Skip to contentSkip to conversation

+ tpm_bios-needs-more-securityfs_-functions.patch added to -mm tree

DISCUSSION Randy.Dunlap 1 participant 1 message Jan 20, 2006 2 messages not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

#2 lore raw
Quoted 67 lines · On Thu, 19 Jan 2006 20:51:15 -0800 Greg KH wrote:
Quoted 66 lines · On Thu, Jan 19, 2006 at 08:40:23PM -0800, [email protected] wrote:

The patch titled

tpm_bios: needs more securityfs_ functions

has been added to the -mm tree. Its filename is

tpm_bios-needs-more-securityfs_-functions.patch

See http://www.zip.com.au/~akpm/linux/patches/stuff/added-to-mm.txt to find out what to do about this

From: Randy Dunlap <[email protected]>

tpm_bios.c needs securityfs_xyz() functions.

Does include/linux/security.h need stubs for these, or should char/tpm/Makefile just be modified to say:

ifdef CONFIG_ACPI
ifdef CONFIG_SECURITY

	obj-$(CONFIG_TCG_TPM) += tpm_bios.o

endif
endif

drivers/char/tpm/tpm_bios.c:494: warning: implicit declaration of function 'securityfs_create_dir' drivers/char/tpm/tpm_bios.c:494: warning: assignment makes pointer from integer without a cast drivers/char/tpm/tpm_bios.c:499: warning: implicit declaration of function 'securityfs_create_file' drivers/char/tpm/tpm_bios.c:501: warning: assignment makes pointer from integer without a cast drivers/char/tpm/tpm_bios.c:508: warning: assignment makes pointer from integer without a cast drivers/char/tpm/tpm_bios.c:523: warning: implicit declaration of function 'securityfs_remove' *** Warning: "securityfs_create_file" [drivers/char/tpm/tpm_bios.ko] undefined!
*** Warning: "securityfs_create_dir" [drivers/char/tpm/tpm_bios.ko] undefined!
*** Warning: "securityfs_remove" [drivers/char/tpm/tpm_bios.ko] undefined!

There are also some gcc and sparse warnings that could be fixed. (see http://www.xenotime.net/linux/doc/build-tpm.out)

---

 include/linux/security.h |   19 +++++++++++++++++++
 1 files changed, 19 insertions(+)

diff -puN include/linux/security.h~tpm_bios-needs-more-securityfs_-functions include/linux/security.h

include/linux/security.h renamed from include/linux/security.h~tpm_bios-needs-more-securityfs_-functions +10 -0
@@ -2617,6 +2617,25 @@ static inline int security_netlink_recv 
 	return cap_netlink_recv (skb);
 }
 
+static inline struct dentry *securityfs_create_dir(const char *name,
+					struct dentry *parent)
+{
+	return NULL;
+}
+
+static inline struct dentry *securityfs_create_file(const char *name,
+						mode_t mode,
+						struct dentry *parent,
+						void *data,
Quoted 8 lines

include/linux/security.h static inline int security_netlink_recv · line 2633

Quoted 4 lines
include/linux/security.h renamed from include/linux/security.h~tpm_bios-needs-more-securityfs_-functions +4 -0
+						struct file_operations *fops)
+{
+	return NULL;
+}

No, these should return ERR_PTR(-ENODEV) so that you can actually test for a real error in the code, and not have to #ifdef if SECURITY is enabled or not (meaning that you can check for NULL for a real error.)

Now that Andrew has changed that (thanks!), the callers in tpm_bios.c look incorrect. They only check for 0/non-zero.

Too late for me to fix that tonight. Maybe *@us.ibm.com can do that.

---
~Randy