Skip to contentSkip to conversation

Put vdso in ramfs-like filesystem (vdsofs)

RFC In review Dmitry Safonov 9 participants 37 messages Aug 25, 2016 – Sep 21, 2016

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

cover Put vdso in ramfs-like filesystem (vdsofs)

This patches set is cleanly RFC and is not supposed to be applied. Also for RFC time it builds only on x86_64.

So, in a mail thread Oleg told that it would be worth to introduce vm_file for vdso mappings as currently uprobes can not be placed on vDSO VMAs [1]. In this patches set I introduce in-kernel filesystem for vdso files. After patches vDSO VMA now has inode and is just a private file mapping:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0 [vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18 [vdso]

Then I introduce interface in uprobe_events to insert uprobes in vdso. FWIW:
[~]# cd kernel/linux
[linux]# readelf --syms arch/x86/entry/vdso/vdso64.so
Symbol table '.dynsym' contains 11 entries:

   Num:    Value          Size Type    Bind   Vis      Ndx Name
     0: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT  UND 
     1: 0000000000000470     0 SECTION LOCAL  DEFAULT    8 
     2: 00000000000008d0   885 FUNC    WEAK   DEFAULT   12 clock_gettime@@LINUX_2.6
     3: 0000000000000c50   472 FUNC    GLOBAL DEFAULT   12 __vdso_gettimeofday@@LINUX_2.6
     4: 0000000000000c50   472 FUNC    WEAK   DEFAULT   12 gettimeofday@@LINUX_2.6
     5: 0000000000000e30    21 FUNC    GLOBAL DEFAULT   12 __vdso_time@@LINUX_2.6
     6: 0000000000000e30    21 FUNC    WEAK   DEFAULT   12 time@@LINUX_2.6
     7: 00000000000008d0   885 FUNC    GLOBAL DEFAULT   12 __vdso_clock_gettime@@LINUX_2.6
     8: 0000000000000000     0 OBJECT  GLOBAL DEFAULT  ABS LINUX_2.6
     9: 0000000000000e50    41 FUNC    GLOBAL DEFAULT   12 __vdso_getcpu@@LINUX_2.6
    10: 0000000000000e50    41 FUNC    WEAK   DEFAULT   12 getcpu@@LINUX_2.6
  [~]# cd /sys/kernel/debug/tracing/
  [tracing]# echo 'p:clock_gettime :vdso:/64:0x8d0' > uprobe_events
  [tracing]# echo 'p:gettimeofday :vdso:/64:0xc50' >> uprobe_events
  [tracing]# echo 'p:time :vdso:/64:0xe30' >> uprobe_events
  [tracing]# echo 1 > events/uprobes/enable
  [tracing]# su test # it has UID=1001
  [tracing]$ date
  Thu Aug 25 17:19:29 MSK 2016
  [tracing]$ exit
  [tracing]# cat trace
  # tracer: nop
  #
  # entries-in-buffer/entries-written: 175/175   #P:4
  #
  #                              _-----=> irqs-off
  #                             / _----=> need-resched
  #                            | / _---=> hardirq/softirq
  #                            || / _--=> preempt-depth
  #                            ||| /     delay
  #           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
  #              | |       |   ||||       |         |
              bash-11560 [001] d...   316.470236: time: (0x7ffcacebae30)
              bash-11560 [001] d...   316.471436: gettimeofday: (0x7ffcacebac50)
              bash-11560 [001] d...   316.477550: time: (0x7ffcacebae30)
              bash-11560 [001] d...   316.477655: time: (0x7ffcacebae30)
            mktemp-11568 [001] d...   316.479589: gettimeofday: (0x7ffc603f0c50)
              date-11571 [001] d...   316.481890: clock_gettime: (0x7ffec9db58d0)

[...]

If this approach will be decided as fine, I will prepare a better version, fixing the following things:
o put vdsofs in generic fs/* dir
o support other archs and vdso blobs
o remove BUG_ON()'s and UID==1001 check
o remove extern's and use headers only
o refactor code in create_trace_uprobe()
o add some state to (struct trace_uprobe), so i.e., cat uprobe_events will print those uprobes as vdso-based
o document this interface in Documentation/trace/uprobetracer.txt
o prepare nice patches set?

So, opinions? Is it worth to add something like this?

[1]: https://lkml.org/lkml/2016/7/12/346

Dmitry Safonov (3):
x86/vdso: create vdso file, use it for mapping
uprobe: drop isdigit() check in create_trace_uprobe
uprobe: add vdso support

Cc Oleg Nesterov <[email protected]>
Cc Al Viro <[email protected]>
Cc Steven Rostedt <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Dmitry Safonov <[email protected]>
 arch/x86/entry/vdso/vma.c   | 148 ++++++++++++++++++++++++++++++++++++++++++--
 kernel/trace/trace_uprobe.c |  50 +++++++++++----
 2 files changed, 180 insertions(+), 18 deletions(-)
Signature · 1 line
2.9.0
#1 lore raw

RFC 1/3 x86/vdso: create vdso file, use it for mapping

I added here a new in-kernel fs with ramfs-like options. Created vdso file in this fs (yet for testing, only 64-bit vdso). Mapped this file to process's mm on setup_additional_pages. Just for testing purpose it's done only for specific UID.

---
 arch/x86/entry/vdso/vma.c | 146 ++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 140 insertions(+), 6 deletions(-)
arch/x86/entry/vdso/vma.c +140 -6
index f840766659a8..e83830e422ae 100644
@@ -13,6 +13,10 @@
 #include <linux/elf.h>
 #include <linux/cpu.h>
 #include <linux/ptrace.h>
+#include <linux/pagemap.h>
+#include <linux/mount.h>
+#include <linux/ramfs.h>
+#include <linux/file.h>
 #include <asm/pvclock.h>
 #include <asm/vgtod.h>
 #include <asm/proto.h>
@@ -35,6 +39,9 @@ void __init init_vdso_image(const struct vdso_image *image)
 						image->alt_len));
 }
 
+static struct vfsmount *vdso_mnt;
+struct file *vdso_file_64;
+
 struct linux_binprm;
 
 /*
@@ -217,12 +224,31 @@ static int map_vdso(const struct vdso_image *image, bool calculate_addr)
 	/*
 	 * MAYWRITE to allow gdb to COW and set breakpoints
 	 */
-	vma = _install_special_mapping(mm,
-				       text_start,
-				       image->size,
-				       VM_READ|VM_EXEC|
-				       VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC,
-				       &vdso_mapping);
+	if (__kuid_val(task_uid(current)) == 1001) {
+		unsigned long n_addr = mmap_region(vdso_file_64, text_start,
+				image->size, VM_READ|VM_EXEC|
+				VM_DONTEXPAND|VM_SOFTDIRTY|
+				VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC, 0);
+		if (text_start != n_addr) {
+			pr_err("Failed to mmap vdso file at %lx, mmap_region returned %lx\n",
+					text_start, n_addr);
+			goto old_way;
+		}
+		vma = find_vma(mm, text_start);
+		if (IS_ERR(vma) || vma->vm_start != text_start) {
+			pr_err("Failed to find vdso mapped vma at %lx\n",
+					text_start);
+			goto old_way;
+		}
+	} else {
+old_way:
+		vma = _install_special_mapping(mm,
+				text_start,
+				image->size,
+				VM_READ|VM_EXEC|
+				VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC,
+				&vdso_mapping);
+	}
 
 	if (IS_ERR(vma)) {
 		ret = PTR_ERR(vma);
@@ -336,6 +362,109 @@ static int vgetcpu_online(unsigned int cpu)
 	return smp_call_function_single(cpu, vgetcpu_cpu_init, NULL, 1);
 }
 
+static __init int add_vdso_pages_to_page_cache(
+		const struct vdso_image *vdso_image, struct inode *inode)
+{
+	unsigned long i;
+	int ret;
+
+	for (i = 0; i < (vdso_image->size / PAGE_SIZE); i++) {
+		struct page *page = virt_to_page(vdso_image->data +
+						(i << PAGE_SHIFT));
+		int ret;
+
+		__SetPageLocked(page);
+		ret = add_to_page_cache_locked(page, inode->i_mapping,
+						i, __GFP_REPEAT);
+		__SetPageUptodate(page);
+		__ClearPageLocked(page);
+		if (unlikely(ret))
+			goto put_pages;
+	}
+	return 0;
+
+put_pages:
+	while (i > 0)
+		put_page(virt_to_page(vdso_image->data + (i << PAGE_SHIFT)));
+	return ret;
+}
+
+static char *vdso_vma_name(struct dentry *dentry, char *buffer, int buflen)
+{
+	return "[vdso]";
+}
+static const struct dentry_operations vdso_dops = {
+	.d_dname = vdso_vma_name,
+};
+
+static __init struct file *init_vdso_file(const struct vdso_image *vdso_image,
+					const char *name)
+{
+	struct super_block *sb;
+	struct qstr name_str;
+	struct inode *inode;
+	struct path path;
+	struct file *res;
+
+	if (IS_ERR(vdso_mnt))
+		return ERR_CAST(vdso_mnt);
+	sb = vdso_mnt->mnt_sb;
+
+	name_str.hash = 0;
+	name_str.len = strlen(name);
+	name_str.name = name;
+
+	res = ERR_PTR(-ENOMEM);
+	path.mnt = mntget(vdso_mnt);
+	path.dentry = d_alloc_pseudo(sb, &name_str);
+	if (!path.dentry)
+		goto put_path;
+	d_set_d_op(path.dentry, &vdso_dops);
+
+	res = ERR_PTR(-ENOSPC);
+	inode = ramfs_get_inode(sb, NULL, S_IFREG | S_IRUGO | S_IXUGO, 0);
+	if (!inode)
+		goto put_path;
+
+	inode->i_flags |= S_PRIVATE;
+	d_instantiate(path.dentry, inode);
+	inode->i_size = vdso_image->size;
+
+	res = ERR_PTR(add_vdso_pages_to_page_cache(vdso_image, inode));
+	if (IS_ERR(res))
+		goto put_path;
+
+	res = alloc_file(&path, FMODE_READ, &ramfs_file_operations);
+	if (!IS_ERR(res))
+		return res;
+
+put_path:
+	path_put(&path);
+	return res;
+}
+
+
+static struct file_system_type vdso_fs_type = {
+	.name		= "vdsofs",
+	.mount		= ramfs_mount,
+	.kill_sb	= kill_litter_super,
+};
+
+static int __init init_vdso_fs(void)
+{
+	int ret;
+
+	ret = register_filesystem(&vdso_fs_type);
+	if (ret)
+		return ret;
+
+	vdso_mnt = kern_mount(&vdso_fs_type);
+	if (IS_ERR(vdso_mnt))
+		return PTR_ERR(vdso_mnt);
+	return 0;
+}
+
+/* XXX: replace BUG_ON with return to old-way vdso handling */
 static int __init init_vdso(void)
 {
 	init_vdso_image(&vdso_image_64);
@@ -344,6 +473,11 @@ static int __init init_vdso(void)
 	init_vdso_image(&vdso_image_x32);
 #endif
 
+	BUG_ON(init_vdso_fs());
+
+	vdso_file_64 = init_vdso_file(&vdso_image_64, "vdso_image_64");
+	BUG_ON(IS_ERR(vdso_file_64));
+
 	/* notifier priority > KVM */
 	return cpuhp_setup_state(CPUHP_AP_X86_VDSO_VMA_ONLINE,
 				 "AP_X86_VDSO_VMA_ONLINE", vgetcpu_online, NULL);
Signature · 1 line
2.9.0
#2 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

2016-08-25 18:21 GMT+03:00 Dmitry Safonov <[email protected]>:

arch/x86/entry/vdso/vma.c static int vgetcpu_online(unsigned int cpu) · line 394

Quoted 3 lines
arch/x86/entry/vdso/vma.c +3 -0
+static char *vdso_vma_name(struct dentry *dentry, char *buffer, int buflen)
+{
+       return "[vdso]";

It should be:
+ return dynamic_dname(dentry, buffer, buflen, "[vdso]");
returned pointer should be inside buffer.

Signature · 1 line
             Dmitry
#3 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

Quoted 9 lines · Dmitry Safonov

2016-08-25 18:21 GMT+03:00 Dmitry Safonov <[email protected]>:

arch/x86/entry/vdso/vma.c static int vgetcpu_online(unsigned int cpu) · line 394

Quoted 3 lines
arch/x86/entry/vdso/vma.c +3 -0
+static char *vdso_vma_name(struct dentry *dentry, char *buffer, int buflen)
+{
+       return "[vdso]";

It should be:
+ return dynamic_dname(dentry, buffer, buflen, "[vdso]");
returned pointer should be inside buffer.

Funny thing: after this fixup, I can easily get vdso blob image with:
[root@localhost ~]# dd if=/proc/11486/map_files/7ffd26596000-7ffd26598000 of=./vdso 16+0 records in
16+0 records out
8192 bytes (8.2 kB) copied, 0.000101732 s, 80.5 MB/s
[root@localhost ~]# objdump -dS ./vdso | head

./vdso: file format elf64-x86-64

Disassembly of section .text:

00000000000008d0 <__vdso_clock_gettime>:

 8d0:	55                   	push   %rbp
 8d1:	83 ff 01             	cmp    $0x1,%edi
 8d4:	48 89 e5             	mov    %rsp,%rbp
Signature · 1 line
             Dmitry
#4 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

In reply to Dmitry Safonov
Quoted 6 lines · Dmitry Safonov

I added here a new in-kernel fs with ramfs-like options. Created vdso file in this fs (yet for testing, only 64-bit vdso). Mapped this file to process's mm on setup_additional_pages. Just for testing purpose it's done only for specific UID.

Dmitry, could you clarify please, why "old way" remains in the code? Even if RFC is not the new approach is supposing to completely remove old one? Or I miss something obvious?

#5 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

Quoted 8 lines · Cyrill Gorcunov
Quoted 6 lines · Dmitry Safonov

I added here a new in-kernel fs with ramfs-like options. Created vdso file in this fs (yet for testing, only 64-bit vdso). Mapped this file to process's mm on setup_additional_pages. Just for testing purpose it's done only for specific UID.

Hi Cyrill,

Quoted 3 lines

Dmitry, could you clarify please, why "old way" remains in the code? Even if RFC is not the new approach is supposing to completely remove old one? Or I miss something obvious?

Well, I left it for RFC, I think on post-RFC we can leave it as cmd-param (alike vdso{64,32}_enabled) so it could be easily disabled if it brokes something in userspace. I guess, that's preferred way to introduce something like that.

Signature · 1 line
             Dmitry
#6 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

In reply to Dmitry Safonov
Quoted 4 lines · Dmitry Safonov

I added here a new in-kernel fs with ramfs-like options. Created vdso file in this fs (yet for testing, only 64-bit vdso). Mapped this file to process's mm on setup_additional_pages. Just for testing purpose it's done only for specific UID.

I'm wondering whether all this code could be easily moved into the core special mapping helpers so that all special mappings get the same benefit. We could embed a struct file * (or struct inode or whatever) in special_mapping if needed.

Also, could this be simplified to use anon_inode?

(I'm not a VFS expert at all, so I could be way off base.)

#7 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

Quoted 10 lines · Andy Lutomirski
Quoted 4 lines · Dmitry Safonov

I added here a new in-kernel fs with ramfs-like options. Created vdso file in this fs (yet for testing, only 64-bit vdso). Mapped this file to process's mm on setup_additional_pages. Just for testing purpose it's done only for specific UID.

I'm wondering whether all this code could be easily moved into the core special mapping helpers so that all special mappings get the same benefit. We could embed a struct file * (or struct inode or whatever) in special_mapping if needed.

Hmm, yes, I guess. The only thing -- we'll still need per-arch changes to initialize those files on booting. But that looks like the proper generic place to move this code.

Quoted 3 lines

Also, could this be simplified to use anon_inode?

(I'm not a VFS expert at all, so I could be way off base.)

Well, I'll try to do the second version with anon_inode.

Thanks,
Dmitry

#8 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

Quoted 17 lines · Dmitry Safonov
Quoted 10 lines · Andy Lutomirski
Quoted 5 lines · Dmitry Safonov

I added here a new in-kernel fs with ramfs-like options. Created vdso file in this fs (yet for testing, only 64-bit vdso). Mapped this file to process's mm on setup_additional_pages. Just for testing purpose it's done only for specific UID.

Quoted 4 lines

I'm wondering whether all this code could be easily moved into the core special mapping helpers so that all special mappings get the same benefit. We could embed a struct file * (or struct inode or whatever) in special_mapping if needed.

Quoted 3 lines

Hmm, yes, I guess. The only thing -- we'll still need per-arch changes to initialize those files on booting. But that looks like the proper generic place to move this code.

You might be able to get away with initializing on first use.

--Andy

#9 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

In reply to Andy Lutomirski
Quoted 6 lines · Andy Lutomirski
Quoted 4 lines · Dmitry Safonov

I added here a new in-kernel fs with ramfs-like options. Created vdso file in this fs (yet for testing, only 64-bit vdso). Mapped this file to process's mm on setup_additional_pages. Just for testing purpose it's done only for specific UID.

Quoted 6 lines

I'm wondering whether all this code could be easily moved into the core special mapping helpers so that all special mappings get the same benefit. We could embed a struct file * (or struct inode or whatever) in special_mapping if needed.

Also, could this be simplified to use anon_inode?

Please, don't. anon_inode is for situations when you don't mind sharing the _same_ inode for different things. This one very clearly isn't that.

#10 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

In reply to Dmitry Safonov
Quoted 36 lines · Dmitry Safonov
arch/x86/entry/vdso/vma.c +35 -0
+static __init struct file *init_vdso_file(const struct vdso_image *vdso_image,
+					const char *name)
+{
+	struct super_block *sb;
+	struct qstr name_str;
+	struct inode *inode;
+	struct path path;
+	struct file *res;
+
+	if (IS_ERR(vdso_mnt))
+		return ERR_CAST(vdso_mnt);
+	sb = vdso_mnt->mnt_sb;
+
+	name_str.hash = 0;
+	name_str.len = strlen(name);
+	name_str.name = name;
+
+	res = ERR_PTR(-ENOMEM);
+	path.mnt = mntget(vdso_mnt);
+	path.dentry = d_alloc_pseudo(sb, &name_str);
+	if (!path.dentry)
+		goto put_path;
+	d_set_d_op(path.dentry, &vdso_dops);
+
+	res = ERR_PTR(-ENOSPC);
+	inode = ramfs_get_inode(sb, NULL, S_IFREG | S_IRUGO | S_IXUGO, 0);
+	if (!inode)
+		goto put_path;
+
+	inode->i_flags |= S_PRIVATE;
+	d_instantiate(path.dentry, inode);
+	inode->i_size = vdso_image->size;
+
+	res = ERR_PTR(add_vdso_pages_to_page_cache(vdso_image, inode));
+	if (IS_ERR(res))

arch/x86/entry/vdso/vma.c static int vgetcpu_online(unsigned int cpu) · line 444

Quoted 10 lines
arch/x86/entry/vdso/vma.c +10 -0
+		goto put_path;
+
+	res = alloc_file(&path, FMODE_READ, &ramfs_file_operations);
+	if (!IS_ERR(res))
+		return res;
+
+put_path:
+	path_put(&path);
+	return res;
+}

Not sure... I think alloc_anon_inode() makes more sense.

Perhaps you can look at fs/aio.c and extract some bits of code into the new helpers which could be used by both aio_fs and vdso_fs?

Oleg.

#11 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

2016-08-30 17:33 GMT+03:00 Oleg Nesterov <[email protected]>:

Quoted 50 lines · Dmitry Safonov
Quoted 36 lines
arch/x86/entry/vdso/vma.c +35 -0
+static __init struct file *init_vdso_file(const struct vdso_image *vdso_image,
+                                     const char *name)
+{
+     struct super_block *sb;
+     struct qstr name_str;
+     struct inode *inode;
+     struct path path;
+     struct file *res;
+
+     if (IS_ERR(vdso_mnt))
+             return ERR_CAST(vdso_mnt);
+     sb = vdso_mnt->mnt_sb;
+
+     name_str.hash = 0;
+     name_str.len = strlen(name);
+     name_str.name = name;
+
+     res = ERR_PTR(-ENOMEM);
+     path.mnt = mntget(vdso_mnt);
+     path.dentry = d_alloc_pseudo(sb, &name_str);
+     if (!path.dentry)
+             goto put_path;
+     d_set_d_op(path.dentry, &vdso_dops);
+
+     res = ERR_PTR(-ENOSPC);
+     inode = ramfs_get_inode(sb, NULL, S_IFREG | S_IRUGO | S_IXUGO, 0);
+     if (!inode)
+             goto put_path;
+
+     inode->i_flags |= S_PRIVATE;
+     d_instantiate(path.dentry, inode);
+     inode->i_size = vdso_image->size;
+
+     res = ERR_PTR(add_vdso_pages_to_page_cache(vdso_image, inode));
+     if (IS_ERR(res))

arch/x86/entry/vdso/vma.c static int vgetcpu_online(unsigned int cpu) · line 444

Quoted 10 lines
arch/x86/entry/vdso/vma.c +10 -0
+             goto put_path;
+
+     res = alloc_file(&path, FMODE_READ, &ramfs_file_operations);
+     if (!IS_ERR(res))
+             return res;
+
+put_path:
+     path_put(&path);
+     return res;
+}

Not sure... I think alloc_anon_inode() makes more sense.

Quoted 2 lines

Perhaps you can look at fs/aio.c and extract some bits of code into the new helpers which could be used by both aio_fs and vdso_fs?

Thanks, will do for the next version.

Signature · 1 line
             Dmitry
#12 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

In reply to Oleg Nesterov
Quoted 3 lines · Oleg Nesterov

arch/x86/entry/vdso/vma.c static int vgetcpu_online(unsigned int cpu) · line 425

Quoted 1 line
arch/x86/entry/vdso/vma.c +1 -0
+	inode = ramfs_get_inode(sb, NULL, S_IFREG | S_IRUGO | S_IXUGO, 0);

Not sure... I think alloc_anon_inode() makes more sense.

Compared to this, you mean? It's going to give you the wrong
permissions/i_op/a_ops, and you'll have to assign them manually. ramfs already gives the right ones for what's wanted, AFAICS.

#13 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

In reply to Dmitry Safonov
Quoted 4 lines · Dmitry Safonov
arch/x86/entry/vdso/vma.c +4 -0
+		unsigned long n_addr = mmap_region(vdso_file_64, text_start,
+				image->size, VM_READ|VM_EXEC|
+				VM_DONTEXPAND|VM_SOFTDIRTY|
+				VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC, 0);

arch/x86/entry/vdso/vma.c static int map_vdso(const struct vdso_image *image, bool calculate_addr) · line 241

Quoted 10 lines
arch/x86/entry/vdso/vma.c +10 -0
+		if (text_start != n_addr) {
+			pr_err("Failed to mmap vdso file at %lx, mmap_region returned %lx\n",
+					text_start, n_addr);
+			goto old_way;
+		}
+		vma = find_vma(mm, text_start);
+		if (IS_ERR(vma) || vma->vm_start != text_start) {
+			pr_err("Failed to find vdso mapped vma at %lx\n",
+					text_start);
+			goto old_way;

Umm... Since when can find_vma() return ERR_PTR()?

arch/x86/entry/vdso/vma.c static int vgetcpu_online(unsigned int cpu) · line 422

Quoted 1 line
arch/x86/entry/vdso/vma.c +1 -0
+	d_set_d_op(path.dentry, &vdso_dops);

Nope. Set ->s_d_op to &vdso_dops and be done with that.

arch/x86/entry/vdso/vma.c static int vgetcpu_online(unsigned int cpu) · line 449

Quoted 3 lines
arch/x86/entry/vdso/vma.c +3 -0
+static struct file_system_type vdso_fs_type = {
+	.name		= "vdsofs",
+	.mount		= ramfs_mount,

Probably the wrong thing here. Just use a simple wrapper using mount_pseudo() for all work; see fs/aio.c:aio_mount().

arch/x86/entry/vdso/vma.c static int vgetcpu_online(unsigned int cpu) · line 457

Quoted 1 line
arch/x86/entry/vdso/vma.c +1 -0
+	ret = register_filesystem(&vdso_fs_type);

Do you really want it user-mountable? If not, no need to register...

#14 lore raw

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

Big thanks on review, Al!

2016-09-03 3:20 GMT+03:00 Al Viro <[email protected]>:

Quoted 16 lines · Dmitry Safonov
Quoted 4 lines
arch/x86/entry/vdso/vma.c +4 -0
+             unsigned long n_addr = mmap_region(vdso_file_64, text_start,
+                             image->size, VM_READ|VM_EXEC|
+                             VM_DONTEXPAND|VM_SOFTDIRTY|
+                             VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC, 0);

arch/x86/entry/vdso/vma.c static int map_vdso(const struct vdso_image *image, bool calculate_addr) · line 241

Quoted 10 lines
arch/x86/entry/vdso/vma.c +10 -0
+             if (text_start != n_addr) {
+                     pr_err("Failed to mmap vdso file at %lx, mmap_region returned %lx\n",
+                                     text_start, n_addr);
+                     goto old_way;
+             }
+             vma = find_vma(mm, text_start);
+             if (IS_ERR(vma) || vma->vm_start != text_start) {
+                     pr_err("Failed to find vdso mapped vma at %lx\n",
+                                     text_start);
+                     goto old_way;
Quoted 1 line

Umm... Since when can find_vma() return ERR_PTR()?

Will fix

Quoted 3 lines
Quoted 1 line

+ d_set_d_op(path.dentry, &vdso_dops);

Nope. Set ->s_d_op to &vdso_dops and be done with that.

Ok

Quoted 6 lines

arch/x86/entry/vdso/vma.c static int vgetcpu_online(unsigned int cpu) · line 449

Quoted 3 lines
arch/x86/entry/vdso/vma.c +3 -0
+static struct file_system_type vdso_fs_type = {
+     .name           = "vdsofs",
+     .mount          = ramfs_mount,

Probably the wrong thing here. Just use a simple wrapper using mount_pseudo() for all work; see fs/aio.c:aio_mount().

Will check, thanks!

Quoted 3 lines
Quoted 1 line

+ ret = register_filesystem(&vdso_fs_type);

Do you really want it user-mountable? If not, no need to register...

Of course, I don't want to, will omit that, thanks.

Signature · 1 line
             Dmitry
#15 lore raw

RFC 2/3 uprobe: drop isdigit() check in create_trace_uprobe

In reply to Dmitry Safonov

It's useless. Before:
[tracing]# echo 'p:test /a:0x0' >> uprobe_events [tracing]# echo 'p:test a:0x0' >> uprobe_events -bash: echo: write error: No such file or directory [tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: Invalid argument

After:
[tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: No such file or directory

---
 kernel/trace/trace_uprobe.c | 4 ----
 1 file changed, 4 deletions(-)
kernel/trace/trace_uprobe.c +0 -4
index c53485441c88..a74f2d9ff379 100644
@@ -427,10 +427,6 @@ static int create_trace_uprobe(int argc, char **argv)
 		pr_info("Probe point is not specified.\n");
 		return -EINVAL;
 	}
-	if (isdigit(argv[1][0])) {
-		pr_info("probe point must be have a filename.\n");
-		return -EINVAL;
-	}
 	arg = strchr(argv[1], ':');
 	if (!arg) {
 		ret = -EINVAL;
Signature · 1 line
2.9.0
maintainer Acked-by on v1 2/3
#16 lore raw

Re: [RFC 2/3] uprobe: drop isdigit() check in create_trace_uprobe

Quoted 7 lines · On Thu, 25 Aug 2016 18:21:09 +0300 Dmitry Safonov <[email protected]>…

It's useless. Before:
[tracing]# echo 'p:test /a:0x0' >> uprobe_events [tracing]# echo 'p:test a:0x0' >> uprobe_events -bash: echo: write error: No such file or directory [tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: Invalid argument

Quoted 5 lines

After:
[tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: No such file or directory

-- Steve

Quoted 19 lines
---
 kernel/trace/trace_uprobe.c | 4 ----
 1 file changed, 4 deletions(-)
kernel/trace/trace_uprobe.c +0 -4
index c53485441c88..a74f2d9ff379 100644
@@ -427,10 +427,6 @@ static int create_trace_uprobe(int argc, char **argv)
 		pr_info("Probe point is not specified.\n");
 		return -EINVAL;
 	}
-	if (isdigit(argv[1][0])) {
-		pr_info("probe point must be have a filename.\n");
-		return -EINVAL;
-	}
 	arg = strchr(argv[1], ':');
 	if (!arg) {
 		ret = -EINVAL;
maintainer
#17 lore raw

Re: [RFC 2/3] uprobe: drop isdigit() check in create_trace_uprobe

Quoted 10 lines · On Mon, 29 Aug 2016 18:58:13 -0400 Steven Rostedt <[email protected]> wrote:
Quoted 7 lines · On Thu, 25 Aug 2016 18:21:09 +0300 Dmitry Safonov <[email protected]>…

It's useless. Before:
[tracing]# echo 'p:test /a:0x0' >> uprobe_events [tracing]# echo 'p:test a:0x0' >> uprobe_events -bash: echo: write error: No such file or directory [tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: Invalid argument

Quoted 7 lines
Quoted 5 lines

After:
[tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: No such file or directory

Actually, this patch seems agnostic to the series. I'll just pull it in now.

-- Steve

Quoted 21 lines
Quoted 20 lines
---
 kernel/trace/trace_uprobe.c | 4 ----
 1 file changed, 4 deletions(-)
kernel/trace/trace_uprobe.c +0 -4
index c53485441c88..a74f2d9ff379 100644
@@ -427,10 +427,6 @@ static int create_trace_uprobe(int argc, char **argv)
 		pr_info("Probe point is not specified.\n");
 		return -EINVAL;
 	}
-	if (isdigit(argv[1][0])) {
-		pr_info("probe point must be have a filename.\n");
-		return -EINVAL;
-	}
 	arg = strchr(argv[1], ':');
 	if (!arg) {
 		ret = -EINVAL;  
#18 lore raw

Re: [RFC 2/3] uprobe: drop isdigit() check in create_trace_uprobe

2016-08-30 1:59 GMT+03:00 Steven Rostedt <[email protected]>:

Quoted 17 lines
Quoted 14 lines · On Mon, 29 Aug 2016 18:58:13 -0400 Steven Rostedt <[email protected]> wrote:
Quoted 11 lines · On Thu, 25 Aug 2016 18:21:09 +0300 Dmitry Safonov <[email protected]>…

It's useless. Before:
[tracing]# echo 'p:test /a:0x0' >> uprobe_events [tracing]# echo 'p:test a:0x0' >> uprobe_events -bash: echo: write error: No such file or directory [tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: Invalid argument

After:
[tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: No such file or directory

Quoted 6 lines
Quoted 3 lines
Quoted 1 line

Actually, this patch seems agnostic to the series. I'll just pull it in now.

Thanks, Steve!

Signature · 1 line
             Dmitry
#19 lore raw

Re: [RFC 2/3] uprobe: drop isdigit() check in create_trace_uprobe

In reply to Steven Rostedt
Quoted 18 lines · Steven Rostedt
Quoted 14 lines · On Mon, 29 Aug 2016 18:58:13 -0400 Steven Rostedt <[email protected]> wrote:
Quoted 11 lines · On Thu, 25 Aug 2016 18:21:09 +0300 Dmitry Safonov <[email protected]>…

It's useless. Before:
[tracing]# echo 'p:test /a:0x0' >> uprobe_events [tracing]# echo 'p:test a:0x0' >> uprobe_events -bash: echo: write error: No such file or directory [tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: Invalid argument

After:
[tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: No such file or directory

Quoted 6 lines
Quoted 3 lines
Quoted 1 line

Actually, this patch seems agnostic to the series. I'll just pull it in now.

Yes, agreed, this isdigit() is pointless, thanks.

Oleg.

maintainer
#20 lore raw

Re: [RFC 2/3] uprobe: drop isdigit() check in create_trace_uprobe

Quoted 19 lines · On Tue, 30 Aug 2016 16:37:28 +0200 Oleg Nesterov <[email protected]> wrote:
Quoted 18 lines · Steven Rostedt
Quoted 14 lines · On Mon, 29 Aug 2016 18:58:13 -0400 Steven Rostedt <[email protected]> wrote:
Quoted 11 lines · On Thu, 25 Aug 2016 18:21:09 +0300 Dmitry Safonov <[email protected]>…

It's useless. Before:
[tracing]# echo 'p:test /a:0x0' >> uprobe_events [tracing]# echo 'p:test a:0x0' >> uprobe_events -bash: echo: write error: No such file or directory [tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: Invalid argument

After:
[tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: No such file or directory

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
Quoted 1 line

Actually, this patch seems agnostic to the series. I'll just pull it in now.

Yes, agreed, this isdigit() is pointless, thanks.

Should I put you down as an Acked-by?

-- Steve

#21 lore raw

Re: [RFC 2/3] uprobe: drop isdigit() check in create_trace_uprobe

Quoted 7 lines · Steven Rostedt
Quoted 3 lines · On Tue, 30 Aug 2016 16:37:28 +0200 Oleg Nesterov <[email protected]> wrote:
Quoted 2 lines · Steven Rostedt
Quoted 6 lines
Quoted 4 lines
Quoted 2 lines

Actually, this patch seems agnostic to the series. I'll just pull it in now.

Yes, agreed, this isdigit() is pointless, thanks.

Should I put you down as an Acked-by?

Thanks, feel free to add my ack.

Oleg.

#22 lore raw

Re: [RFC 2/3] uprobe: drop isdigit() check in create_trace_uprobe

In reply to Steven Rostedt
Quoted 10 lines · Steven Rostedt
Quoted 7 lines · On Thu, 25 Aug 2016 18:21:09 +0300 Dmitry Safonov <[email protected]>…

It's useless. Before:
[tracing]# echo 'p:test /a:0x0' >> uprobe_events [tracing]# echo 'p:test a:0x0' >> uprobe_events -bash: echo: write error: No such file or directory [tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: Invalid argument

Quoted 7 lines
Quoted 5 lines

After:
[tracing]# echo 'p:test 1:0x0' >> uprobe_events -bash: echo: write error: No such file or directory

Agree.

Signature · 2 lines
Thanks and Regards
Srikar Dronamraju
#23 lore raw

RFC 3/3 uprobe: add vdso support

In reply to Dmitry Safonov

For RFC, this patch is full of hacks:
- non-generic vdso_mnt -- it will not built on !x86_64
- used externs in trace_uprobe.c, rather than clean header-includes

This patch converts vdsofs from lookup-less, renames vdso file name to just "64". It also adds support to uprobe_events for setting uprobe to vdso image file.

One can set vdso uprobe now like this:
echo 'p:gettimeofday :vdso:/64:0xc50' >> uprobe_events

---
 arch/x86/entry/vdso/vma.c   | 28 ++++++++++++++-------------
 kernel/trace/trace_uprobe.c | 46 +++++++++++++++++++++++++++++++++++++--------
 2 files changed, 53 insertions(+), 21 deletions(-)
arch/x86/entry/vdso/vma.c +15 -13
index e83830e422ae..9e966b649028 100644
@@ -39,7 +39,7 @@ void __init init_vdso_image(const struct vdso_image *image)
 						image->alt_len));
 }
 
-static struct vfsmount *vdso_mnt;
+struct vfsmount *vdso_mnt;
 struct file *vdso_file_64;
 
 struct linux_binprm;
@@ -401,38 +401,40 @@ static __init struct file *init_vdso_file(const struct vdso_image *vdso_image,
 					const char *name)
 {
 	struct super_block *sb;
-	struct qstr name_str;
 	struct inode *inode;
 	struct path path;
 	struct file *res;
+	int ret;
 
 	if (IS_ERR(vdso_mnt))
 		return ERR_CAST(vdso_mnt);
 	sb = vdso_mnt->mnt_sb;
 
-	name_str.hash = 0;
-	name_str.len = strlen(name);
-	name_str.name = name;
-
 	res = ERR_PTR(-ENOMEM);
 	path.mnt = mntget(vdso_mnt);
-	path.dentry = d_alloc_pseudo(sb, &name_str);
-	if (!path.dentry)
+	path.dentry = d_alloc_name(vdso_mnt->mnt_root, name);
+	if (!path.dentry) {
+		pr_err("Failed to allocate dentry for %s vdso file\n", name);
 		goto put_path;
+	}
 	d_set_d_op(path.dentry, &vdso_dops);
 
 	res = ERR_PTR(-ENOSPC);
 	inode = ramfs_get_inode(sb, NULL, S_IFREG | S_IRUGO | S_IXUGO, 0);
-	if (!inode)
+	if (!inode) {
+		pr_err("Failed to get inode for %s vdso file\n", name);
 		goto put_path;
+	}
 
 	inode->i_flags |= S_PRIVATE;
-	d_instantiate(path.dentry, inode);
+	d_add(path.dentry, inode);
 	inode->i_size = vdso_image->size;
 
-	res = ERR_PTR(add_vdso_pages_to_page_cache(vdso_image, inode));
-	if (IS_ERR(res))
+	ret = add_vdso_pages_to_page_cache(vdso_image, inode);
+	if (ret) {
+		pr_err("Failed to put %s pages in page cache: %d\n", name, ret);
 		goto put_path;
+	}
 
 	res = alloc_file(&path, FMODE_READ, &ramfs_file_operations);
 	if (!IS_ERR(res))
@@ -475,7 +477,7 @@ static int __init init_vdso(void)
 
 	BUG_ON(init_vdso_fs());
 
-	vdso_file_64 = init_vdso_file(&vdso_image_64, "vdso_image_64");
+	vdso_file_64 = init_vdso_file(&vdso_image_64, "64");
 	BUG_ON(IS_ERR(vdso_file_64));
 
 	/* notifier priority > KVM */
kernel/trace/trace_uprobe.c +38 -8

65 lines not on this page · raw message

Signature · 1 line
2.9.0
#24 lore raw
In reply to Dmitry Safonov
Quoted 101 lines · Dmitry Safonov

This patches set is cleanly RFC and is not supposed to be applied. Also for RFC time it builds only on x86_64.

So, in a mail thread Oleg told that it would be worth to introduce vm_file
for vdso mappings as currently uprobes can not be placed on vDSO VMAs [1].
In this patches set I introduce in-kernel filesystem for vdso files. After patches vDSO VMA now has inode and is just a private file mapping:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0
[vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18
[vdso]

Then I introduce interface in uprobe_events to insert uprobes in vdso. FWIW:
[~]# cd kernel/linux
[linux]# readelf --syms arch/x86/entry/vdso/vdso64.so
Symbol table '.dynsym' contains 11 entries:

  Num:    Value          Size Type    Bind   Vis      Ndx Name
    0: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT  UND 
    1: 0000000000000470     0 SECTION LOCAL  DEFAULT    8 

2: 00000000000008d0 885 FUNC WEAK DEFAULT 12
clock_gettime@@LINUX_2.6
3: 0000000000000c50 472 FUNC GLOBAL DEFAULT 12
__vdso_gettimeofday@@LINUX_2.6
4: 0000000000000c50 472 FUNC WEAK DEFAULT 12
gettimeofday@@LINUX_2.6
5: 0000000000000e30 21 FUNC GLOBAL DEFAULT 12
__vdso_time@@LINUX_2.6
6: 0000000000000e30 21 FUNC WEAK DEFAULT 12 time@@LINUX_2.6
7: 00000000000008d0 885 FUNC GLOBAL DEFAULT 12
__vdso_clock_gettime@@LINUX_2.6

    8: 0000000000000000     0 OBJECT  GLOBAL DEFAULT  ABS LINUX_2.6

9: 0000000000000e50 41 FUNC GLOBAL DEFAULT 12
__vdso_getcpu@@LINUX_2.6
10: 0000000000000e50 41 FUNC WEAK DEFAULT 12
getcpu@@LINUX_2.6
[~]# cd /sys/kernel/debug/tracing/
[tracing]# echo 'p:clock_gettime :vdso:/64:0x8d0' > uprobe_events [tracing]# echo 'p:gettimeofday :vdso:/64:0xc50' >> uprobe_events [tracing]# echo 'p:time :vdso:/64:0xe30' >> uprobe_events [tracing]# echo 1 > events/uprobes/enable
[tracing]# su test # it has UID=1001
[tracing]$ date
Thu Aug 25 17:19:29 MSK 2016
[tracing]$ exit
[tracing]# cat trace
# tracer: nop
#
# entries-in-buffer/entries-written: 175/175 #P:4
#
# _-----=> irqs-off
# / _----=> need-resched

 #                            | / _---=> hardirq/softirq
 #                            || / _--=> preempt-depth
 #                            ||| /     delay
 #           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
 #              | |       |   ||||       |         |
            bash-11560 [001] d...   316.470236: time: (0x7ffcacebae30)
    bash-11560 [001] d...   316.471436: gettimeofday: (0x7ffcacebac50)
            bash-11560 [001] d...   316.477550: time: (0x7ffcacebae30)
            bash-11560 [001] d...   316.477655: time: (0x7ffcacebae30)

mktemp-11568 [001] d... 316.479589: gettimeofday: (0x7ffc603f0c50)
date-11571 [001] d... 316.481890: clock_gettime: (0x7ffec9db58d0)
[...]

If this approach will be decided as fine, I will prepare a better version,
fixing the following things:
o put vdsofs in generic fs/* dir
o support other archs and vdso blobs
o remove BUG_ON()'s and UID==1001 check
o remove extern's and use headers only
o refactor code in create_trace_uprobe()
o add some state to (struct trace_uprobe), so i.e., cat uprobe_events will
print those uprobes as vdso-based
o document this interface in Documentation/trace/uprobetracer.txt
o prepare nice patches set?

So, opinions? Is it worth to add something like this?

[1]: https://lkml.org/lkml/2016/7/12/346

Dmitry Safonov (3):
x86/vdso: create vdso file, use it for mapping
uprobe: drop isdigit() check in create_trace_uprobe
uprobe: add vdso support

Cc Oleg Nesterov <[email protected]>
Cc Al Viro <[email protected]>
Cc Steven Rostedt <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Dmitry Safonov <[email protected]>
Quoted 4 lines

arch/x86/entry/vdso/vma.c | 148
++++++++++++++++++++++++++++++++++++++++++--
kernel/trace/trace_uprobe.c | 50 +++++++++++----

2 files changed, 180 insertions(+), 18 deletions(-)

I think there is a lot to be said for this idea. However, a private mapping is definitely wrong for the vvar data; for the vdso code it could be considered either way I suppose.

Signature · 1 line
Sent from my Android device with K-9 Mail. Please excuse brevity and formatting.
#25 lore raw

2016-08-25 23:49 GMT+03:00 H. Peter Anvin <[email protected]>:

Quoted 102 lines · Dmitry Safonov
Quoted 101 lines

This patches set is cleanly RFC and is not supposed to be applied. Also for RFC time it builds only on x86_64.

So, in a mail thread Oleg told that it would be worth to introduce vm_file
for vdso mappings as currently uprobes can not be placed on vDSO VMAs [1].
In this patches set I introduce in-kernel filesystem for vdso files. After patches vDSO VMA now has inode and is just a private file mapping:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0
[vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18
[vdso]

Then I introduce interface in uprobe_events to insert uprobes in vdso. FWIW:
[~]# cd kernel/linux
[linux]# readelf --syms arch/x86/entry/vdso/vdso64.so
Symbol table '.dynsym' contains 11 entries:

  Num:    Value          Size Type    Bind   Vis      Ndx Name
    0: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT  UND
    1: 0000000000000470     0 SECTION LOCAL  DEFAULT    8

2: 00000000000008d0 885 FUNC WEAK DEFAULT 12
clock_gettime@@LINUX_2.6
3: 0000000000000c50 472 FUNC GLOBAL DEFAULT 12
__vdso_gettimeofday@@LINUX_2.6
4: 0000000000000c50 472 FUNC WEAK DEFAULT 12
gettimeofday@@LINUX_2.6
5: 0000000000000e30 21 FUNC GLOBAL DEFAULT 12
__vdso_time@@LINUX_2.6
6: 0000000000000e30 21 FUNC WEAK DEFAULT 12 time@@LINUX_2.6
7: 00000000000008d0 885 FUNC GLOBAL DEFAULT 12
__vdso_clock_gettime@@LINUX_2.6

    8: 0000000000000000     0 OBJECT  GLOBAL DEFAULT  ABS LINUX_2.6

9: 0000000000000e50 41 FUNC GLOBAL DEFAULT 12
__vdso_getcpu@@LINUX_2.6
10: 0000000000000e50 41 FUNC WEAK DEFAULT 12
getcpu@@LINUX_2.6
[~]# cd /sys/kernel/debug/tracing/
[tracing]# echo 'p:clock_gettime :vdso:/64:0x8d0' > uprobe_events [tracing]# echo 'p:gettimeofday :vdso:/64:0xc50' >> uprobe_events [tracing]# echo 'p:time :vdso:/64:0xe30' >> uprobe_events [tracing]# echo 1 > events/uprobes/enable
[tracing]# su test # it has UID=1001
[tracing]$ date
Thu Aug 25 17:19:29 MSK 2016
[tracing]$ exit
[tracing]# cat trace
# tracer: nop
#
# entries-in-buffer/entries-written: 175/175 #P:4
#
# _-----=> irqs-off
# / _----=> need-resched

 #                            | / _---=> hardirq/softirq
 #                            || / _--=> preempt-depth
 #                            ||| /     delay
 #           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
 #              | |       |   ||||       |         |
            bash-11560 [001] d...   316.470236: time: (0x7ffcacebae30)
    bash-11560 [001] d...   316.471436: gettimeofday: (0x7ffcacebac50)
            bash-11560 [001] d...   316.477550: time: (0x7ffcacebae30)
            bash-11560 [001] d...   316.477655: time: (0x7ffcacebae30)

mktemp-11568 [001] d... 316.479589: gettimeofday: (0x7ffc603f0c50)
date-11571 [001] d... 316.481890: clock_gettime: (0x7ffec9db58d0)
[...]

If this approach will be decided as fine, I will prepare a better version,
fixing the following things:
o put vdsofs in generic fs/* dir
o support other archs and vdso blobs
o remove BUG_ON()'s and UID==1001 check
o remove extern's and use headers only
o refactor code in create_trace_uprobe()
o add some state to (struct trace_uprobe), so i.e., cat uprobe_events will
print those uprobes as vdso-based
o document this interface in Documentation/trace/uprobetracer.txt
o prepare nice patches set?

So, opinions? Is it worth to add something like this?

[1]: https://lkml.org/lkml/2016/7/12/346

Dmitry Safonov (3):
x86/vdso: create vdso file, use it for mapping
uprobe: drop isdigit() check in create_trace_uprobe
uprobe: add vdso support

Cc Oleg Nesterov <[email protected]>
Cc Al Viro <[email protected]>
Cc Steven Rostedt <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Dmitry Safonov <[email protected]>
Quoted 6 lines
Quoted 4 lines

arch/x86/entry/vdso/vma.c | 148
++++++++++++++++++++++++++++++++++++++++++--
kernel/trace/trace_uprobe.c | 50 +++++++++++----

2 files changed, 180 insertions(+), 18 deletions(-)

I think there is a lot to be said for this idea. However, a private mapping is definitely wrong for the vvar data; for the vdso code it could be considered either way I suppose.

Thanks on your reply.
As you could see, I preserved pure mapping of pfn for vvar:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0 [vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18 [vdso]
(no inode number).
I also think it would be useless to do the same to vvar as it has just data and there is no point in probing it.

Signature · 1 line
             Dmitry
#26 lore raw
Quoted 122 lines · Dmitry Safonov

2016-08-25 23:49 GMT+03:00 H. Peter Anvin <[email protected]>:

Quoted 1 line

On August 25, 2016 8:21:07 AM PDT, Dmitry Safonov

Quoted 15 lines · <[email protected]> wrote:
Quoted 15 lines

This patches set is cleanly RFC and is not supposed to be applied. Also for RFC time it builds only on x86_64.

So, in a mail thread Oleg told that it would be worth to introduce vm_file
for vdso mappings as currently uprobes can not be placed on vDSO VMAs [1].
In this patches set I introduce in-kernel filesystem for vdso files. After patches vDSO VMA now has inode and is just a private file mapping:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0
[vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18
[vdso]

Quoted 1 line
Quoted 1 line

Then I introduce interface in uprobe_events to insert uprobes in

vdso.

Quoted 8 lines
Quoted 8 lines

FWIW:
[~]# cd kernel/linux
[linux]# readelf --syms arch/x86/entry/vdso/vdso64.so
Symbol table '.dynsym' contains 11 entries:

  Num:    Value          Size Type    Bind   Vis      Ndx Name
    0: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT  UND
    1: 0000000000000470     0 SECTION LOCAL  DEFAULT    8

2: 00000000000008d0 885 FUNC WEAK DEFAULT 12

Quoted 8 lines
Quoted 8 lines

clock_gettime@@LINUX_2.6
3: 0000000000000c50 472 FUNC GLOBAL DEFAULT 12
__vdso_gettimeofday@@LINUX_2.6
4: 0000000000000c50 472 FUNC WEAK DEFAULT 12
gettimeofday@@LINUX_2.6
5: 0000000000000e30 21 FUNC GLOBAL DEFAULT 12
__vdso_time@@LINUX_2.6
6: 0000000000000e30 21 FUNC WEAK DEFAULT 12

time@@LINUX_2.6

Quoted 21 lines
Quoted 21 lines

7: 00000000000008d0 885 FUNC GLOBAL DEFAULT 12
__vdso_clock_gettime@@LINUX_2.6

    8: 0000000000000000     0 OBJECT  GLOBAL DEFAULT  ABS LINUX_2.6

9: 0000000000000e50 41 FUNC GLOBAL DEFAULT 12
__vdso_getcpu@@LINUX_2.6
10: 0000000000000e50 41 FUNC WEAK DEFAULT 12
getcpu@@LINUX_2.6
[~]# cd /sys/kernel/debug/tracing/
[tracing]# echo 'p:clock_gettime :vdso:/64:0x8d0' > uprobe_events [tracing]# echo 'p:gettimeofday :vdso:/64:0xc50' >> uprobe_events [tracing]# echo 'p:time :vdso:/64:0xe30' >> uprobe_events [tracing]# echo 1 > events/uprobes/enable
[tracing]# su test # it has UID=1001
[tracing]$ date
Thu Aug 25 17:19:29 MSK 2016
[tracing]$ exit
[tracing]# cat trace
# tracer: nop
#
# entries-in-buffer/entries-written: 175/175 #P:4
#

Quoted 8 lines
Quoted 8 lines

# _-----=> irqs-off
# / _----=> need-resched

 #                            | / _---=> hardirq/softirq
 #                            || / _--=> preempt-depth
 #                            ||| /     delay
 #           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
 #              | |       |   ||||       |         |
            bash-11560 [001] d...   316.470236: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line
    bash-11560 [001] d...   316.471436: gettimeofday:

(0x7ffcacebac50)

Quoted 1 line
Quoted 1 line
            bash-11560 [001] d...   316.477550: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line
            bash-11560 [001] d...   316.477655: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line

mktemp-11568 [001] d... 316.479589: gettimeofday:

(0x7ffc603f0c50)

Quoted 1 line
Quoted 1 line
   date-11571 [001] d...   316.481890: clock_gettime:

(0x7ffec9db58d0)

Quoted 11 lines
Quoted 11 lines

[...]

If this approach will be decided as fine, I will prepare a better version,
fixing the following things:
o put vdsofs in generic fs/* dir
o support other archs and vdso blobs
o remove BUG_ON()'s and UID==1001 check
o remove extern's and use headers only
o refactor code in create_trace_uprobe()
o add some state to (struct trace_uprobe), so i.e., `cat

uprobe_events`

Quoted 24 lines
Quoted 24 lines

will
print those uprobes as vdso-based
o document this interface in Documentation/trace/uprobetracer.txt
o prepare nice patches set?

So, opinions? Is it worth to add something like this?

[1]: https://lkml.org/lkml/2016/7/12/346

Dmitry Safonov (3):
x86/vdso: create vdso file, use it for mapping
uprobe: drop isdigit() check in create_trace_uprobe
uprobe: add vdso support

Cc Oleg Nesterov <[email protected]>
Cc Al Viro <[email protected]>
Cc Steven Rostedt <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Dmitry Safonov <[email protected]>
Quoted 6 lines
Quoted 4 lines

arch/x86/entry/vdso/vma.c | 148
++++++++++++++++++++++++++++++++++++++++++--
kernel/trace/trace_uprobe.c | 50 +++++++++++----

2 files changed, 180 insertions(+), 18 deletions(-)

I think there is a lot to be said for this idea. However, a private

mapping is definitely wrong for the vvar data; for the vdso code it could be considered either way I suppose.

Quoted 9 lines

Thanks on your reply.
As you could see, I preserved pure mapping of pfn for vvar:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0 [vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18 [vdso]
(no inode number).
I also think it would be useless to do the same to vvar as it has just data and there is no point in probing it.

Well, it would things like mremap() just work and so on. Let's get rid of special cases if we are.

Signature · 1 line
Sent from my Android device with K-9 Mail. Please excuse brevity and formatting.
#27 lore raw

2016-08-26 2:00 GMT+03:00 H. Peter Anvin <[email protected]>:

Quoted 133 lines · Dmitry Safonov
Quoted 122 lines

2016-08-25 23:49 GMT+03:00 H. Peter Anvin <[email protected]>:

Quoted 1 line

On August 25, 2016 8:21:07 AM PDT, Dmitry Safonov

Quoted 15 lines · <[email protected]> wrote:
Quoted 15 lines

This patches set is cleanly RFC and is not supposed to be applied. Also for RFC time it builds only on x86_64.

So, in a mail thread Oleg told that it would be worth to introduce vm_file
for vdso mappings as currently uprobes can not be placed on vDSO VMAs [1].
In this patches set I introduce in-kernel filesystem for vdso files. After patches vDSO VMA now has inode and is just a private file mapping:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0
[vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18
[vdso]

Quoted 1 line
Quoted 1 line

Then I introduce interface in uprobe_events to insert uprobes in

vdso.

Quoted 8 lines
Quoted 8 lines

FWIW:
[~]# cd kernel/linux
[linux]# readelf --syms arch/x86/entry/vdso/vdso64.so
Symbol table '.dynsym' contains 11 entries:

  Num:    Value          Size Type    Bind   Vis      Ndx Name
    0: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT  UND
    1: 0000000000000470     0 SECTION LOCAL  DEFAULT    8

2: 00000000000008d0 885 FUNC WEAK DEFAULT 12

Quoted 8 lines
Quoted 8 lines

clock_gettime@@LINUX_2.6
3: 0000000000000c50 472 FUNC GLOBAL DEFAULT 12
__vdso_gettimeofday@@LINUX_2.6
4: 0000000000000c50 472 FUNC WEAK DEFAULT 12
gettimeofday@@LINUX_2.6
5: 0000000000000e30 21 FUNC GLOBAL DEFAULT 12
__vdso_time@@LINUX_2.6
6: 0000000000000e30 21 FUNC WEAK DEFAULT 12

time@@LINUX_2.6

Quoted 21 lines
Quoted 21 lines

7: 00000000000008d0 885 FUNC GLOBAL DEFAULT 12
__vdso_clock_gettime@@LINUX_2.6

    8: 0000000000000000     0 OBJECT  GLOBAL DEFAULT  ABS LINUX_2.6

9: 0000000000000e50 41 FUNC GLOBAL DEFAULT 12
__vdso_getcpu@@LINUX_2.6
10: 0000000000000e50 41 FUNC WEAK DEFAULT 12
getcpu@@LINUX_2.6
[~]# cd /sys/kernel/debug/tracing/
[tracing]# echo 'p:clock_gettime :vdso:/64:0x8d0' > uprobe_events [tracing]# echo 'p:gettimeofday :vdso:/64:0xc50' >> uprobe_events [tracing]# echo 'p:time :vdso:/64:0xe30' >> uprobe_events [tracing]# echo 1 > events/uprobes/enable
[tracing]# su test # it has UID=1001
[tracing]$ date
Thu Aug 25 17:19:29 MSK 2016
[tracing]$ exit
[tracing]# cat trace
# tracer: nop
#
# entries-in-buffer/entries-written: 175/175 #P:4
#

Quoted 8 lines
Quoted 8 lines

# _-----=> irqs-off
# / _----=> need-resched

 #                            | / _---=> hardirq/softirq
 #                            || / _--=> preempt-depth
 #                            ||| /     delay
 #           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
 #              | |       |   ||||       |         |
            bash-11560 [001] d...   316.470236: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line
    bash-11560 [001] d...   316.471436: gettimeofday:

(0x7ffcacebac50)

Quoted 1 line
Quoted 1 line
            bash-11560 [001] d...   316.477550: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line
            bash-11560 [001] d...   316.477655: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line

mktemp-11568 [001] d... 316.479589: gettimeofday:

(0x7ffc603f0c50)

Quoted 1 line
Quoted 1 line
   date-11571 [001] d...   316.481890: clock_gettime:

(0x7ffec9db58d0)

Quoted 11 lines
Quoted 11 lines

[...]

If this approach will be decided as fine, I will prepare a better version,
fixing the following things:
o put vdsofs in generic fs/* dir
o support other archs and vdso blobs
o remove BUG_ON()'s and UID==1001 check
o remove extern's and use headers only
o refactor code in create_trace_uprobe()
o add some state to (struct trace_uprobe), so i.e., `cat

uprobe_events`

Quoted 24 lines
Quoted 24 lines

will
print those uprobes as vdso-based
o document this interface in Documentation/trace/uprobetracer.txt
o prepare nice patches set?

So, opinions? Is it worth to add something like this?

[1]: https://lkml.org/lkml/2016/7/12/346

Dmitry Safonov (3):
x86/vdso: create vdso file, use it for mapping
uprobe: drop isdigit() check in create_trace_uprobe
uprobe: add vdso support

Cc Oleg Nesterov <[email protected]>
Cc Al Viro <[email protected]>
Cc Steven Rostedt <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Dmitry Safonov <[email protected]>
Quoted 6 lines
Quoted 4 lines

arch/x86/entry/vdso/vma.c | 148
++++++++++++++++++++++++++++++++++++++++++--
kernel/trace/trace_uprobe.c | 50 +++++++++++----

2 files changed, 180 insertions(+), 18 deletions(-)

I think there is a lot to be said for this idea. However, a private

mapping is definitely wrong for the vvar data; for the vdso code it could be considered either way I suppose.

Quoted 9 lines

Thanks on your reply.
As you could see, I preserved pure mapping of pfn for vvar:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0 [vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18 [vdso]
(no inode number).
I also think it would be useless to do the same to vvar as it has just data and there is no point in probing it.

Quoted 1 line

Well, it would things like mremap() just work and so on. Let's get rid of special cases if we are.

Well, for RFC it wouldn't move context.vdso pointer on mremap(), but as RFC is for x86_64 only, it will work on it.
Anyway, I don't think it would be hard to fix and make mremap() work on other archs on post-RFC.

The only corner-case I see for now is that /proc/self/map_files/<vdso_range> will point to [vdso] which is broken link. But one could read this file and dump/read vdso blob.
So, in the other words: if some program assumes that /proc/self/map_files/* should always point to correct file, it may be confused. Not sure, maybe it would be confused by orphane-file mappings, so having dangling link there is just fine.

Signature · 1 line
             Dmitry
#28 lore raw
Quoted 141 lines · Dmitry Safonov

2016-08-26 2:00 GMT+03:00 H. Peter Anvin <[email protected]>:

Quoted 133 lines · Dmitry Safonov
Quoted 122 lines

2016-08-25 23:49 GMT+03:00 H. Peter Anvin <[email protected]>:

Quoted 1 line

On August 25, 2016 8:21:07 AM PDT, Dmitry Safonov

Quoted 15 lines · <[email protected]> wrote:
Quoted 15 lines

This patches set is cleanly RFC and is not supposed to be applied. Also for RFC time it builds only on x86_64.

So, in a mail thread Oleg told that it would be worth to introduce vm_file
for vdso mappings as currently uprobes can not be placed on vDSO VMAs [1].
In this patches set I introduce in-kernel filesystem for vdso files. After patches vDSO VMA now has inode and is just a private file mapping:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0
[vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18
[vdso]

Quoted 1 line
Quoted 1 line

Then I introduce interface in uprobe_events to insert uprobes in

vdso.

Quoted 8 lines
Quoted 8 lines

FWIW:
[~]# cd kernel/linux
[linux]# readelf --syms arch/x86/entry/vdso/vdso64.so
Symbol table '.dynsym' contains 11 entries:

  Num:    Value          Size Type    Bind   Vis      Ndx Name
    0: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT  UND
    1: 0000000000000470     0 SECTION LOCAL  DEFAULT    8

2: 00000000000008d0 885 FUNC WEAK DEFAULT 12

Quoted 8 lines
Quoted 8 lines

clock_gettime@@LINUX_2.6
3: 0000000000000c50 472 FUNC GLOBAL DEFAULT 12
__vdso_gettimeofday@@LINUX_2.6
4: 0000000000000c50 472 FUNC WEAK DEFAULT 12
gettimeofday@@LINUX_2.6
5: 0000000000000e30 21 FUNC GLOBAL DEFAULT 12
__vdso_time@@LINUX_2.6
6: 0000000000000e30 21 FUNC WEAK DEFAULT 12

time@@LINUX_2.6

Quoted 21 lines
Quoted 21 lines

7: 00000000000008d0 885 FUNC GLOBAL DEFAULT 12
__vdso_clock_gettime@@LINUX_2.6

    8: 0000000000000000     0 OBJECT  GLOBAL DEFAULT  ABS LINUX_2.6

9: 0000000000000e50 41 FUNC GLOBAL DEFAULT 12
__vdso_getcpu@@LINUX_2.6
10: 0000000000000e50 41 FUNC WEAK DEFAULT 12
getcpu@@LINUX_2.6
[~]# cd /sys/kernel/debug/tracing/
[tracing]# echo 'p:clock_gettime :vdso:/64:0x8d0' > uprobe_events [tracing]# echo 'p:gettimeofday :vdso:/64:0xc50' >> uprobe_events [tracing]# echo 'p:time :vdso:/64:0xe30' >> uprobe_events [tracing]# echo 1 > events/uprobes/enable
[tracing]# su test # it has UID=1001
[tracing]$ date
Thu Aug 25 17:19:29 MSK 2016
[tracing]$ exit
[tracing]# cat trace
# tracer: nop
#
# entries-in-buffer/entries-written: 175/175 #P:4
#

Quoted 8 lines
Quoted 8 lines

# _-----=> irqs-off
# / _----=> need-resched

 #                            | / _---=> hardirq/softirq
 #                            || / _--=> preempt-depth
 #                            ||| /     delay
 #           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
 #              | |       |   ||||       |         |
            bash-11560 [001] d...   316.470236: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line
    bash-11560 [001] d...   316.471436: gettimeofday:

(0x7ffcacebac50)

Quoted 1 line
Quoted 1 line
            bash-11560 [001] d...   316.477550: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line
            bash-11560 [001] d...   316.477655: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line

mktemp-11568 [001] d... 316.479589: gettimeofday:

(0x7ffc603f0c50)

Quoted 1 line
Quoted 1 line
   date-11571 [001] d...   316.481890: clock_gettime:

(0x7ffec9db58d0)

Quoted 11 lines
Quoted 11 lines

[...]

If this approach will be decided as fine, I will prepare a better version,
fixing the following things:
o put vdsofs in generic fs/* dir
o support other archs and vdso blobs
o remove BUG_ON()'s and UID==1001 check
o remove extern's and use headers only
o refactor code in create_trace_uprobe()
o add some state to (struct trace_uprobe), so i.e., `cat

uprobe_events`

Quoted 24 lines
Quoted 24 lines

will
print those uprobes as vdso-based
o document this interface in Documentation/trace/uprobetracer.txt
o prepare nice patches set?

So, opinions? Is it worth to add something like this?

[1]: https://lkml.org/lkml/2016/7/12/346

Dmitry Safonov (3):
x86/vdso: create vdso file, use it for mapping
uprobe: drop isdigit() check in create_trace_uprobe
uprobe: add vdso support

Cc Oleg Nesterov <[email protected]>
Cc Al Viro <[email protected]>
Cc Steven Rostedt <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Dmitry Safonov <[email protected]>
Quoted 6 lines
Quoted 4 lines

arch/x86/entry/vdso/vma.c | 148
++++++++++++++++++++++++++++++++++++++++++--
kernel/trace/trace_uprobe.c | 50 +++++++++++----

2 files changed, 180 insertions(+), 18 deletions(-)

I think there is a lot to be said for this idea. However, a private

mapping is definitely wrong for the vvar data; for the vdso code it could be considered either way I suppose.

Quoted 9 lines

Thanks on your reply.
As you could see, I preserved pure mapping of pfn for vvar:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0 [vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18 [vdso]
(no inode number).
I also think it would be useless to do the same to vvar as it has just data and there is no point in probing it.

Quoted 1 line

Well, it would things like mremap() just work and so on. Let's get rid of special cases if we are.

Well, for RFC it wouldn't move context.vdso pointer on mremap(), but as RFC is for x86_64 only, it will work on it.
Anyway, I don't think it would be hard to fix and make mremap() work on other archs on post-RFC.

Quoted 7 lines

The only corner-case I see for now is that /proc/self/map_files/<vdso_range> will point to [vdso] which is broken link. But one could read this file and dump/read vdso blob.
So, in the other words: if some program assumes that /proc/self/map_files/* should always point to correct file, it may be confused. Not sure, maybe it would be confused by orphane-file mappings, so having dangling link there is just fine.

I don't see anything a priori wrong with having map_files point somewhere, but it could be worth special casing it for special mappings to preserve existing behavior (no file at all).

--Andy

#29 lore raw

2016-08-26 17:32 GMT+03:00 Andy Lutomirski <[email protected]>:

Quoted 150 lines · Dmitry Safonov
Quoted 141 lines

2016-08-26 2:00 GMT+03:00 H. Peter Anvin <[email protected]>:

Quoted 133 lines · Dmitry Safonov
Quoted 122 lines

2016-08-25 23:49 GMT+03:00 H. Peter Anvin <[email protected]>:

Quoted 1 line

On August 25, 2016 8:21:07 AM PDT, Dmitry Safonov

Quoted 15 lines · <[email protected]> wrote:
Quoted 15 lines

This patches set is cleanly RFC and is not supposed to be applied. Also for RFC time it builds only on x86_64.

So, in a mail thread Oleg told that it would be worth to introduce vm_file
for vdso mappings as currently uprobes can not be placed on vDSO VMAs [1].
In this patches set I introduce in-kernel filesystem for vdso files. After patches vDSO VMA now has inode and is just a private file mapping:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0
[vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18
[vdso]

Quoted 1 line
Quoted 1 line

Then I introduce interface in uprobe_events to insert uprobes in

vdso.

Quoted 8 lines
Quoted 8 lines

FWIW:
[~]# cd kernel/linux
[linux]# readelf --syms arch/x86/entry/vdso/vdso64.so
Symbol table '.dynsym' contains 11 entries:

  Num:    Value          Size Type    Bind   Vis      Ndx Name
    0: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT  UND
    1: 0000000000000470     0 SECTION LOCAL  DEFAULT    8

2: 00000000000008d0 885 FUNC WEAK DEFAULT 12

Quoted 8 lines
Quoted 8 lines

clock_gettime@@LINUX_2.6
3: 0000000000000c50 472 FUNC GLOBAL DEFAULT 12
__vdso_gettimeofday@@LINUX_2.6
4: 0000000000000c50 472 FUNC WEAK DEFAULT 12
gettimeofday@@LINUX_2.6
5: 0000000000000e30 21 FUNC GLOBAL DEFAULT 12
__vdso_time@@LINUX_2.6
6: 0000000000000e30 21 FUNC WEAK DEFAULT 12

time@@LINUX_2.6

Quoted 21 lines
Quoted 21 lines

7: 00000000000008d0 885 FUNC GLOBAL DEFAULT 12
__vdso_clock_gettime@@LINUX_2.6

    8: 0000000000000000     0 OBJECT  GLOBAL DEFAULT  ABS LINUX_2.6

9: 0000000000000e50 41 FUNC GLOBAL DEFAULT 12
__vdso_getcpu@@LINUX_2.6
10: 0000000000000e50 41 FUNC WEAK DEFAULT 12
getcpu@@LINUX_2.6
[~]# cd /sys/kernel/debug/tracing/
[tracing]# echo 'p:clock_gettime :vdso:/64:0x8d0' > uprobe_events [tracing]# echo 'p:gettimeofday :vdso:/64:0xc50' >> uprobe_events [tracing]# echo 'p:time :vdso:/64:0xe30' >> uprobe_events [tracing]# echo 1 > events/uprobes/enable
[tracing]# su test # it has UID=1001
[tracing]$ date
Thu Aug 25 17:19:29 MSK 2016
[tracing]$ exit
[tracing]# cat trace
# tracer: nop
#
# entries-in-buffer/entries-written: 175/175 #P:4
#

Quoted 8 lines
Quoted 8 lines

# _-----=> irqs-off
# / _----=> need-resched

 #                            | / _---=> hardirq/softirq
 #                            || / _--=> preempt-depth
 #                            ||| /     delay
 #           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
 #              | |       |   ||||       |         |
            bash-11560 [001] d...   316.470236: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line
    bash-11560 [001] d...   316.471436: gettimeofday:

(0x7ffcacebac50)

Quoted 1 line
Quoted 1 line
            bash-11560 [001] d...   316.477550: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line
            bash-11560 [001] d...   316.477655: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line

mktemp-11568 [001] d... 316.479589: gettimeofday:

(0x7ffc603f0c50)

Quoted 1 line
Quoted 1 line
   date-11571 [001] d...   316.481890: clock_gettime:

(0x7ffec9db58d0)

Quoted 11 lines
Quoted 11 lines

[...]

If this approach will be decided as fine, I will prepare a better version,
fixing the following things:
o put vdsofs in generic fs/* dir
o support other archs and vdso blobs
o remove BUG_ON()'s and UID==1001 check
o remove extern's and use headers only
o refactor code in create_trace_uprobe()
o add some state to (struct trace_uprobe), so i.e., `cat

uprobe_events`

Quoted 24 lines
Quoted 24 lines

will
print those uprobes as vdso-based
o document this interface in Documentation/trace/uprobetracer.txt
o prepare nice patches set?

So, opinions? Is it worth to add something like this?

[1]: https://lkml.org/lkml/2016/7/12/346

Dmitry Safonov (3):
x86/vdso: create vdso file, use it for mapping
uprobe: drop isdigit() check in create_trace_uprobe
uprobe: add vdso support

Cc Oleg Nesterov <[email protected]>
Cc Al Viro <[email protected]>
Cc Steven Rostedt <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Dmitry Safonov <[email protected]>
Quoted 6 lines
Quoted 4 lines

arch/x86/entry/vdso/vma.c | 148
++++++++++++++++++++++++++++++++++++++++++--
kernel/trace/trace_uprobe.c | 50 +++++++++++----

2 files changed, 180 insertions(+), 18 deletions(-)

I think there is a lot to be said for this idea. However, a private

mapping is definitely wrong for the vvar data; for the vdso code it could be considered either way I suppose.

Quoted 9 lines

Thanks on your reply.
As you could see, I preserved pure mapping of pfn for vvar:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0 [vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18 [vdso]
(no inode number).
I also think it would be useless to do the same to vvar as it has just data and there is no point in probing it.

Quoted 1 line

Well, it would things like mremap() just work and so on. Let's get rid of special cases if we are.

Well, for RFC it wouldn't move context.vdso pointer on mremap(), but as RFC is for x86_64 only, it will work on it.
Anyway, I don't think it would be hard to fix and make mremap() work on other archs on post-RFC.

Quoted 7 lines

The only corner-case I see for now is that /proc/self/map_files/<vdso_range> will point to [vdso] which is broken link. But one could read this file and dump/read vdso blob.
So, in the other words: if some program assumes that /proc/self/map_files/* should always point to correct file, it may be confused. Not sure, maybe it would be confused by orphane-file mappings, so having dangling link there is just fine.

Quoted 3 lines

I don't see anything a priori wrong with having map_files point somewhere, but it could be worth special casing it for special mappings to preserve existing behavior (no file at all).

Yep, that could be easily done, will do.
Anyway, just curious - what may it break?

Thanks on the reply, Andy. Does the patches set look sane for you?

Signature · 1 line
             Dmitry
#30 lore raw

2016-08-26 17:42 GMT+03:00 Dmitry Safonov <[email protected]>:

Quoted 151 lines

2016-08-26 17:32 GMT+03:00 Andy Lutomirski <[email protected]>:

Quoted 150 lines · Dmitry Safonov
Quoted 148 lines

2016-08-26 2:00 GMT+03:00 H. Peter Anvin <[email protected]>:

Quoted 133 lines · Dmitry Safonov
Quoted 122 lines

2016-08-25 23:49 GMT+03:00 H. Peter Anvin <[email protected]>:

Quoted 1 line

On August 25, 2016 8:21:07 AM PDT, Dmitry Safonov

Quoted 15 lines · <[email protected]> wrote:
Quoted 15 lines

This patches set is cleanly RFC and is not supposed to be applied. Also for RFC time it builds only on x86_64.

So, in a mail thread Oleg told that it would be worth to introduce vm_file
for vdso mappings as currently uprobes can not be placed on vDSO VMAs [1].
In this patches set I introduce in-kernel filesystem for vdso files. After patches vDSO VMA now has inode and is just a private file mapping:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0
[vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18
[vdso]

Quoted 1 line
Quoted 1 line

Then I introduce interface in uprobe_events to insert uprobes in

vdso.

Quoted 8 lines
Quoted 8 lines

FWIW:
[~]# cd kernel/linux
[linux]# readelf --syms arch/x86/entry/vdso/vdso64.so
Symbol table '.dynsym' contains 11 entries:

  Num:    Value          Size Type    Bind   Vis      Ndx Name
    0: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT  UND
    1: 0000000000000470     0 SECTION LOCAL  DEFAULT    8

2: 00000000000008d0 885 FUNC WEAK DEFAULT 12

Quoted 8 lines
Quoted 8 lines

clock_gettime@@LINUX_2.6
3: 0000000000000c50 472 FUNC GLOBAL DEFAULT 12
__vdso_gettimeofday@@LINUX_2.6
4: 0000000000000c50 472 FUNC WEAK DEFAULT 12
gettimeofday@@LINUX_2.6
5: 0000000000000e30 21 FUNC GLOBAL DEFAULT 12
__vdso_time@@LINUX_2.6
6: 0000000000000e30 21 FUNC WEAK DEFAULT 12

time@@LINUX_2.6

Quoted 21 lines
Quoted 21 lines

7: 00000000000008d0 885 FUNC GLOBAL DEFAULT 12
__vdso_clock_gettime@@LINUX_2.6

    8: 0000000000000000     0 OBJECT  GLOBAL DEFAULT  ABS LINUX_2.6

9: 0000000000000e50 41 FUNC GLOBAL DEFAULT 12
__vdso_getcpu@@LINUX_2.6
10: 0000000000000e50 41 FUNC WEAK DEFAULT 12
getcpu@@LINUX_2.6
[~]# cd /sys/kernel/debug/tracing/
[tracing]# echo 'p:clock_gettime :vdso:/64:0x8d0' > uprobe_events [tracing]# echo 'p:gettimeofday :vdso:/64:0xc50' >> uprobe_events [tracing]# echo 'p:time :vdso:/64:0xe30' >> uprobe_events [tracing]# echo 1 > events/uprobes/enable
[tracing]# su test # it has UID=1001
[tracing]$ date
Thu Aug 25 17:19:29 MSK 2016
[tracing]$ exit
[tracing]# cat trace
# tracer: nop
#
# entries-in-buffer/entries-written: 175/175 #P:4
#

Quoted 8 lines
Quoted 8 lines

# _-----=> irqs-off
# / _----=> need-resched

 #                            | / _---=> hardirq/softirq
 #                            || / _--=> preempt-depth
 #                            ||| /     delay
 #           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
 #              | |       |   ||||       |         |
            bash-11560 [001] d...   316.470236: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line
    bash-11560 [001] d...   316.471436: gettimeofday:

(0x7ffcacebac50)

Quoted 1 line
Quoted 1 line
            bash-11560 [001] d...   316.477550: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line
            bash-11560 [001] d...   316.477655: time:

(0x7ffcacebae30)

Quoted 1 line
Quoted 1 line

mktemp-11568 [001] d... 316.479589: gettimeofday:

(0x7ffc603f0c50)

Quoted 1 line
Quoted 1 line
   date-11571 [001] d...   316.481890: clock_gettime:

(0x7ffec9db58d0)

Quoted 11 lines
Quoted 11 lines

[...]

If this approach will be decided as fine, I will prepare a better version,
fixing the following things:
o put vdsofs in generic fs/* dir
o support other archs and vdso blobs
o remove BUG_ON()'s and UID==1001 check
o remove extern's and use headers only
o refactor code in create_trace_uprobe()
o add some state to (struct trace_uprobe), so i.e., `cat

uprobe_events`

Quoted 24 lines
Quoted 24 lines

will
print those uprobes as vdso-based
o document this interface in Documentation/trace/uprobetracer.txt
o prepare nice patches set?

So, opinions? Is it worth to add something like this?

[1]: https://lkml.org/lkml/2016/7/12/346

Dmitry Safonov (3):
x86/vdso: create vdso file, use it for mapping
uprobe: drop isdigit() check in create_trace_uprobe
uprobe: add vdso support

Cc Oleg Nesterov <[email protected]>
Cc Al Viro <[email protected]>
Cc Steven Rostedt <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Dmitry Safonov <[email protected]>
Quoted 6 lines
Quoted 4 lines

arch/x86/entry/vdso/vma.c | 148
++++++++++++++++++++++++++++++++++++++++++--
kernel/trace/trace_uprobe.c | 50 +++++++++++----

2 files changed, 180 insertions(+), 18 deletions(-)

I think there is a lot to be said for this idea. However, a private

mapping is definitely wrong for the vvar data; for the vdso code it could be considered either way I suppose.

Quoted 9 lines

Thanks on your reply.
As you could see, I preserved pure mapping of pfn for vvar:
7ffcc4b2b000-7ffcc4b2d000 r--p 00000000 00:00 0 [vvar]
7ffcc4b2d000-7ffcc4b2f000 r-xp 00000000 00:09 18 [vdso]
(no inode number).
I also think it would be useless to do the same to vvar as it has just data and there is no point in probing it.

Quoted 1 line

Well, it would things like mremap() just work and so on. Let's get rid of special cases if we are.

Well, for RFC it wouldn't move context.vdso pointer on mremap(), but as RFC is for x86_64 only, it will work on it.
Anyway, I don't think it would be hard to fix and make mremap() work on other archs on post-RFC.

The only corner-case I see for now is that /proc/self/map_files/<vdso_range> will point to [vdso] which is broken link. But one could read this file and dump/read vdso blob.
So, in the other words: if some program assumes that /proc/self/map_files/* should always point to correct file, it may be confused. Not sure, maybe it would be confused by orphane-file mappings, so having dangling link there is just fine.

Quoted 8 lines
Quoted 3 lines

I don't see anything a priori wrong with having map_files point somewhere, but it could be worth special casing it for special mappings to preserve existing behavior (no file at all).

Yep, that could be easily done, will do.
Anyway, just curious - what may it break?

Thanks on the reply, Andy. Does the patches set look sane for you?

I mean, the idea, not current big TODO list and nitpicks :)

Signature · 1 line
             Dmitry
#31 lore raw
In reply to Dmitry Safonov

The more I'm thinking about this, why don't we simply have these (the various possible vdsos as well as vvar) as actual files in sysfs instead of introducing a new filesystem? I don't believe sysfs actually has to be mounted in order for sysfs files to have an inode.

It could also be in procfs, I guess, but sysfs probably makes more sense.

I'm thinking something like:

/sys/kernel/vdso/{i386,x86_64,x32,vvar}

Not only would this let the container people and so on do weird things much easier, but it ought to eliminate a whole slew of special cases.

-hpa

#32 lore raw
Quoted 5 lines · H. Peter Anvin

The more I'm thinking about this, why don't we simply have these (the various possible vdsos as well as vvar) as actual files in sysfs instead of introducing a new filesystem? I don't believe sysfs actually has to be mounted in order for sysfs files to have an inode.

Quoted 9 lines

It could also be in procfs, I guess, but sysfs probably makes more sense.

I'm thinking something like:

/sys/kernel/vdso/{i386,x86_64,x32,vvar}

Not only would this let the container people and so on do weird things much easier, but it ought to eliminate a whole slew of special cases.

Even crazier idea: instead of a separate vvar file, have the vvar page just be a part of these files (as a shared page)... I'm wondering if we can even use load_elf_interp() since after all it is an ELF shared library image...

-hpa

#33 lore raw
Quoted 12 lines · H. Peter Anvin
Quoted 11 lines · H. Peter Anvin

The more I'm thinking about this, why don't we simply have these (the various possible vdsos as well as vvar) as actual files in sysfs instead of introducing a new filesystem? I don't believe sysfs actually has to be mounted in order for sysfs files to have an inode.

It could also be in procfs, I guess, but sysfs probably makes more sense.

I'm thinking something like:

/sys/kernel/vdso/{i386,x86_64,x32,vvar}

Quoted 8 lines
Quoted 3 lines

Not only would this let the container people and so on do weird things much easier, but it ought to eliminate a whole slew of special cases.

Even crazier idea: instead of a separate vvar file, have the vvar page just be a part of these files (as a shared page)... I'm wondering if we can even use load_elf_interp() since after all it is an ELF shared library image...

I think that may be too crazy:

- If vvar is in the same inode, then that inode won't be a valid ELF
image, because the ELF header won't be in the right place.

- vvar is highly magical. IMO letting it get mapped with VM_MAYWRITE
is asking for trouble, as anything that writes it will COW it, leading to strange malfunctions.

- vvar can, and has, had IO pages in it. This means that the actual
cache types can vary page-to-page in the vvar area, which is not something that ordinary files do.

Also, if we let the users get an fd pointing to the vdso, then we're more or less committing to never having contents in the vdso text that vary per-process. Are we okay with that.

Dmitry's patches have the vdso using the page cache, and I'm not sure that even that is needed. I think that a file with no backing address_space that simply provides vm_fault instead may be sufficient, especially for vvar. I don't know if uprobes would be okay with that, though.

My personal preference is to let them both be real struct file * objects (possibly shared between all processes of the same vdso ABI) but to prevent user code from ever creating an fd referring to one of these files.

--Andy

#34 lore raw
Quoted 2 lines · Andy Lutomirski

- If vvar is in the same inode, then that inode won't be a valid ELF
image, because the ELF header won't be in the right place.

So the vvar ought to move into an actual ELF segment, which is probably The Right Thing anyway.

Quoted 7 lines

- vvar is highly magical. IMO letting it get mapped with VM_MAYWRITE
is asking for trouble, as anything that writes it will COW it, leading to strange malfunctions.

- vvar can, and has, had IO pages in it. This means that the actual
cache types can vary page-to-page in the vvar area, which is not something that ordinary files do.

Neither of these are any different than many devices, or various files in procfs.

Quoted 4 lines

My personal preference is to let them both be real struct file * objects (possibly shared between all processes of the same vdso ABI) but to prevent user code from ever creating an fd referring to one of these files.

Why? It would help people doing weird things like process snapshotting or bimodal execution enormously. We want to share an inode, obviously; the pointer is another issue.

Quoted 3 lines

Also, if we let the users get an fd pointing to the vdso, then we're more or less committing to never having contents in the vdso text that vary per-process. Are we okay with that.

This might be a reason to put these objects in procfs rather than sysfs, but I have to admit that this seems *extremely* far fetched to me. Obviously they vary per process in the sense that there are already several to choose from. In the case of process-unique vdsos there would be a large number of them, of course.

-hpa

#35 lore raw
Quoted 5 lines · H. Peter Anvin
Quoted 5 lines

- vvar is highly magical. IMO letting it get mapped with VM_MAYWRITE
is asking for trouble, as anything that writes it will COW it, leading to strange malfunctions.

The vvar page obviously needs to be mapped MAP_SHARED, and the underlying file needs to reject writes. A solution where this area doesn't end up MAP_SHARED is obviously defective.

As far as keeping the user from doing really stupid things... they can map a RAM page over the vvar area and there is nothing the kernel really can do to keep them from doing something like that without doing things that are probably way worse than the disease. At least it will be obvious looking at the mapping file what is going on.

-hpa

#36 lore raw
Quoted 10 lines · H. Peter Anvin
Quoted 5 lines
Quoted 5 lines

- vvar is highly magical. IMO letting it get mapped with VM_MAYWRITE
is asking for trouble, as anything that writes it will COW it, leading to strange malfunctions.

The vvar page obviously needs to be mapped MAP_SHARED, and the underlying file needs to reject writes. A solution where this area doesn't end up MAP_SHARED is obviously defective.

Hmm, maybe. But it does certainly work now, and I'm not sure what we gain by making it more file-like than needed. Using a non-null vm_file removes tons of special cases, but giving it a real address_space doesn't seem very useful to me.

Quoted 6 lines

As far as keeping the user from doing really stupid things... they can map a RAM page over the vvar area and there is nothing the kernel really can do to keep them from doing something like that without doing things that are probably way worse than the disease. At least it will be obvious looking at the mapping file what is going on.

I don't want an overly clever debugger to poke the page. MAP_SHARED might be sufficient.