Bin Meng <bmeng.cn@gmail.com> wrote:
> >>
> >> This adds 'info mem' command for RISC-V, to show virtual memory
> >> mappings that aids debugging.
> >>
> >> Rather than showing every valid PTE, the command compacts the
> >> output by merging all contiguous physical address mappings into
> >> one block and only shows the merged block mapping details.
> >>
> >> Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
> >> Acked-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
> >> Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
> >>
> >> ---
> >>
> >> Changes in v4:
> >> - restore to v2, that does not print all harts's PTE, since we
> >>   should switch to a cpu context via the 'cpu' command
> >>
> >> Changes in v3:
> >> - print PTEs for all harts instead of just current hart
> >>
> >> Changes in v2:
> >> - promote ppn to hwaddr when doing page table address calculation
> >>
> >>  hmp-commands-info.hx       |   2 +-
> >>  target/riscv/Makefile.objs |   4 +
> >>  target/riscv/monitor.c     | 229 +++++++++++++++++++++++++++++++++++++++++++++
> >>  3 files changed, 234 insertions(+), 1 deletion(-)
> >>  create mode 100644 target/riscv/monitor.c
> >>
> >
> > Ping?
> >
> > What's the status of this patch?
>
> This is in my patch queue (for-master on github).  I'm still a bit backed up on
> email, but when I get caught back up I'll send a PR.

I double checked your git repo, and found you applied an older version
of this patch.

Please drop that, and apply this v4 one.
http://patchwork.ozlabs.org/patch/1147104/

Regards,
Bin
Palmer Dabbelt Sept. 3, 2019, 6:46 p.m. UTC | #4
On Tue, 27 Aug 2019 18:31:18 PDT (-0700), bmeng.cn@gmail.com wrote:
> Hi Palmer,
>
> On Wed, Aug 28, 2019 at 7:18 AM Palmer Dabbelt <palmer@sifive.com> wrote:
>>
>> On Sun, 18 Aug 2019 22:59:54 PDT (-0700), bmeng.cn@gmail.com wrote:
>> > On Wed, Aug 14, 2019 at 11:33 PM Bin Meng <bmeng.cn@gmail.com> wrote:
>> >>
>> >> This adds 'info mem' command for RISC-V, to show virtual memory
>> >> mappings that aids debugging.
>> >>
>> >> Rather than showing every valid PTE, the command compacts the
>> >> output by merging all contiguous physical address mappings into
>> >> one block and only shows the merged block mapping details.
>> >>
>> >> Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
>> >> Acked-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
>> >> Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
>> >>
>> >> ---
>> >>
>> >> Changes in v4:
>> >> - restore to v2, that does not print all harts's PTE, since we
>> >>   should switch to a cpu context via the 'cpu' command
>> >>
>> >> Changes in v3:
>> >> - print PTEs for all harts instead of just current hart
>> >>
>> >> Changes in v2:
>> >> - promote ppn to hwaddr when doing page table address calculation
>> >>
>> >>  hmp-commands-info.hx       |   2 +-
>> >>  target/riscv/Makefile.objs |   4 +
>> >>  target/riscv/monitor.c     | 229 +++++++++++++++++++++++++++++++++++++++++++++
>> >>  3 files changed, 234 insertions(+), 1 deletion(-)
>> >>  create mode 100644 target/riscv/monitor.c
>> >>
>> >
>> > Ping?
>> >
>> > What's the status of this patch?
>>
>> This is in my patch queue (for-master on github).  I'm still a bit backed up on
>> email, but when I get caught back up I'll send a PR.
>
> I double checked your git repo, and found you applied an older version
> of this patch.
>
> Please drop that, and apply this v4 one.
> http://patchwork.ozlabs.org/patch/1147104/

Sorry about that, it should be fixed now.
diff mbox series

Patch

diff --git a/hmp-commands-info.hx b/hmp-commands-info.hx
index c59444c..257ee7d 100644
--- a/hmp-commands-info.hx
+++ b/hmp-commands-info.hx
@@ -249,7 +249,7 @@  STEXI
 Show virtual to physical memory mappings.
 ETEXI
 
-#if defined(TARGET_I386)
+#if defined(TARGET_I386) || defined(TARGET_RISCV)
     {
         .name       = "mem",
         .args_type  = "",
diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
index b1c79bc..a8ceccd 100644
--- a/target/riscv/Makefile.objs
+++ b/target/riscv/Makefile.objs
@@ -1,5 +1,9 @@ 
 obj-y += translate.o op_helper.o cpu_helper.o cpu.o csr.o fpu_helper.o gdbstub.o pmp.o
 
+ifeq ($(CONFIG_SOFTMMU),y)
+obj-y += monitor.o
+endif
+
 DECODETREE = $(SRC_PATH)/scripts/decodetree.py
 
 decode32-y = $(SRC_PATH)/target/riscv/insn32.decode
diff --git a/target/riscv/monitor.c b/target/riscv/monitor.c
new file mode 100644
index 0000000..d725a7a
--- /dev/null
+++ b/target/riscv/monitor.c
@@ -0,0 +1,229 @@ 
+/*
+ * QEMU monitor for RISC-V
+ *
+ * Copyright (c) 2019 Bin Meng <bmeng.cn@gmail.com>
+ *
+ * RISC-V specific monitor commands implementation
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "qemu/osdep.h"
+#include "cpu.h"
+#include "cpu_bits.h"
+#include "monitor/monitor.h"
+#include "monitor/hmp-target.h"
+
+#ifdef TARGET_RISCV64
+#define PTE_HEADER_FIELDS       "vaddr            paddr            "\
+                                "size             attr\n"
+#define PTE_HEADER_DELIMITER    "---------------- ---------------- "\
+                                "---------------- -------\n"
+#else
+#define PTE_HEADER_FIELDS       "vaddr    paddr            size     attr\n"
+#define PTE_HEADER_DELIMITER    "-------- ---------------- -------- -------\n"
+#endif
+
+/* Perform linear address sign extension */
+static target_ulong addr_canonical(int va_bits, target_ulong addr)
+{
+#ifdef TARGET_RISCV64
+    if (addr & (1UL << (va_bits - 1))) {
+        addr |= (hwaddr)-(1L << va_bits);
+    }
+#endif
+
+    return addr;
+}
+
+static void print_pte_header(Monitor *mon)
+{
+    monitor_printf(mon, PTE_HEADER_FIELDS);
+    monitor_printf(mon, PTE_HEADER_DELIMITER);
+}
+
+static void print_pte(Monitor *mon, int va_bits, target_ulong vaddr,
+                      hwaddr paddr, target_ulong size, int attr)
+{
+    /* santity check on vaddr */
+    if (vaddr >= (1UL << va_bits)) {
+        return;
+    }
+
+    if (!size) {
+        return;
+    }
+
+    monitor_printf(mon, TARGET_FMT_lx " " TARGET_FMT_plx " " TARGET_FMT_lx
+                   " %c%c%c%c%c%c%c\n",
+                   addr_canonical(va_bits, vaddr),
+                   paddr, size,
+                   attr & PTE_R ? 'r' : '-',
+                   attr & PTE_W ? 'w' : '-',
+                   attr & PTE_X ? 'x' : '-',
+                   attr & PTE_U ? 'u' : '-',
+                   attr & PTE_G ? 'g' : '-',
+                   attr & PTE_A ? 'a' : '-',
+                   attr & PTE_D ? 'd' : '-');
+}
+
+static void walk_pte(Monitor *mon, hwaddr base, target_ulong start,
+                     int level, int ptidxbits, int ptesize, int va_bits,
+                     target_ulong *vbase, hwaddr *pbase, hwaddr *last_paddr,
+                     target_ulong *last_size, int *last_attr)
+{
+    hwaddr pte_addr;
+    hwaddr paddr;
+    target_ulong pgsize;
+    target_ulong pte;
+    int ptshift;
+    int attr;
+    int idx;
+
+    if (level < 0) {
+        return;
+    }
+
+    ptshift = level * ptidxbits;
+    pgsize = 1UL << (PGSHIFT + ptshift);
+
+    for (idx = 0; idx < (1UL << ptidxbits); idx++) {
+        pte_addr = base + idx * ptesize;
+        cpu_physical_memory_read(pte_addr, &pte, ptesize);
+
+        paddr = (hwaddr)(pte >> PTE_PPN_SHIFT) << PGSHIFT;
+        attr = pte & 0xff;
+
+        /* PTE has to be valid */
+        if (attr & PTE_V) {
+            if (attr & (PTE_R | PTE_W | PTE_X)) {
+                /*
+                 * A leaf PTE has been found
+                 *
+                 * If current PTE's permission bits differ from the last one,
+                 * or current PTE's ppn does not make a contiguous physical
+                 * address block together with the last one, print out the last
+                 * contiguous mapped block details.
+                 */
+                if ((*last_attr != attr) ||
+                    (*last_paddr + *last_size != paddr)) {
+                    print_pte(mon, va_bits, *vbase, *pbase,
+                              *last_paddr + *last_size - *pbase, *last_attr);
+
+                    *vbase = start;
+                    *pbase = paddr;
+                    *last_attr = attr;
+                }
+
+                *last_paddr = paddr;
+                *last_size = pgsize;
+            } else {
+                /* pointer to the next level of the page table */
+                walk_pte(mon, paddr, start, level - 1, ptidxbits, ptesize,
+                         va_bits, vbase, pbase, last_paddr,
+                         last_size, last_attr);
+            }
+        }
+
+        start += pgsize;
+    }
+
+}
+
+static void mem_info_svxx(Monitor *mon, CPUArchState *env)
+{
+    int levels, ptidxbits, ptesize, vm, va_bits;
+    hwaddr base;
+    target_ulong vbase;
+    hwaddr pbase;
+    hwaddr last_paddr;
+    target_ulong last_size;
+    int last_attr;
+
+    base = (hwaddr)get_field(env->satp, SATP_PPN) << PGSHIFT;
+
+    vm = get_field(env->satp, SATP_MODE);
+    switch (vm) {
+    case VM_1_10_SV32:
+        levels = 2;
+        ptidxbits = 10;
+        ptesize = 4;
+        break;
+    case VM_1_10_SV39:
+        levels = 3;
+        ptidxbits = 9;
+        ptesize = 8;
+        break;
+    case VM_1_10_SV48:
+        levels = 4;
+        ptidxbits = 9;
+        ptesize = 8;
+        break;
+    case VM_1_10_SV57:
+        levels = 5;
+        ptidxbits = 9;
+        ptesize = 8;
+        break;
+    default:
+        g_assert_not_reached();
+        break;
+    }
+
+    /* calculate virtual address bits */
+    va_bits = PGSHIFT + levels * ptidxbits;
+
+    /* print header */
+    print_pte_header(mon);
+
+    vbase = -1;
+    pbase = -1;
+    last_paddr = -1;
+    last_size = 0;
+    last_attr = 0;
+
+    /* walk page tables, starting from address 0 */
+    walk_pte(mon, base, 0, levels - 1, ptidxbits, ptesize, va_bits,
+             &vbase, &pbase, &last_paddr, &last_size, &last_attr);
+
+    /* don't forget the last one */
+    print_pte(mon, va_bits, vbase, pbase,
+              last_paddr + last_size - pbase, last_attr);
+}
+
+void hmp_info_mem(Monitor *mon, const QDict *qdict)
+{
+    CPUArchState *env;
+
+    env = mon_get_cpu_env();
+    if (!env) {
+        monitor_printf(mon, "No CPU available\n");
+        return;
+    }
+
+    if (!riscv_feature(env, RISCV_FEATURE_MMU)) {
+        monitor_printf(mon, "S-mode MMU unavailable\n");
+        return;
+    }
+
+    if (env->priv_ver < PRIV_VERSION_1_10_0) {
+        monitor_printf(mon, "Privileged mode < 1.10 unsupported\n");
+        return;
+    }
+
+    if (!(env->satp & SATP_MODE)) {
+        monitor_printf(mon, "No translation or protection\n");
+        return;
+    }
+
+    mem_info_svxx(mon, env);
+}