[PATCH] mtd: add module license to mtdbdi

[PATCH] mtd: add module license to mtdbdi

[Mariusz Kozlowski]

Hello,

	This seems to be a known issue [1] about Photron PNC-2000 module called
pnc2000. To reproduce this just 'modprobe pnc2000'. The patch attached at least
makes the kernel to be not tainted when mtdbdi gets loaded.

Signed-off-by: Mariusz Kozlowski <m.kozlowski@tuxland.pl>

 drivers/mtd/mtdbdi.c |    2 ++
 1 file changed, 2 insertions(+)

--- linux-2.6.22-rc6-bo-a/drivers/mtd/mtdbdi.c	2007-06-27 22:05:34.000000000 +0200
+++ linux-2.6.22-rc6-bo-b/drivers/mtd/mtdbdi.c	2007-06-27 23:20:57.000000000 +0200
@@ -46,3 +46,5 @@ struct backing_dev_info mtd_bdi_rw_mappa
 			   BDI_CAP_WRITE_MAP),
 };
 EXPORT_SYMBOL_GPL(mtd_bdi_rw_mappable);
+
+MODULE_LICENSE("GPL v2");



Regards,

	Mariusz

[1] http://groups.google.pl/group/linux.kernel/browse_frm/thread/77d9f9273ab14382

a.txt - unpatched
b.txt - patched

--- a.txt       2007-06-27 23:32:48.000000000 +0200
+++ b.txt       2007-06-27 23:33:33.000000000 +0200
@@ -1,4 +1,3 @@
-mtdbdi: module license 'unspecified' taints kernel.
 Photron PNC-2000 flash mapping: 400000 at bf000000
 BUG: unable to handle kernel paging request at virtual address bf000000
  printing eip:
@@ -6,27 +5,29 @@ ded752ff
 *pde = 00000000
 Oops: 0002 [#1]
 PREEMPT
-Modules linked in: cfi_probe gen_probe pnc2000 mtdpart mtdcore mtdbdi(P) chipreg eeprom i2c_ali1535 rtc_cmos rtc_core rtc_lib i2c_ali15x3 8139too 8250_pci 8250 serial_core yenta_socket rsrc_nonstatic
+Modules linked in: cfi_probe gen_probe pnc2000 mtdpart mtdcore mtdbdi chipreg eeprom i2c_ali1535 rtc_cmos rtc_core rtc_lib 8139too i2c_ali1
+5x3 yenta_socket rsrc_nonstatic 8250_pci 8250 serial_core
 CPU:    0
-EIP:    0060:[<ded752ff>]    Tainted: P        VLI
+EIP:    0060:[<ded752ff>]    Not tainted VLI
 EFLAGS: 00010286   (2.6.22-rc6-mm1 #2)
 EIP is at cfi_probe_chip+0xac/0xe61 [cfi_probe]
 eax: bf000000   ebx: 00000004   ecx: 00000010   edx: 00f000f0
-esi: ded564e0   edi: 00000002   ebp: c3511e14   esp: c3511db0
+esi: ded574e0   edi: 00000002   ebp: c354de14   esp: c354ddb0
 ds: 007b   es: 007b   fs: 0000  gs: 0033  ss: 0068
-Process modprobe (pid: 6075, ti=c3510000 task=c1b70ac0 task.ti=c3510000)
-Stack: 00000000 00000000 c3511e2c c0126d84 00000000 c3511dd0 00000000 00000000
-       00000000 00000001 dead4ead ffffffff ffffffff c043cc58 00000000 00000046
-       c3511df0 c3511df0 00000000 00000001 dead4ead ffffffff ded75253 00000002
+Process modprobe (pid: 6235, ti=c354c000 task=c263c030 task.ti=c354c000)
+Stack: 00000000 00000000 c354de2c c0126d84 00000000 c354ddd0 00000000 00000000
+       00000000 00000001 dead4ead ffffffff ffffffff c06aee70 00000000 c04f3d98
+       c354ddf0 c354ddf0 00000000 00000001 dead4ead ffffffff ded75253 00000002
 Call Trace:
- [<ded54131>] mtd_do_chip_probe+0x83/0x32a [gen_probe]
+ [<ded55131>] mtd_do_chip_probe+0x83/0x32a [gen_probe]
  [<ded75251>] cfi_probe+0xd/0xf [cfi_probe]
- [<de872185>] do_map_probe+0x1c/0x61 [chipreg]
- [<ded5c049>] init_pnc2000+0x49/0x76 [pnc2000]
+ [<ded51185>] do_map_probe+0x1c/0x61 [chipreg]
+ [<ded5d049>] init_pnc2000+0x49/0x76 [pnc2000]
  [<c0137c50>] sys_init_module+0x11b/0x160b
  [<c010401e>] sysenter_past_esp+0x5f/0x85
  [<ffffe410>] 0xffffe410
  =======================
 INFO: lockdep is turned off.
-Code: f0 8d 0c fd 00 00 00 00 89 c2 d3 e2 09 c2 83 fb 01 0f 84 af 0d 00 00 83 fb 02 0f 84 95 0d 00 00 83 fb 04 75 08 8b 45 b8 03 46 0c <89> 10 0f ae f0 89 f6 8b 5e 14 8b 55 08 8b 4a 08 89 d8 99 f7 f9
-EIP: [<ded752ff>] cfi_probe_chip+0xac/0xe61 [cfi_probe] SS:ESP 0068:c3511db0
+Code: f0 8d 0c fd 00 00 00 00 89 c2 d3 e2 09 c2 83 fb 01 0f 84 af 0d 00 00 83 fb 02 0f 84 95 0d 00 00 83 fb 04 75 08 8b 45 b8 03 46 0c <89>
+ 10 0f ae f0 89 f6 8b 5e 14 8b 55 08 8b 4a 08 89 d8 99 f7 f9
+EIP: [<ded752ff>] cfi_probe_chip+0xac/0xe61 [cfi_probe] SS:ESP 0068:c354ddb0

Re: [PATCH] mtd: add module license to mtdbdi

[David Howells]

Mariusz Kozlowski <m.kozlowski@tuxland.pl> wrote:

> 	This seems to be a known issue [1] about Photron PNC-2000 module called
> pnc2000. To reproduce this just 'modprobe pnc2000'. The patch attached at least
> makes the kernel to be not tainted when mtdbdi gets loaded.

This is the wrong solution.  This file should not be its own module, but
should rather be linked with mtdcore.

There was a patch for this, but I don't know what happened to it.

David