Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1313132 > unrolled thread

[PATCH 4/5] scripts/gdb: Add mount point list command

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  [PATCH 4/5] scripts/gdb: Add mount point list command Kieran Bingham <kieran.bingham@linaro.org> - 2016-01-20 12:20 +0100
    Re: [PATCH 4/5] scripts/gdb: Add mount point list command Jan Kiszka <jan.kiszka@siemens.com> - 2016-01-20 12:50 +0100
      Re: [PATCH 4/5] scripts/gdb: Add mount point list command Kieran Bingham <kieran.bingham@linaro.org> - 2016-01-20 13:00 +0100
        Re: [PATCH 4/5] scripts/gdb: Add mount point list command Jan Kiszka <jan.kiszka@siemens.com> - 2016-01-20 13:10 +0100
    Re: [PATCH 4/5] scripts/gdb: Add mount point list command Jan Kiszka <jan.kiszka@siemens.com> - 2016-01-23 13:40 +0100
    Re: [PATCH 4/5] scripts/gdb: Add mount point list command Jan Kiszka <jan.kiszka@siemens.com> - 2016-01-23 16:30 +0100
      Re: [PATCH 4/5] scripts/gdb: Add mount point list command Kieran Bingham <kieran.bingham@linaro.org> - 2016-01-24 01:30 +0100

#1313132 — [PATCH 4/5] scripts/gdb: Add mount point list command

lx-mounts will identify current mount points based on the 'init_task'
namespace by default, as we do not yet have a kernel thread list
implementation to select the current running thread.

Optionally, a user can specify a PID to list from that process'
namespace

This is somewhat limited vs the /proc/mounts file, as that calls into
vfs hooks through the s_op functions to obtain extra information.

Signed-off-by: Kieran Bingham <kieran.bingham@linaro.org>
---


In this patch, I'm interested in your opinions on coding styles.
Would you prefer to see the function helpers, (dentry_name, info_opts) where
they are, or inside the command as class members? Or perhaps defined in utils?

This also shows where I need to take constant information from the kernel.
In this case, they are simple numerical bitflags, and unlikely to change but
I didn't want to duplicate their values.


 scripts/gdb/linux/constants.py.in |  21 ++++++++
 scripts/gdb/linux/proc.py         | 110 ++++++++++++++++++++++++++++++++++++++
 2 files changed, 131 insertions(+)

diff --git a/scripts/gdb/linux/constants.py.in b/scripts/gdb/linux/constants.py.in
index d84084ac945b..739a15d2e984 100644
--- a/scripts/gdb/linux/constants.py.in
+++ b/scripts/gdb/linux/constants.py.in
@@ -12,7 +12,11 @@
  *
  */
 
+#include <linux/fs.h>
+#include <linux/mount.h>
+
 /* We need to stringify expanded macros so that they can be parsed */
 #define STRING(x) #x
 #define XSTRING(x) STRING(x)
 
@@ -20,3 +24,20 @@
 <!-- end-c-headers -->
 
 import gdb
+
+/* linux/fs.h */
+LX_MS_RDONLY = MS_RDONLY
+LX_MS_SYNCHRONOUS = MS_SYNCHRONOUS
+LX_MS_MANDLOCK = MS_MANDLOCK
+LX_MS_DIRSYNC = MS_DIRSYNC
+LX_MS_NOATIME = MS_NOATIME
+LX_MS_NODIRATIME = MS_NODIRATIME
+
+/* linux/mount.h */
+LX_MNT_NOSUID = MNT_NOSUID
+LX_MNT_NODEV = MNT_NODEV
+LX_MNT_NOEXEC = MNT_NOEXEC
+LX_MNT_NOATIME = MNT_NOATIME
+LX_MNT_NODIRATIME = MNT_NODIRATIME
+LX_MNT_RELATIME = MNT_RELATIME
+
diff --git a/scripts/gdb/linux/proc.py b/scripts/gdb/linux/proc.py
index d855b2fd9a06..b79ce2a33a3d 100644
--- a/scripts/gdb/linux/proc.py
+++ b/scripts/gdb/linux/proc.py
@@ -12,6 +12,10 @@
 #
 
 import gdb
+from linux import constants
+from linux import utils
+from linux import tasks
+from linux import lists
 
 
 class LxCmdLine(gdb.Command):
@@ -96,3 +100,109 @@ Equivalent to cat /proc/ioports on a running target"""
         return show_lx_resources("ioport_resource")
 
 LxIOPorts()
+
+
+# Mount namespace viewer
+#  /proc/mounts
+
+
+def dentry_name(d):
+    if d['d_parent'] == d:
+        return ""
+    p = dentry_name(d['d_parent']) + "/"
+    return p + d['d_iname'].string()
+
+
+def info_opts(lst, opt):
+    opts = ""
+    for key, string in lst.items():
+        if opt & key:
+            opts += string
+    return opts
+
+
+FS_INFO = {constants.LX_MS_SYNCHRONOUS: ",sync",
+           constants.LX_MS_MANDLOCK: ",mand",
+           constants.LX_MS_DIRSYNC: ",dirsync",
+           constants.LX_MS_NOATIME: ",noatime",
+           constants.LX_MS_NODIRATIME: ",nodiratime"}
+
+MNT_INFO = {constants.LX_MNT_NOSUID: ",nosuid",
+            constants.LX_MNT_NODEV: ",nodev",
+            constants.LX_MNT_NOEXEC: ",noexec",
+            constants.LX_MNT_NOATIME: ",noatime",
+            constants.LX_MNT_NODIRATIME: ",nodiratime",
+            constants.LX_MNT_RELATIME: ",relatime"}
+
+mount_type = utils.CachedType("struct mount")
+mount_ptr_type = mount_type.get_type().pointer()
+
+
+class LxMounts(gdb.Command):
+    """Report the VFS mounts of the current process namespace.
+
+Equivalent to cat /proc/mounts on a running target
+An integer value can be supplied to display the mount
+values of that process namespace"""
+
+    def __init__(self):
+        super(LxMounts, self).__init__("lx-mounts", gdb.COMMAND_DATA)
+
+    # Equivalent to proc_namespace.c:show_vfsmnt
+    # However, that has the ability to call into s_op functions
+    # whereas we cannot and must make do with the information we can obtain.
+    def invoke(self, arg, from_tty):
+        argv = gdb.string_to_argv(arg)
+        if len(argv) >= 1:
+            try:
+                pid = int(argv[0])
+            except:
+                raise gdb.GdbError("Provide a PID as integer value")
+        else:
+            pid = 1
+
+        task = tasks.get_task_by_pid(pid)
+        if not task:
+            raise gdb.GdbError("Couldn't find a process with PID {}"
+                               .format(pid))
+
+        namespace = task['nsproxy']['mnt_ns']
+        if not namespace:
+            raise gdb.GdbError("No namespace for current process")
+
+        for vfs in lists.items(mount_ptr_type, "mnt_list", namespace['list']):
+            # There appears to be a null entry at the end of the list...
+            if not vfs['mnt_parent']:
+                break
+
+            devname = vfs['mnt_devname'].string()
+            devname = devname if devname else "none"
+
+            pathname = ""
+            parent = vfs
+            while True:
+                mntpoint = parent['mnt_mountpoint']
+                pathname = dentry_name(mntpoint) + pathname
+                if (parent == parent['mnt_parent']):
+                    break
+                parent = parent['mnt_parent']
+
+            if (pathname == ""):
+                pathname = "/"
+
+            superblock = vfs['mnt']['mnt_sb']
+            fstype = superblock['s_type']['name'].string()
+            s_flags = int(superblock['s_flags'])
+            m_flags = int(vfs['mnt']['mnt_flags'])
+            rd = "ro" if (s_flags & constants.LX_MS_RDONLY) else "rw"
+
+            gdb.write(
+                "{} {} {} {}{}{} 0 0\n"
+                .format(devname,
+                        pathname,
+                        fstype,
+                        rd,
+                        info_opts(FS_INFO, s_flags),
+                        info_opts(MNT_INFO, m_flags)))
+
+LxMounts()
-- 
2.5.0

[toc] | [next] | [standalone]

#1313148

Hi Kieran,

just a couple of quick comments:

On 2016-01-20 12:15, Kieran Bingham wrote:
> lx-mounts will identify current mount points based on the 'init_task'
> namespace by default, as we do not yet have a kernel thread list
> implementation to select the current running thread.

current_task? See LxCurrentFunc, could be factored out if usable. Or
what are you looking for?

> 
> Optionally, a user can specify a PID to list from that process'
> namespace
> 
> This is somewhat limited vs the /proc/mounts file, as that calls into
> vfs hooks through the s_op functions to obtain extra information.
> 
> Signed-off-by: Kieran Bingham <kieran.bingham@linaro.org>
> ---
> 
> 
> In this patch, I'm interested in your opinions on coding styles.
> Would you prefer to see the function helpers, (dentry_name, info_opts) where
> they are, or inside the command as class members? Or perhaps defined in utils?

Need to look into this.

> 
> This also shows where I need to take constant information from the kernel.
> In this case, they are simple numerical bitflags, and unlikely to change but
> I didn't want to duplicate their values.

Maybe we can generate python files with the required constants from the
C headers during build? Similar to asm-offsets.c stuff.

> 
> 
>  scripts/gdb/linux/constants.py.in |  21 ++++++++
>  scripts/gdb/linux/proc.py         | 110 ++++++++++++++++++++++++++++++++++++++
>  2 files changed, 131 insertions(+)
> 
> diff --git a/scripts/gdb/linux/constants.py.in b/scripts/gdb/linux/constants.py.in
> index d84084ac945b..739a15d2e984 100644
> --- a/scripts/gdb/linux/constants.py.in
> +++ b/scripts/gdb/linux/constants.py.in
> @@ -12,7 +12,11 @@
>   *
>   */
>  
> +#include <linux/fs.h>
> +#include <linux/mount.h>
> +
>  /* We need to stringify expanded macros so that they can be parsed */
>  #define STRING(x) #x
>  #define XSTRING(x) STRING(x)
>  
> @@ -20,3 +24,20 @@
>  <!-- end-c-headers -->
>  
>  import gdb
> +
> +/* linux/fs.h */
> +LX_MS_RDONLY = MS_RDONLY
> +LX_MS_SYNCHRONOUS = MS_SYNCHRONOUS
> +LX_MS_MANDLOCK = MS_MANDLOCK
> +LX_MS_DIRSYNC = MS_DIRSYNC
> +LX_MS_NOATIME = MS_NOATIME
> +LX_MS_NODIRATIME = MS_NODIRATIME
> +
> +/* linux/mount.h */
> +LX_MNT_NOSUID = MNT_NOSUID
> +LX_MNT_NODEV = MNT_NODEV
> +LX_MNT_NOEXEC = MNT_NOEXEC
> +LX_MNT_NOATIME = MNT_NOATIME
> +LX_MNT_NODIRATIME = MNT_NODIRATIME
> +LX_MNT_RELATIME = MNT_RELATIME
> +
> diff --git a/scripts/gdb/linux/proc.py b/scripts/gdb/linux/proc.py
> index d855b2fd9a06..b79ce2a33a3d 100644
> --- a/scripts/gdb/linux/proc.py
> +++ b/scripts/gdb/linux/proc.py
> @@ -12,6 +12,10 @@
>  #
>  
>  import gdb
> +from linux import constants
> +from linux import utils
> +from linux import tasks
> +from linux import lists
>  
>  
>  class LxCmdLine(gdb.Command):
> @@ -96,3 +100,109 @@ Equivalent to cat /proc/ioports on a running target"""
>          return show_lx_resources("ioport_resource")
>  
>  LxIOPorts()
> +
> +
> +# Mount namespace viewer
> +#  /proc/mounts
> +
> +
> +def dentry_name(d):
> +    if d['d_parent'] == d:
> +        return ""
> +    p = dentry_name(d['d_parent']) + "/"
> +    return p + d['d_iname'].string()
> +
> +
> +def info_opts(lst, opt):
> +    opts = ""
> +    for key, string in lst.items():
> +        if opt & key:
> +            opts += string
> +    return opts
> +
> +
> +FS_INFO = {constants.LX_MS_SYNCHRONOUS: ",sync",
> +           constants.LX_MS_MANDLOCK: ",mand",
> +           constants.LX_MS_DIRSYNC: ",dirsync",
> +           constants.LX_MS_NOATIME: ",noatime",
> +           constants.LX_MS_NODIRATIME: ",nodiratime"}
> +
> +MNT_INFO = {constants.LX_MNT_NOSUID: ",nosuid",
> +            constants.LX_MNT_NODEV: ",nodev",
> +            constants.LX_MNT_NOEXEC: ",noexec",
> +            constants.LX_MNT_NOATIME: ",noatime",
> +            constants.LX_MNT_NODIRATIME: ",nodiratime",
> +            constants.LX_MNT_RELATIME: ",relatime"}
> +
> +mount_type = utils.CachedType("struct mount")
> +mount_ptr_type = mount_type.get_type().pointer()
> +
> +
> +class LxMounts(gdb.Command):
> +    """Report the VFS mounts of the current process namespace.
> +
> +Equivalent to cat /proc/mounts on a running target
> +An integer value can be supplied to display the mount
> +values of that process namespace"""
> +
> +    def __init__(self):
> +        super(LxMounts, self).__init__("lx-mounts", gdb.COMMAND_DATA)
> +
> +    # Equivalent to proc_namespace.c:show_vfsmnt
> +    # However, that has the ability to call into s_op functions
> +    # whereas we cannot and must make do with the information we can obtain.
> +    def invoke(self, arg, from_tty):
> +        argv = gdb.string_to_argv(arg)
> +        if len(argv) >= 1:
> +            try:
> +                pid = int(argv[0])
> +            except:
> +                raise gdb.GdbError("Provide a PID as integer value")
> +        else:
> +            pid = 1
> +
> +        task = tasks.get_task_by_pid(pid)
> +        if not task:
> +            raise gdb.GdbError("Couldn't find a process with PID {}"
> +                               .format(pid))
> +
> +        namespace = task['nsproxy']['mnt_ns']
> +        if not namespace:
> +            raise gdb.GdbError("No namespace for current process")
> +
> +        for vfs in lists.items(mount_ptr_type, "mnt_list", namespace['list']):
> +            # There appears to be a null entry at the end of the list...
> +            if not vfs['mnt_parent']:
> +                break
> +
> +            devname = vfs['mnt_devname'].string()
> +            devname = devname if devname else "none"
> +
> +            pathname = ""
> +            parent = vfs
> +            while True:
> +                mntpoint = parent['mnt_mountpoint']
> +                pathname = dentry_name(mntpoint) + pathname
> +                if (parent == parent['mnt_parent']):
> +                    break
> +                parent = parent['mnt_parent']
> +
> +            if (pathname == ""):
> +                pathname = "/"
> +
> +            superblock = vfs['mnt']['mnt_sb']
> +            fstype = superblock['s_type']['name'].string()
> +            s_flags = int(superblock['s_flags'])
> +            m_flags = int(vfs['mnt']['mnt_flags'])
> +            rd = "ro" if (s_flags & constants.LX_MS_RDONLY) else "rw"
> +
> +            gdb.write(
> +                "{} {} {} {}{}{} 0 0\n"
> +                .format(devname,
> +                        pathname,
> +                        fstype,
> +                        rd,
> +                        info_opts(FS_INFO, s_flags),
> +                        info_opts(MNT_INFO, m_flags)))
> +
> +LxMounts()
> 

-- 
Siemens AG, Corporate Technology, CT RDA ITP SES-DE
Corporate Competence Center Embedded Linux

[toc] | [prev] | [next] | [standalone]

#1313158

Hi Jan,

On 20/01/16 11:42, Jan Kiszka wrote:
> Hi Kieran,
> 
> just a couple of quick comments:
> 
> On 2016-01-20 12:15, Kieran Bingham wrote:
>> lx-mounts will identify current mount points based on the 'init_task'
>> namespace by default, as we do not yet have a kernel thread list
>> implementation to select the current running thread.
> 
> current_task? See LxCurrentFunc, could be factored out if usable. Or
> what are you looking for?

LxCurrentFunc relies on gdb.parse_and_eval("&current_task") which is not
available on ARM.

Although that is not what I was referring to in the comment. My meaning
was that once we have gdb-thread objects created, (my next phase of
work) then this command would operate on the thread currently selected
in gdb, based on the inferior_ptid.


>>
>> Optionally, a user can specify a PID to list from that process'
>> namespace
>>
>> This is somewhat limited vs the /proc/mounts file, as that calls into
>> vfs hooks through the s_op functions to obtain extra information.
>>
>> Signed-off-by: Kieran Bingham <kieran.bingham@linaro.org>
>> ---
>>
>>
>> In this patch, I'm interested in your opinions on coding styles.
>> Would you prefer to see the function helpers, (dentry_name, info_opts) where
>> they are, or inside the command as class members? Or perhaps defined in utils?
> 
> Need to look into this.
> 
>>
>> This also shows where I need to take constant information from the kernel.
>> In this case, they are simple numerical bitflags, and unlikely to change but
>> I didn't want to duplicate their values.
> 
> Maybe we can generate python files with the required constants from the
> C headers during build? Similar to asm-offsets.c stuff.
> 


Yes, this is what I've implemented in [PATCH 1/5] ? Perhaps the mails
have reached you out-of-order.



>>
>>
>>  scripts/gdb/linux/constants.py.in |  21 ++++++++
>>  scripts/gdb/linux/proc.py         | 110 ++++++++++++++++++++++++++++++++++++++
>>  2 files changed, 131 insertions(+)
>>
>> diff --git a/scripts/gdb/linux/constants.py.in b/scripts/gdb/linux/constants.py.in
>> index d84084ac945b..739a15d2e984 100644
>> --- a/scripts/gdb/linux/constants.py.in
>> +++ b/scripts/gdb/linux/constants.py.in
>> @@ -12,7 +12,11 @@
>>   *
>>   */
>>  
>> +#include <linux/fs.h>
>> +#include <linux/mount.h>
>> +
>>  /* We need to stringify expanded macros so that they can be parsed */
>>  #define STRING(x) #x
>>  #define XSTRING(x) STRING(x)
>>  
>> @@ -20,3 +24,20 @@
>>  <!-- end-c-headers -->
>>  
>>  import gdb
>> +
>> +/* linux/fs.h */
>> +LX_MS_RDONLY = MS_RDONLY
>> +LX_MS_SYNCHRONOUS = MS_SYNCHRONOUS
>> +LX_MS_MANDLOCK = MS_MANDLOCK
>> +LX_MS_DIRSYNC = MS_DIRSYNC
>> +LX_MS_NOATIME = MS_NOATIME
>> +LX_MS_NODIRATIME = MS_NODIRATIME
>> +
>> +/* linux/mount.h */
>> +LX_MNT_NOSUID = MNT_NOSUID
>> +LX_MNT_NODEV = MNT_NODEV
>> +LX_MNT_NOEXEC = MNT_NOEXEC
>> +LX_MNT_NOATIME = MNT_NOATIME
>> +LX_MNT_NODIRATIME = MNT_NODIRATIME
>> +LX_MNT_RELATIME = MNT_RELATIME
>> +
>> diff --git a/scripts/gdb/linux/proc.py b/scripts/gdb/linux/proc.py
>> index d855b2fd9a06..b79ce2a33a3d 100644
>> --- a/scripts/gdb/linux/proc.py
>> +++ b/scripts/gdb/linux/proc.py
>> @@ -12,6 +12,10 @@
>>  #
>>  
>>  import gdb
>> +from linux import constants
>> +from linux import utils
>> +from linux import tasks
>> +from linux import lists
>>  
>>  
>>  class LxCmdLine(gdb.Command):
>> @@ -96,3 +100,109 @@ Equivalent to cat /proc/ioports on a running target"""
>>          return show_lx_resources("ioport_resource")
>>  
>>  LxIOPorts()
>> +
>> +
>> +# Mount namespace viewer
>> +#  /proc/mounts
>> +
>> +
>> +def dentry_name(d):
>> +    if d['d_parent'] == d:
>> +        return ""
>> +    p = dentry_name(d['d_parent']) + "/"
>> +    return p + d['d_iname'].string()
>> +
>> +
>> +def info_opts(lst, opt):
>> +    opts = ""
>> +    for key, string in lst.items():
>> +        if opt & key:
>> +            opts += string
>> +    return opts
>> +
>> +
>> +FS_INFO = {constants.LX_MS_SYNCHRONOUS: ",sync",
>> +           constants.LX_MS_MANDLOCK: ",mand",
>> +           constants.LX_MS_DIRSYNC: ",dirsync",
>> +           constants.LX_MS_NOATIME: ",noatime",
>> +           constants.LX_MS_NODIRATIME: ",nodiratime"}
>> +
>> +MNT_INFO = {constants.LX_MNT_NOSUID: ",nosuid",
>> +            constants.LX_MNT_NODEV: ",nodev",
>> +            constants.LX_MNT_NOEXEC: ",noexec",
>> +            constants.LX_MNT_NOATIME: ",noatime",
>> +            constants.LX_MNT_NODIRATIME: ",nodiratime",
>> +            constants.LX_MNT_RELATIME: ",relatime"}
>> +
>> +mount_type = utils.CachedType("struct mount")
>> +mount_ptr_type = mount_type.get_type().pointer()
>> +
>> +
>> +class LxMounts(gdb.Command):
>> +    """Report the VFS mounts of the current process namespace.
>> +
>> +Equivalent to cat /proc/mounts on a running target
>> +An integer value can be supplied to display the mount
>> +values of that process namespace"""
>> +
>> +    def __init__(self):
>> +        super(LxMounts, self).__init__("lx-mounts", gdb.COMMAND_DATA)
>> +
>> +    # Equivalent to proc_namespace.c:show_vfsmnt
>> +    # However, that has the ability to call into s_op functions
>> +    # whereas we cannot and must make do with the information we can obtain.
>> +    def invoke(self, arg, from_tty):
>> +        argv = gdb.string_to_argv(arg)
>> +        if len(argv) >= 1:
>> +            try:
>> +                pid = int(argv[0])
>> +            except:
>> +                raise gdb.GdbError("Provide a PID as integer value")
>> +        else:
>> +            pid = 1
>> +
>> +        task = tasks.get_task_by_pid(pid)
>> +        if not task:
>> +            raise gdb.GdbError("Couldn't find a process with PID {}"
>> +                               .format(pid))
>> +
>> +        namespace = task['nsproxy']['mnt_ns']
>> +        if not namespace:
>> +            raise gdb.GdbError("No namespace for current process")
>> +
>> +        for vfs in lists.items(mount_ptr_type, "mnt_list", namespace['list']):
>> +            # There appears to be a null entry at the end of the list...
>> +            if not vfs['mnt_parent']:
>> +                break
>> +
>> +            devname = vfs['mnt_devname'].string()
>> +            devname = devname if devname else "none"
>> +
>> +            pathname = ""
>> +            parent = vfs
>> +            while True:
>> +                mntpoint = parent['mnt_mountpoint']
>> +                pathname = dentry_name(mntpoint) + pathname
>> +                if (parent == parent['mnt_parent']):
>> +                    break
>> +                parent = parent['mnt_parent']
>> +
>> +            if (pathname == ""):
>> +                pathname = "/"
>> +
>> +            superblock = vfs['mnt']['mnt_sb']
>> +            fstype = superblock['s_type']['name'].string()
>> +            s_flags = int(superblock['s_flags'])
>> +            m_flags = int(vfs['mnt']['mnt_flags'])
>> +            rd = "ro" if (s_flags & constants.LX_MS_RDONLY) else "rw"
>> +
>> +            gdb.write(
>> +                "{} {} {} {}{}{} 0 0\n"
>> +                .format(devname,
>> +                        pathname,
>> +                        fstype,
>> +                        rd,
>> +                        info_opts(FS_INFO, s_flags),
>> +                        info_opts(MNT_INFO, m_flags)))
>> +
>> +LxMounts()
>>
> 

[toc] | [prev] | [next] | [standalone]

#1313165

On 2016-01-20 12:51, Kieran Bingham wrote:
> Hi Jan,
> 
> On 20/01/16 11:42, Jan Kiszka wrote:
>> Hi Kieran,
>>
>> just a couple of quick comments:
>>
>> On 2016-01-20 12:15, Kieran Bingham wrote:
>>> lx-mounts will identify current mount points based on the 'init_task'
>>> namespace by default, as we do not yet have a kernel thread list
>>> implementation to select the current running thread.
>>
>> current_task? See LxCurrentFunc, could be factored out if usable. Or
>> what are you looking for?
> 
> LxCurrentFunc relies on gdb.parse_and_eval("&current_task") which is not
> available on ARM.

Good point, not only affecting ARM. Needs to be fixed - lx-current() is
an important service.

> 
> Although that is not what I was referring to in the comment. My meaning
> was that once we have gdb-thread objects created, (my next phase of
> work) then this command would operate on the thread currently selected
> in gdb, based on the inferior_ptid.

I see. However, to have thread selection run with reasonable defaults,
we will need current() support as well.

> 
> 
>>>
>>> Optionally, a user can specify a PID to list from that process'
>>> namespace
>>>
>>> This is somewhat limited vs the /proc/mounts file, as that calls into
>>> vfs hooks through the s_op functions to obtain extra information.
>>>
>>> Signed-off-by: Kieran Bingham <kieran.bingham@linaro.org>
>>> ---
>>>
>>>
>>> In this patch, I'm interested in your opinions on coding styles.
>>> Would you prefer to see the function helpers, (dentry_name, info_opts) where
>>> they are, or inside the command as class members? Or perhaps defined in utils?
>>
>> Need to look into this.
>>
>>>
>>> This also shows where I need to take constant information from the kernel.
>>> In this case, they are simple numerical bitflags, and unlikely to change but
>>> I didn't want to duplicate their values.
>>
>> Maybe we can generate python files with the required constants from the
>> C headers during build? Similar to asm-offsets.c stuff.
>>
> 
> 
> Yes, this is what I've implemented in [PATCH 1/5] ? Perhaps the mails
> have reached you out-of-order.
> 

No, they were only processed in random order. Perfect!

Jan

-- 
Siemens AG, Corporate Technology, CT RDA ITP SES-DE
Corporate Competence Center Embedded Linux

[toc] | [prev] | [next] | [standalone]

#1315583

On 2016-01-20 12:15, Kieran Bingham wrote:
> lx-mounts will identify current mount points based on the 'init_task'
> namespace by default, as we do not yet have a kernel thread list
> implementation to select the current running thread.
> 
> Optionally, a user can specify a PID to list from that process'
> namespace
> 
> This is somewhat limited vs the /proc/mounts file, as that calls into
> vfs hooks through the s_op functions to obtain extra information.
> 
> Signed-off-by: Kieran Bingham <kieran.bingham@linaro.org>
> ---
> 
> 
> In this patch, I'm interested in your opinions on coding styles.
> Would you prefer to see the function helpers, (dentry_name, info_opts) where
> they are, or inside the command as class members? Or perhaps defined in utils?
> 
> This also shows where I need to take constant information from the kernel.
> In this case, they are simple numerical bitflags, and unlikely to change but
> I didn't want to duplicate their values.
> 
> 
>  scripts/gdb/linux/constants.py.in |  21 ++++++++
>  scripts/gdb/linux/proc.py         | 110 ++++++++++++++++++++++++++++++++++++++
>  2 files changed, 131 insertions(+)
> 
> diff --git a/scripts/gdb/linux/constants.py.in b/scripts/gdb/linux/constants.py.in
> index d84084ac945b..739a15d2e984 100644
> --- a/scripts/gdb/linux/constants.py.in
> +++ b/scripts/gdb/linux/constants.py.in
> @@ -12,7 +12,11 @@
>   *
>   */
>  
> +#include <linux/fs.h>
> +#include <linux/mount.h>
> +
>  /* We need to stringify expanded macros so that they can be parsed */
>  #define STRING(x) #x
>  #define XSTRING(x) STRING(x)
>  

This hunk is malformed, indicating that 4 lines are added while there
are actually only 3 new ones. Could you check what went wrong? I'm
applying it manually for now.

Jan

-- 
Siemens AG, Corporate Technology, CT RDA ITP SES-DE
Corporate Competence Center Embedded Linux

[toc] | [prev] | [next] | [standalone]

#1315663

On 2016-01-20 12:15, Kieran Bingham wrote:
> lx-mounts will identify current mount points based on the 'init_task'
> namespace by default, as we do not yet have a kernel thread list
> implementation to select the current running thread.
> 
> Optionally, a user can specify a PID to list from that process'
> namespace
> 
> This is somewhat limited vs the /proc/mounts file, as that calls into
> vfs hooks through the s_op functions to obtain extra information.
> 
> Signed-off-by: Kieran Bingham <kieran.bingham@linaro.org>
> ---
> 
> 
> In this patch, I'm interested in your opinions on coding styles.
> Would you prefer to see the function helpers, (dentry_name, info_opts) where
> they are, or inside the command as class members? Or perhaps defined in utils?

Do you think they could be useful beyond this class? If not, stick them
inside. Refactoring can still be done once they are needed.

> 
> This also shows where I need to take constant information from the kernel.
> In this case, they are simple numerical bitflags, and unlikely to change but
> I didn't want to duplicate their values.
> 
> 
>  scripts/gdb/linux/constants.py.in |  21 ++++++++
>  scripts/gdb/linux/proc.py         | 110 ++++++++++++++++++++++++++++++++++++++
>  2 files changed, 131 insertions(+)
> 
> diff --git a/scripts/gdb/linux/constants.py.in b/scripts/gdb/linux/constants.py.in
> index d84084ac945b..739a15d2e984 100644
> --- a/scripts/gdb/linux/constants.py.in
> +++ b/scripts/gdb/linux/constants.py.in
> @@ -12,7 +12,11 @@
>   *
>   */
>  
> +#include <linux/fs.h>
> +#include <linux/mount.h>
> +
>  /* We need to stringify expanded macros so that they can be parsed */
>  #define STRING(x) #x
>  #define XSTRING(x) STRING(x)
>  
> @@ -20,3 +24,20 @@
>  <!-- end-c-headers -->
>  
>  import gdb
> +
> +/* linux/fs.h */
> +LX_MS_RDONLY = MS_RDONLY
> +LX_MS_SYNCHRONOUS = MS_SYNCHRONOUS
> +LX_MS_MANDLOCK = MS_MANDLOCK
> +LX_MS_DIRSYNC = MS_DIRSYNC
> +LX_MS_NOATIME = MS_NOATIME
> +LX_MS_NODIRATIME = MS_NODIRATIME
> +
> +/* linux/mount.h */
> +LX_MNT_NOSUID = MNT_NOSUID
> +LX_MNT_NODEV = MNT_NODEV
> +LX_MNT_NOEXEC = MNT_NOEXEC
> +LX_MNT_NOATIME = MNT_NOATIME
> +LX_MNT_NODIRATIME = MNT_NODIRATIME
> +LX_MNT_RELATIME = MNT_RELATIME
> +
> diff --git a/scripts/gdb/linux/proc.py b/scripts/gdb/linux/proc.py
> index d855b2fd9a06..b79ce2a33a3d 100644
> --- a/scripts/gdb/linux/proc.py
> +++ b/scripts/gdb/linux/proc.py
> @@ -12,6 +12,10 @@
>  #
>  
>  import gdb
> +from linux import constants
> +from linux import utils
> +from linux import tasks
> +from linux import lists
>  
>  
>  class LxCmdLine(gdb.Command):
> @@ -96,3 +100,109 @@ Equivalent to cat /proc/ioports on a running target"""
>          return show_lx_resources("ioport_resource")
>  
>  LxIOPorts()
> +
> +
> +# Mount namespace viewer
> +#  /proc/mounts
> +
> +
> +def dentry_name(d):
> +    if d['d_parent'] == d:
> +        return ""
> +    p = dentry_name(d['d_parent']) + "/"
> +    return p + d['d_iname'].string()
> +
> +
> +def info_opts(lst, opt):
> +    opts = ""
> +    for key, string in lst.items():
> +        if opt & key:
> +            opts += string
> +    return opts
> +
> +
> +FS_INFO = {constants.LX_MS_SYNCHRONOUS: ",sync",
> +           constants.LX_MS_MANDLOCK: ",mand",
> +           constants.LX_MS_DIRSYNC: ",dirsync",
> +           constants.LX_MS_NOATIME: ",noatime",
> +           constants.LX_MS_NODIRATIME: ",nodiratime"}
> +
> +MNT_INFO = {constants.LX_MNT_NOSUID: ",nosuid",
> +            constants.LX_MNT_NODEV: ",nodev",
> +            constants.LX_MNT_NOEXEC: ",noexec",
> +            constants.LX_MNT_NOATIME: ",noatime",
> +            constants.LX_MNT_NODIRATIME: ",nodiratime",
> +            constants.LX_MNT_RELATIME: ",relatime"}
> +
> +mount_type = utils.CachedType("struct mount")
> +mount_ptr_type = mount_type.get_type().pointer()
> +
> +
> +class LxMounts(gdb.Command):
> +    """Report the VFS mounts of the current process namespace.
> +
> +Equivalent to cat /proc/mounts on a running target
> +An integer value can be supplied to display the mount
> +values of that process namespace"""
> +
> +    def __init__(self):
> +        super(LxMounts, self).__init__("lx-mounts", gdb.COMMAND_DATA)
> +
> +    # Equivalent to proc_namespace.c:show_vfsmnt
> +    # However, that has the ability to call into s_op functions
> +    # whereas we cannot and must make do with the information we can obtain.
> +    def invoke(self, arg, from_tty):
> +        argv = gdb.string_to_argv(arg)
> +        if len(argv) >= 1:
> +            try:
> +                pid = int(argv[0])
> +            except:
> +                raise gdb.GdbError("Provide a PID as integer value")
> +        else:
> +            pid = 1
> +
> +        task = tasks.get_task_by_pid(pid)
> +        if not task:
> +            raise gdb.GdbError("Couldn't find a process with PID {}"
> +                               .format(pid))
> +
> +        namespace = task['nsproxy']['mnt_ns']
> +        if not namespace:
> +            raise gdb.GdbError("No namespace for current process")
> +
> +        for vfs in lists.items(mount_ptr_type, "mnt_list", namespace['list']):
> +            # There appears to be a null entry at the end of the list...

"There appears to be" - hmm... Did you check this against the code?

> +            if not vfs['mnt_parent']:
> +                break
> +
> +            devname = vfs['mnt_devname'].string()
> +            devname = devname if devname else "none"
> +
> +            pathname = ""
> +            parent = vfs
> +            while True:
> +                mntpoint = parent['mnt_mountpoint']
> +                pathname = dentry_name(mntpoint) + pathname

I'm getting an error in this line:

(gdb) lx-mounts
devtmpfs /dev devtmpfs rw,relatime 0 0
tmpfs /dev/shm tmpfs rw,relatime 0 0
devpts /dev/pts devpts rw,relatime 0 0
/dev/sda2 / ext4 rw,relatime 0 0
proc /proc proc rw,nodiratime,relatime 0 0
sysfs /sys sysfs rw,relatime 0 0
debugfs /sys/kernel/debug debugfs rw,relatime 0 0
securityfs /sys/kernel/security securityfs rw,relatime 0 0
fusectl /sys/fs/fuse/connections fusectl rw,relatime 0 0
gvfs-fuse-daemon /home/jan/.gvfs fuse rw,relatime,nosuid,nodev 0 0
tracefs /sys/kernel/debug/tracing tracefs rw,relatime 0 0
Traceback (most recent call last):
  File "/data/linux/build-dbg/scripts/gdb/linux/proc.py", line 185, in invoke
    pathname = dentry_name(mntpoint) + pathname
  File "/data/linux/build-dbg/scripts/gdb/linux/proc.py", line 112, in dentry_name
    p = dentry_name(d['d_parent']) + "/"
  File "/data/linux/build-dbg/scripts/gdb/linux/proc.py", line 110, in dentry_name
    if d['d_parent'] == d:
gdb.MemoryError: Cannot access memory at address 0x40
Error occurred in Python command: Cannot access memory at address 0x40

In this case, the dump should have stopped after the tracefs line.

> +                if (parent == parent['mnt_parent']):
> +                    break
> +                parent = parent['mnt_parent']
> +
> +            if (pathname == ""):
> +                pathname = "/"
> +
> +            superblock = vfs['mnt']['mnt_sb']
> +            fstype = superblock['s_type']['name'].string()
> +            s_flags = int(superblock['s_flags'])
> +            m_flags = int(vfs['mnt']['mnt_flags'])
> +            rd = "ro" if (s_flags & constants.LX_MS_RDONLY) else "rw"
> +
> +            gdb.write(
> +                "{} {} {} {}{}{} 0 0\n"
> +                .format(devname,
> +                        pathname,
> +                        fstype,
> +                        rd,
> +                        info_opts(FS_INFO, s_flags),
> +                        info_opts(MNT_INFO, m_flags)))
> +
> +LxMounts()
> 

Jan

-- 
Siemens AG, Corporate Technology, CT RDA ITP SES-DE
Corporate Competence Center Embedded Linux

[toc] | [prev] | [next] | [standalone]

#1315794

On 23/01/16 15:27, Jan Kiszka wrote:
> On 2016-01-20 12:15, Kieran Bingham wrote:
>> lx-mounts will identify current mount points based on the 'init_task'
>> namespace by default, as we do not yet have a kernel thread list
>> implementation to select the current running thread.
>>
>> Optionally, a user can specify a PID to list from that process'
>> namespace
>>
>> This is somewhat limited vs the /proc/mounts file, as that calls into
>> vfs hooks through the s_op functions to obtain extra information.
>>
>> Signed-off-by: Kieran Bingham <kieran.bingham@linaro.org>
>> ---
>>
>>
>> In this patch, I'm interested in your opinions on coding styles.
>> Would you prefer to see the function helpers, (dentry_name, info_opts) where
>> they are, or inside the command as class members? Or perhaps defined in utils?
> 
> Do you think they could be useful beyond this class? If not, stick them
> inside. Refactoring can still be done once they are needed.
> 

Ok - that's good logic. Keep them close until needed elsewhere :)


>>
>> This also shows where I need to take constant information from the kernel.
>> In this case, they are simple numerical bitflags, and unlikely to change but
>> I didn't want to duplicate their values.
>>
>>
>>  scripts/gdb/linux/constants.py.in |  21 ++++++++
>>  scripts/gdb/linux/proc.py         | 110 ++++++++++++++++++++++++++++++++++++++
>>  2 files changed, 131 insertions(+)
>>
>> diff --git a/scripts/gdb/linux/constants.py.in b/scripts/gdb/linux/constants.py.in
>> index d84084ac945b..739a15d2e984 100644
>> --- a/scripts/gdb/linux/constants.py.in
>> +++ b/scripts/gdb/linux/constants.py.in
>> @@ -12,7 +12,11 @@
>>   *
>>   */
>>  
>> +#include <linux/fs.h>
>> +#include <linux/mount.h>
>> +
>>  /* We need to stringify expanded macros so that they can be parsed */
>>  #define STRING(x) #x
>>  #define XSTRING(x) STRING(x)
>>  
>> @@ -20,3 +24,20 @@
>>  <!-- end-c-headers -->
>>  
>>  import gdb
>> +
>> +/* linux/fs.h */
>> +LX_MS_RDONLY = MS_RDONLY
>> +LX_MS_SYNCHRONOUS = MS_SYNCHRONOUS
>> +LX_MS_MANDLOCK = MS_MANDLOCK
>> +LX_MS_DIRSYNC = MS_DIRSYNC
>> +LX_MS_NOATIME = MS_NOATIME
>> +LX_MS_NODIRATIME = MS_NODIRATIME
>> +
>> +/* linux/mount.h */
>> +LX_MNT_NOSUID = MNT_NOSUID
>> +LX_MNT_NODEV = MNT_NODEV
>> +LX_MNT_NOEXEC = MNT_NOEXEC
>> +LX_MNT_NOATIME = MNT_NOATIME
>> +LX_MNT_NODIRATIME = MNT_NODIRATIME
>> +LX_MNT_RELATIME = MNT_RELATIME
>> +
>> diff --git a/scripts/gdb/linux/proc.py b/scripts/gdb/linux/proc.py
>> index d855b2fd9a06..b79ce2a33a3d 100644
>> --- a/scripts/gdb/linux/proc.py
>> +++ b/scripts/gdb/linux/proc.py
>> @@ -12,6 +12,10 @@
>>  #
>>  
>>  import gdb
>> +from linux import constants
>> +from linux import utils
>> +from linux import tasks
>> +from linux import lists
>>  
>>  
>>  class LxCmdLine(gdb.Command):
>> @@ -96,3 +100,109 @@ Equivalent to cat /proc/ioports on a running target"""
>>          return show_lx_resources("ioport_resource")
>>  
>>  LxIOPorts()
>> +
>> +
>> +# Mount namespace viewer
>> +#  /proc/mounts
>> +
>> +
>> +def dentry_name(d):
>> +    if d['d_parent'] == d:
>> +        return ""
>> +    p = dentry_name(d['d_parent']) + "/"
>> +    return p + d['d_iname'].string()
>> +
>> +
>> +def info_opts(lst, opt):
>> +    opts = ""
>> +    for key, string in lst.items():
>> +        if opt & key:
>> +            opts += string
>> +    return opts
>> +
>> +
>> +FS_INFO = {constants.LX_MS_SYNCHRONOUS: ",sync",
>> +           constants.LX_MS_MANDLOCK: ",mand",
>> +           constants.LX_MS_DIRSYNC: ",dirsync",
>> +           constants.LX_MS_NOATIME: ",noatime",
>> +           constants.LX_MS_NODIRATIME: ",nodiratime"}
>> +
>> +MNT_INFO = {constants.LX_MNT_NOSUID: ",nosuid",
>> +            constants.LX_MNT_NODEV: ",nodev",
>> +            constants.LX_MNT_NOEXEC: ",noexec",
>> +            constants.LX_MNT_NOATIME: ",noatime",
>> +            constants.LX_MNT_NODIRATIME: ",nodiratime",
>> +            constants.LX_MNT_RELATIME: ",relatime"}
>> +
>> +mount_type = utils.CachedType("struct mount")
>> +mount_ptr_type = mount_type.get_type().pointer()
>> +
>> +
>> +class LxMounts(gdb.Command):
>> +    """Report the VFS mounts of the current process namespace.
>> +
>> +Equivalent to cat /proc/mounts on a running target
>> +An integer value can be supplied to display the mount
>> +values of that process namespace"""
>> +
>> +    def __init__(self):
>> +        super(LxMounts, self).__init__("lx-mounts", gdb.COMMAND_DATA)
>> +
>> +    # Equivalent to proc_namespace.c:show_vfsmnt
>> +    # However, that has the ability to call into s_op functions
>> +    # whereas we cannot and must make do with the information we can obtain.
>> +    def invoke(self, arg, from_tty):
>> +        argv = gdb.string_to_argv(arg)
>> +        if len(argv) >= 1:
>> +            try:
>> +                pid = int(argv[0])
>> +            except:
>> +                raise gdb.GdbError("Provide a PID as integer value")
>> +        else:
>> +            pid = 1
>> +
>> +        task = tasks.get_task_by_pid(pid)
>> +        if not task:
>> +            raise gdb.GdbError("Couldn't find a process with PID {}"
>> +                               .format(pid))
>> +
>> +        namespace = task['nsproxy']['mnt_ns']
>> +        if not namespace:
>> +            raise gdb.GdbError("No namespace for current process")
>> +
>> +        for vfs in lists.items(mount_ptr_type, "mnt_list", namespace['list']):
>> +            # There appears to be a null entry at the end of the list...
> 
> "There appears to be" - hmm... Did you check this against the code?


Not properly no... I'll check and see what happened.

I'll have a think as to the best defences to put in.
This one stopped a NULL dereference on mine, but as below clearly it
didn't stop them all!

> 
>> +            if not vfs['mnt_parent']:
>> +                break
>> +
>> +            devname = vfs['mnt_devname'].string()
>> +            devname = devname if devname else "none"
>> +
>> +            pathname = ""
>> +            parent = vfs
>> +            while True:
>> +                mntpoint = parent['mnt_mountpoint']
>> +                pathname = dentry_name(mntpoint) + pathname
> 
> I'm getting an error in this line:
> 
> (gdb) lx-mounts
> devtmpfs /dev devtmpfs rw,relatime 0 0
> tmpfs /dev/shm tmpfs rw,relatime 0 0
> devpts /dev/pts devpts rw,relatime 0 0
> /dev/sda2 / ext4 rw,relatime 0 0
> proc /proc proc rw,nodiratime,relatime 0 0
> sysfs /sys sysfs rw,relatime 0 0
> debugfs /sys/kernel/debug debugfs rw,relatime 0 0
> securityfs /sys/kernel/security securityfs rw,relatime 0 0
> fusectl /sys/fs/fuse/connections fusectl rw,relatime 0 0
> gvfs-fuse-daemon /home/jan/.gvfs fuse rw,relatime,nosuid,nodev 0 0
> tracefs /sys/kernel/debug/tracing tracefs rw,relatime 0 0
> Traceback (most recent call last):
>   File "/data/linux/build-dbg/scripts/gdb/linux/proc.py", line 185, in invoke
>     pathname = dentry_name(mntpoint) + pathname
>   File "/data/linux/build-dbg/scripts/gdb/linux/proc.py", line 112, in dentry_name
>     p = dentry_name(d['d_parent']) + "/"
>   File "/data/linux/build-dbg/scripts/gdb/linux/proc.py", line 110, in dentry_name
>     if d['d_parent'] == d:
> gdb.MemoryError: Cannot access memory at address 0x40
> Error occurred in Python command: Cannot access memory at address 0x40
> 
> In this case, the dump should have stopped after the tracefs line.
> 

Interesting, I'll have to be more defensive.

I'm sure when I looked the proc show command seemed to have only the
list end and the dev name checks...

But we could by our nature run at any time - so it should be a bit more
defensive anyway

>> +                if (parent == parent['mnt_parent']):
>> +                    break
>> +                parent = parent['mnt_parent']
>> +
>> +            if (pathname == ""):
>> +                pathname = "/"
>> +
>> +            superblock = vfs['mnt']['mnt_sb']
>> +            fstype = superblock['s_type']['name'].string()
>> +            s_flags = int(superblock['s_flags'])
>> +            m_flags = int(vfs['mnt']['mnt_flags'])
>> +            rd = "ro" if (s_flags & constants.LX_MS_RDONLY) else "rw"
>> +
>> +            gdb.write(
>> +                "{} {} {} {}{}{} 0 0\n"
>> +                .format(devname,
>> +                        pathname,
>> +                        fstype,
>> +                        rd,
>> +                        info_opts(FS_INFO, s_flags),
>> +                        info_opts(MNT_INFO, m_flags)))
>> +
>> +LxMounts()
>>
> 
> Jan
> 

Kieran

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web