Skip to contentSkip to conversation

floppy: fix usercopy direction

PATCH In review Jann Horn 7 participants 10 messages Mar 26, 2019 – Aug 13, 2019

Subsystems (3)
lore
Share & tools
lore

Jann Horn

author
#0 lore raw

PATCH 1/1 floppy: fix usercopy direction

As sparse points out, these two copy_from_user() should actually be copy_to_user().

Fixes 229b53c9bf4e ("take floppy compat ioctls to sodding floppy.c")

---
compile-tested only

 drivers/block/floppy.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/block/floppy.c +2 -2
index 95f608d1a098..8c641245ff12 100644
@@ -3749,7 +3749,7 @@ static int compat_getdrvprm(int drive,
 	v.native_format = UDP->native_format;
 	mutex_unlock(&floppy_mutex);
 
-	if (copy_from_user(arg, &v, sizeof(struct compat_floppy_drive_params)))
+	if (copy_to_user(arg, &v, sizeof(struct compat_floppy_drive_params)))
 		return -EFAULT;
 	return 0;
 }
@@ -3785,7 +3785,7 @@ static int compat_getdrvstat(int drive, bool poll,
 	v.bufblocks = UDRS->bufblocks;
 	mutex_unlock(&floppy_mutex);
 
-	if (copy_from_user(arg, &v, sizeof(struct compat_floppy_drive_struct)))
+	if (copy_to_user(arg, &v, sizeof(struct compat_floppy_drive_struct)))
 		return -EFAULT;
 	return 0;
 Eintr:
Signature · 1 line
2.21.0.392.gf8f6787159e-goog
Reviewed-by on v1 1/1
#1 lore raw
Quoted 6 lines · Jann Horn

As sparse points out, these two copy_from_user() should actually be copy_to_user().

Fixes 229b53c9bf4e ("take floppy compat ioctls to sodding floppy.c")
Quoted 28 lines

---
compile-tested only

  drivers/block/floppy.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)
drivers/block/floppy.c +2 -2
index 95f608d1a098..8c641245ff12 100644
@@ -3749,7 +3749,7 @@ static int compat_getdrvprm(int drive,
  	v.native_format = UDP->native_format;
  	mutex_unlock(&floppy_mutex);
 
-	if (copy_from_user(arg, &v, sizeof(struct compat_floppy_drive_params)))
+	if (copy_to_user(arg, &v, sizeof(struct compat_floppy_drive_params)))
  		return -EFAULT;
  	return 0;
  }
@@ -3785,7 +3785,7 @@ static int compat_getdrvstat(int drive, bool poll,
  	v.bufblocks = UDRS->bufblocks;
  	mutex_unlock(&floppy_mutex);
 
-	if (copy_from_user(arg, &v, sizeof(struct compat_floppy_drive_struct)))
+	if (copy_to_user(arg, &v, sizeof(struct compat_floppy_drive_struct)))
  		return -EFAULT;
  	return 0;
  Eintr:
Acked-by on v1 1/1
#2 lore raw
In reply to Jann Horn

Hello everyone!

Quoted 2 lines · Jann Horn

As sparse points out, these two copy_from_user() should actually be copy_to_user().

I've spent some time on these bugs, but it turned out that they are already public.

I think Jann's patch is lost, it is not applied to the mainline. So I add a new floppy maintainer Denis Efremov to CC.

These bugs on x86_64 cause memset for the userspace memory from the kernelspace. That is funny:
- access_ok for the copy_from_user source (2nd argument) returns zero;
- copy_from_user tries to erase the destination (1st argument);
- but the destination is in the userspace instead of kernelspace.

So we have:

[   40.937098] BUG: unable to handle page fault for address: 0000000041414242
[   40.938714] #PF: supervisor write access in kernel mode
[   40.939951] #PF: error_code(0x0002) - not-present page
[   40.941121] PGD 7963f067 P4D 7963f067 PUD 0
[   40.942107] Oops: 0002 [#1] SMP NOPTI
[   40.942968] CPU: 0 PID: 292 Comm: d Not tainted 5.3.0-rc3+ #7
[   40.944288] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
Ubuntu-1.8.2-1ubuntu1 04/01/2014
[   40.946478] RIP: 0010:__memset+0x24/0x30
[   40.947394] Code: 90 90 90 90 90 90 0f 1f 44 00 00 49 89 f9 48 89 d1 83 e2 07
48 c1 e9 03 40 0f b6 f6 48 b8 01 01 01 01 01 01 01 01 48 0f af c6 <f3> 48 ab 89
d1 f3 aa 4c 89 c8 c3 90 49 89 f9 40 88 f0 48 89 d1 f3
[   40.951721] RSP: 0018:ffffc900003dbd58 EFLAGS: 00010206
[   40.952941] RAX: 0000000000000000 RBX: 0000000000000034 RCX: 0000000000000006
[   40.954592] RDX: 0000000000000004 RSI: 0000000000000000 RDI: 0000000041414242
[   40.956169] RBP: 0000000041414242 R08: ffffffff8200bd80 R09: 0000000041414242
[   40.957753] R10: 0000000000121806 R11: ffff88807da28ab0 R12: ffffc900003dbd7c
[   40.959407] R13: 0000000000000001 R14: 0000000041414242 R15: 0000000041414242
[   40.961062] FS:  00007f91115c4440(0000) GS:ffff88807da00000(0000)
knlGS:0000000000000000
[   40.962603] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   40.963695] CR2: 0000000041414242 CR3: 000000007c584000 CR4: 00000000000006f0
[   40.965004] Call Trace:
[   40.965459]  _copy_from_user+0x51/0x60
[   40.966141]  compat_getdrvstat+0x124/0x170
[   40.966781]  fd_compat_ioctl+0x69c/0x6d0
[   40.967423]  ? selinux_file_ioctl+0x16f/0x210
[   40.968117]  compat_blkdev_ioctl+0x21d/0x8f0
[   40.968864]  __x32_compat_sys_ioctl+0x99/0x250
[   40.969659]  do_syscall_64+0x4a/0x110
[   40.970337]  entry_SYSCALL_64_after_hwframe+0x44/0xa9

I haven't found a way to exploit it.

Quoted 6 lines
Fixes 229b53c9bf4e ("take floppy compat ioctls to sodding floppy.c")

---
compile-tested only

Quoted 18 lines
 drivers/block/floppy.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/block/floppy.c +1 -1
index 95f608d1a098..8c641245ff12 100644
@@ -3749,7 +3749,7 @@ static int compat_getdrvprm(int drive,
 	v.native_format = UDP->native_format;
 	mutex_unlock(&floppy_mutex);
 
-	if (copy_from_user(arg, &v, sizeof(struct compat_floppy_drive_params)))
+	if (copy_to_user(arg, &v, sizeof(struct compat_floppy_drive_params)))
 		return -EFAULT;
 	return 0;
 }

drivers/block/floppy.c static int compat_getdrvstat(int drive, bool poll, · line 3791

Quoted 8 lines
drivers/block/floppy.c +1 -1
@@ -3785,7 +3785,7 @@ static int compat_getdrvstat(int drive, bool poll,
 	v.bufblocks = UDRS->bufblocks;
 	mutex_unlock(&floppy_mutex);
 
-	if (copy_from_user(arg, &v, sizeof(struct compat_floppy_drive_struct)))
+	if (copy_to_user(arg, &v, sizeof(struct compat_floppy_drive_struct)))
 		return -EFAULT;
 	return 0;
 Eintr:

I also wrote a coccinelle rule for detecting similar bugs (adding coccinelle experts to CC).

virtual report

@cfu@
identifier f;
type t;
identifier v;
position decl_p;
position copy_p;
@@

f(..., t v@decl_p, ...)
{
<+...
copy_from_user@copy_p(v, ...)
...+>
}

@script:python@
f << cfu.f;
t << cfu.t;
v << cfu.v;
decl_p << cfu.decl_p;
copy_p << cfu.copy_p;
@@

if '__user' in t:
msg0 = "function \"" + f + "\" has arg \"" + v + "\" of type \"" + t + "\"" coccilib.report.print_report(decl_p[0], msg0)
msg1 = "copy_from_user uses \"" + v + "\" as the destination. What a shame!\n" coccilib.report.print_report(copy_p[0], msg1)

The rule output:

./drivers/block/floppy.c:3756:49-52: function "compat_getdrvprm" has arg "arg" of type "struct compat_floppy_drive_params __user *" ./drivers/block/floppy.c:3783:5-19: copy_from_user uses "arg" as the destination. What a shame!

./drivers/block/floppy.c:3789:49-52: function "compat_getdrvstat" has arg "arg" of type "struct compat_floppy_drive_struct __user *" ./drivers/block/floppy.c:3819:5-19: copy_from_user uses "arg" as the destination. What a shame!

Best regards,
Alexander

maintainer
#3 lore raw
Quoted 10 lines · Alexander Popov

Hello everyone!

Quoted 2 lines · Jann Horn

As sparse points out, these two copy_from_user() should actually be copy_to_user().

I've spent some time on these bugs, but it turned out that they are already public.

I think Jann's patch is lost, it is not applied to the mainline. So I add a new floppy maintainer Denis Efremov to CC.

Looks like it got lost indeed, I will just apply this one directly for 5.4.

Signature · 1 line
Jens Axboe
maintainer
#4 lore raw
Quoted 5 lines
Quoted 1 line

So I add a new floppy maintainer Denis Efremov to CC.

Looks like it got lost indeed, I will just apply this one directly for 5.4.

Thank you!

Denis

#5 lore raw
In reply to Alexander Popov
Quoted 93 lines · Alexander Popov

Hello everyone!

Quoted 2 lines · Jann Horn

As sparse points out, these two copy_from_user() should actually be copy_to_user().

I've spent some time on these bugs, but it turned out that they are already public.

I think Jann's patch is lost, it is not applied to the mainline. So I add a new floppy maintainer Denis Efremov to CC.

These bugs on x86_64 cause memset for the userspace memory from the kernelspace. That is funny:
- access_ok for the copy_from_user source (2nd argument) returns zero;
- copy_from_user tries to erase the destination (1st argument);
- but the destination is in the userspace instead of kernelspace.

So we have:

[   40.937098] BUG: unable to handle page fault for address: 0000000041414242
[   40.938714] #PF: supervisor write access in kernel mode
[   40.939951] #PF: error_code(0x0002) - not-present page
[   40.941121] PGD 7963f067 P4D 7963f067 PUD 0
[   40.942107] Oops: 0002 [#1] SMP NOPTI
[   40.942968] CPU: 0 PID: 292 Comm: d Not tainted 5.3.0-rc3+ #7
[   40.944288] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
Ubuntu-1.8.2-1ubuntu1 04/01/2014
[   40.946478] RIP: 0010:__memset+0x24/0x30
[   40.947394] Code: 90 90 90 90 90 90 0f 1f 44 00 00 49 89 f9 48 89 d1 83 e2 07
48 c1 e9 03 40 0f b6 f6 48 b8 01 01 01 01 01 01 01 01 48 0f af c6 <f3> 48 ab 89
d1 f3 aa 4c 89 c8 c3 90 49 89 f9 40 88 f0 48 89 d1 f3
[   40.951721] RSP: 0018:ffffc900003dbd58 EFLAGS: 00010206
[   40.952941] RAX: 0000000000000000 RBX: 0000000000000034 RCX: 0000000000000006
[   40.954592] RDX: 0000000000000004 RSI: 0000000000000000 RDI: 0000000041414242
[   40.956169] RBP: 0000000041414242 R08: ffffffff8200bd80 R09: 0000000041414242
[   40.957753] R10: 0000000000121806 R11: ffff88807da28ab0 R12: ffffc900003dbd7c
[   40.959407] R13: 0000000000000001 R14: 0000000041414242 R15: 0000000041414242
[   40.961062] FS:  00007f91115c4440(0000) GS:ffff88807da00000(0000)
knlGS:0000000000000000
[   40.962603] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   40.963695] CR2: 0000000041414242 CR3: 000000007c584000 CR4: 00000000000006f0
[   40.965004] Call Trace:
[   40.965459]  _copy_from_user+0x51/0x60
[   40.966141]  compat_getdrvstat+0x124/0x170
[   40.966781]  fd_compat_ioctl+0x69c/0x6d0
[   40.967423]  ? selinux_file_ioctl+0x16f/0x210
[   40.968117]  compat_blkdev_ioctl+0x21d/0x8f0
[   40.968864]  __x32_compat_sys_ioctl+0x99/0x250
[   40.969659]  do_syscall_64+0x4a/0x110
[   40.970337]  entry_SYSCALL_64_after_hwframe+0x44/0xa9

I haven't found a way to exploit it.

Quoted 6 lines
Fixes 229b53c9bf4e ("take floppy compat ioctls to sodding floppy.c")

---
compile-tested only

Quoted 18 lines
 drivers/block/floppy.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/block/floppy.c +1 -1
index 95f608d1a098..8c641245ff12 100644
@@ -3749,7 +3749,7 @@ static int compat_getdrvprm(int drive,
 	v.native_format = UDP->native_format;
 	mutex_unlock(&floppy_mutex);
 
-	if (copy_from_user(arg, &v, sizeof(struct compat_floppy_drive_params)))
+	if (copy_to_user(arg, &v, sizeof(struct compat_floppy_drive_params)))
 		return -EFAULT;
 	return 0;
 }

drivers/block/floppy.c static int compat_getdrvstat(int drive, bool poll, · line 3791

Quoted 8 lines
drivers/block/floppy.c +1 -1
@@ -3785,7 +3785,7 @@ static int compat_getdrvstat(int drive, bool poll,
 	v.bufblocks = UDRS->bufblocks;
 	mutex_unlock(&floppy_mutex);
 
-	if (copy_from_user(arg, &v, sizeof(struct compat_floppy_drive_struct)))
+	if (copy_to_user(arg, &v, sizeof(struct compat_floppy_drive_struct)))
 		return -EFAULT;
 	return 0;
 Eintr:

I also wrote a coccinelle rule for detecting similar bugs (adding coccinelle experts to CC).

Quoted 3 lines

virtual report

@cfu@

You can replace the above line with @cfu exists@. You want to find the existence of such a call, not ensure that the call occurs on every control-flow path, which is the default.

Do you want this rule to go into the kernel?

julia

Quoted 45 lines

identifier f;
type t;
identifier v;
position decl_p;
position copy_p;
@@

f(..., t v@decl_p, ...)
{
<+...
copy_from_user@copy_p(v, ...)
...+>
}

@script:python@
f << cfu.f;
t << cfu.t;
v << cfu.v;
decl_p << cfu.decl_p;
copy_p << cfu.copy_p;
@@

if '__user' in t:
msg0 = "function \"" + f + "\" has arg \"" + v + "\" of type \"" + t + "\"" coccilib.report.print_report(decl_p[0], msg0)
msg1 = "copy_from_user uses \"" + v + "\" as the destination. What a shame!\n" coccilib.report.print_report(copy_p[0], msg1)

The rule output:

./drivers/block/floppy.c:3756:49-52: function "compat_getdrvprm" has arg "arg" of type "struct compat_floppy_drive_params __user *" ./drivers/block/floppy.c:3783:5-19: copy_from_user uses "arg" as the destination. What a shame!

./drivers/block/floppy.c:3789:49-52: function "compat_getdrvstat" has arg "arg" of type "struct compat_floppy_drive_struct __user *" ./drivers/block/floppy.c:3819:5-19: copy_from_user uses "arg" as the destination. What a shame!

Best regards,
Alexander

#6 lore raw
Quoted 9 lines · Julia Lawall
Quoted 8 lines · Alexander Popov
Quoted 2 lines · Jann Horn

As sparse points out, these two copy_from_user() should actually be copy_to_user().

I also wrote a coccinelle rule for detecting similar bugs (adding coccinelle experts to CC).

Quoted 7 lines
Quoted 3 lines

virtual report

@cfu@

You can replace the above line with @cfu exists@. You want to find the existence of such a call, not ensure that the call occurs on every control-flow path, which is the default.

Thanks Julia, I see exists allows to drop <+ +>, right?

Quoted 1 line

Do you want this rule to go into the kernel?

It turned out that sparse already can find these bugs. Is this rule useful anyway? If so, I can prepare a patch.

Quoted 40 lines
Quoted 40 lines

identifier f;
type t;
identifier v;
position decl_p;
position copy_p;
@@

f(..., t v@decl_p, ...)
{
<+...
copy_from_user@copy_p(v, ...)
...+>
}

@script:python@
f << cfu.f;
t << cfu.t;
v << cfu.v;
decl_p << cfu.decl_p;
copy_p << cfu.copy_p;
@@

if '__user' in t:
msg0 = "function \"" + f + "\" has arg \"" + v + "\" of type \"" + t + "\"" coccilib.report.print_report(decl_p[0], msg0)
msg1 = "copy_from_user uses \"" + v + "\" as the destination. What a shame!\n" coccilib.report.print_report(copy_p[0], msg1)

The rule output:

./drivers/block/floppy.c:3756:49-52: function "compat_getdrvprm" has arg "arg" of type "struct compat_floppy_drive_params __user *" ./drivers/block/floppy.c:3783:5-19: copy_from_user uses "arg" as the destination. What a shame!

./drivers/block/floppy.c:3789:49-52: function "compat_getdrvstat" has arg "arg" of type "struct compat_floppy_drive_struct __user *" ./drivers/block/floppy.c:3819:5-19: copy_from_user uses "arg" as the destination. What a shame!

#7 lore raw
Quoted 1 line

…, I see exists allows to drop <+ +>, right?

I would interpret the combination of such SmPL specifications in a different way.

Quoted 1 line

It turned out that sparse already can find these bugs.

This is generally nice, isn't it?

Quoted 1 line

Is this rule useful anyway?

I hope so.

Can scripts for the semantic patch language help any more?

Quoted 1 line

If so, I can prepare a patch.

Would you like to take corresponding adjustments into account?

Regards,
Markus

#8 lore raw
In reply to Alexander Popov
Quoted 12 lines · Alexander Popov
Quoted 11 lines · Julia Lawall
Quoted 10 lines · Alexander Popov
Quoted 2 lines · Jann Horn

As sparse points out, these two copy_from_user() should actually be copy_to_user().

I also wrote a coccinelle rule for detecting similar bugs (adding coccinelle experts to CC).

virtual report

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

@cfu@

You can replace the above line with @cfu exists@. You want to find the existence of such a call, not ensure that the call occurs on every control-flow path, which is the default.

Thanks Julia, I see exists allows to drop <+ +>, right?

Exists is more efficient when it is possible. It just finds the existence of a path that has the property rather than collecting information about all paths. It is related to <+... ...+> because for that there has to exist at least one match. You could probably do something like

... when any
copy_from_user
... when any

Then with exists you will consider each call one at a time.

Quoted 4 lines
Quoted 2 lines

Do you want this rule to go into the kernel?

It turned out that sparse already can find these bugs.

If sparse is already doing this, then perhaps that's sufficient. Someone just has to be running it :)

julia

Quoted 43 lines

Is this rule useful anyway? If so, I can prepare a patch.

Quoted 40 lines
Quoted 40 lines

identifier f;
type t;
identifier v;
position decl_p;
position copy_p;
@@

f(..., t v@decl_p, ...)
{
<+...
copy_from_user@copy_p(v, ...)
...+>
}

@script:python@
f << cfu.f;
t << cfu.t;
v << cfu.v;
decl_p << cfu.decl_p;
copy_p << cfu.copy_p;
@@

if '__user' in t:
msg0 = "function \"" + f + "\" has arg \"" + v + "\" of type \"" + t + "\"" coccilib.report.print_report(decl_p[0], msg0)
msg1 = "copy_from_user uses \"" + v + "\" as the destination. What a shame!\n" coccilib.report.print_report(copy_p[0], msg1)

The rule output:

./drivers/block/floppy.c:3756:49-52: function "compat_getdrvprm" has arg "arg" of type "struct compat_floppy_drive_params __user *" ./drivers/block/floppy.c:3783:5-19: copy_from_user uses "arg" as the destination. What a shame!

./drivers/block/floppy.c:3789:49-52: function "compat_getdrvstat" has arg "arg" of type "struct compat_floppy_drive_struct __user *" ./drivers/block/floppy.c:3819:5-19: copy_from_user uses "arg" as the destination. What a shame!

#9 lore raw
In reply to Alexander Popov
Quoted 9 lines

@script:python@
f << cfu.f;
t << cfu.t;
v << cfu.v;
decl_p << cfu.decl_p;
copy_p << cfu.copy_p;
@@

if '__user' in t:

Can this check be specified as a constraint for a metavariable in the initial SmPL rule?
Would you like to move it to an other place?

Regards,
Markus