On Fri, May 31, 2024 at 5:33 PM <devel-request@lists.crash-utility.osci.io> wrote:
Date: Fri, 31 May 2024 17:19:34 +0800
From: Tao Liu <ltao@redhat.com>
Subject: [Crash-utility] [PATCH v4 11/16] Fix cpumask_t recursive
        dependence issue
To: devel@lists.crash-utility.osci.io
Cc: Mahesh J Salgaonkar <mahesh@linux.ibm.com>, "Naveen N . Rao"
        <naveen.n.rao@linux.vnet.ibm.com>, Lianbo Jiang <lijiang@redhat.com>,
        Alexey Makhalov <alexey.makhalov@broadcom.com>
Message-ID: <20240531091939.97828-12-ltao@redhat.com>
Content-Type: text/plain; charset=UTF-8

There is recursive dependence for cpumask_t and will exhause the stack,
see the following stack trace:

(gdb) bt
 ...snip...
 #61965 0x00000000005de98c in datatype_info (name=name@entry=0xa5b1fd "cpumask_t", member=member@entry=0x0, dm=dm@entry=0xfffffffffffffffc) at symbols.c:6694
 #61966 0x000000000057e4ea in cpu_map_size ...
 #61967 0x000000000058e7bd in get_cpus_online ...
 #61968 0x000000000061fa4b in diskdump_get_prstatus_percpu ...
 #61969 0x0000000000616d74 in get_netdump_regs_x86_64 ...
 #61970 0x0000000000585290 in get_dumpfile_regs ...
 #61971 0x00000000005b7a3c in x86_64_get_current_task_reg ...
 #61972 0x0000000000650389 in crash_target::fetch_registers ...
 #61973 0x00000000008f385a in target_fetch_registers ...
 #61974 0x000000000086ecda in regcache::raw_update ...
 #61975 regcache::raw_update ...
 #61976 0x000000000086ed7a in readable_regcache::raw_read ...
 #61977 0x000000000086f063 in readable_regcache::cooked_read_value ...
 #61978 0x000000000089c4ee in sentinel_frame_prev_register ...
 #61979 0x0000000000786c76 in frame_unwind_register_value ...
 #61980 0x0000000000786f18 in frame_register_unwind ...
 #61981 0x0000000000787267 in frame_unwind_register ...
 #61982 0x00000000007ad9b0 in i386_unwind_pc ...
 #61983 0x00000000007866c0 in frame_unwind_pc ...
 #61984 0x000000000078679c in get_frame_pc ...
 #61985 get_frame_address_in_block ...
 #61986 0x0000000000786849 in get_frame_address_in_block_if_available ...
 #61987 0x0000000000691466 in get_frame_block ...
 #61988 0x00000000008b9430 in get_selected_block ...
 #61989 0x000000000084f8f2 in parse_exp_in_context ...
 #61990 0x000000000084f9e5 in parse_exp_1 ...
 #61991 parse_expression ...
 #61992 0x00000000008d44da in gdb_get_datatype ...
 #61993 gdb_command_funnel_1 ...
 #61994 0x00000000008d48ae in gdb_command_funnel ...
 #61995 0x000000000059cc42 in gdb_interface ...
 #61996 0x00000000005de98c in datatype_info (name=name@entry=0xa5b1fd "cpumask_t", member=member@entry=0x0, dm=dm@entry=0xfffffffffffffffc) at symbols.c:6694
 #61997 0x000000000057e4ea in cpu_map_size ...
 #61998 0x000000000058e7bd in get_cpus_online () ...
 #61999 0x000000000061fa4b in diskdump_get_prstatus_percpu ...
 #62000 0x0000000000616d74 in get_netdump_regs_x86_64 ...
 #62001 0x0000000000585290 in get_dumpfile_regs ...
 #62002 0x00000000005b7a3c in x86_64_get_current_task_reg ...
 #62003 0x0000000000650389 in crash_target::fetch_registers ...

The cpumask_t will be recursively evaluated. This patch will
fix the bug.

Cc: Sourabh Jain <sourabhjain@linux.ibm.com>
Cc: Hari Bathini <hbathini@linux.ibm.com>
Cc: Mahesh J Salgaonkar <mahesh@linux.ibm.com>
Cc: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Cc: Lianbo Jiang <lijiang@redhat.com>
Cc: HAGIO KAZUHITO(萩尾 一仁) <k-hagio-ab@nec.com>
Cc: Tao Liu <ltao@redhat.com>
Cc: Alexey Makhalov <alexey.makhalov@broadcom.com>
Signed-off-by: Tao Liu <ltao@redhat.com>
---
 defs.h   |  1 +
 kernel.c | 17 ++++++++++-------
 2 files changed, 11 insertions(+), 7 deletions(-)

diff --git a/defs.h b/defs.h
index ed52cc3..fd00462 100644
--- a/defs.h
+++ b/defs.h
@@ -2429,6 +2429,7 @@ struct size_table {         /* stash of commonly-used sizes */
        long maple_tree;
        long maple_node;
        long module_memory;
+       long cpumask_t;
 };


Can you add the cpumask_t into the dump_offset_table()?

 struct array_table {
diff --git a/kernel.c b/kernel.c
index 3730c55..2cae305 100644
--- a/kernel.c
+++ b/kernel.c
@@ -157,6 +157,7 @@ kernel_init()
         if (!(kt->cpu_flags = (ulong *)calloc(NR_CPUS, sizeof(ulong))))
                 error(FATAL, "cannot malloc cpu_flags array");

+       STRUCT_SIZE_INIT(cpumask_t, "cpumask_t");
        cpu_maps_init();

        kt->stext = symbol_value("_stext");
@@ -913,9 +914,10 @@ cpu_map_size(const char *type)
        struct gnu_request req;

         if (LKCD_KERNTYPES()) {
-                if ((len = STRUCT_SIZE("cpumask_t")) < 0)
-                        error(FATAL, "cannot determine type cpumask_t\n");
-               return len;
+               if (INVALID_SIZE(cpumask_t))
+                       error(FATAL, "cannot determine type cpumask_t\n");
+               else
+                       return SIZE(cpumask_t);

Let's remove the 'else', keep the original style:

+               if (INVALID_SIZE(cpumask_t))
+                       error(FATAL, "Invalid type cpumask_t\n");
+
+               return SIZE(cpumask_t);

        }

        sprintf(map_symbol, "cpu_%s_map", type);
@@ -925,11 +927,10 @@ cpu_map_size(const char *type)
                return len;
        }

-       len = STRUCT_SIZE("cpumask_t");
-       if (len < 0)
+       if (INVALID_SIZE(cpumask_t))
                return sizeof(ulong);
        else
-               return len;
+               return SIZE(cpumask_t);

Ditto.
+       if (INVALID_SIZE(cpumask_t))
             return sizeof(ulong);
-        else
-            return len;
+
+        return SIZE(cpumask_t);
 
 }

 /*
@@ -952,8 +953,10 @@ cpu_maps_init(void)
                { ACTIVE_MAP, "active" },
        };

-       if ((len = STRUCT_SIZE("cpumask_t")) < 0)
+       if (INVALID_SIZE(cpumask_t))
                len = sizeof(ulong);
+       else
+               len = SIZE(cpumask_t);

        buf = GETBUF(len);

In addition, I still found two similar calls as below. Can you try to make similar changes in those two functions?
 
1.  kernel.c:
void
generic_get_irq_affinity(int irq)
{
...
        if ((len = STRUCT_SIZE("cpumask_t")) < 0)
                len = DIV_ROUND_UP(kt->cpus, BITS_PER_LONG) * sizeof(ulong);
...
}

2. tools.
ulong *
get_cpumask_buf(void)
{
        int cpulen;
        if ((cpulen = STRUCT_SIZE("cpumask_t")) < 0)
                cpulen = DIV_ROUND_UP(kt->cpus, BITS_PER_LONG) * sizeof(ulong);
        return (ulong *)GETBUF(cpulen);
}
 

Thanks
Lianbo


--
2.40.1