Skip to contentSkip to conversation

binfmt_misc: pass binfmt_misc P flag to the interpreter

PATCH In review YunQiang Su 3 participants 11 messages Mar 6, 2020 – Mar 11, 2020

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

PATCH 1/1 binfmt_misc: pass binfmt_misc P flag to the interpreter

From: Laurent Vivier <[email protected]>

It can be useful to the interpreter to know which flags are in use.

For instance, knowing if the preserve-argv[0] is in use would allow to skip the pathname argument.

This patch uses an unused auxiliary vector, AT_FLAGS, to add a flag to inform interpreter if the preserve-argv[0] is enabled.

---

Notes:
This patch uses AT_FLAGS, which has never been used. We need to make a decision. since it is about the fundmental API of userland.

This can be tested with QEMU from my branch:

https://github.com/vivier/qemu/commits/binfmt-argv0

With something like:

# cp ..../qemu-ppc /chroot/powerpc/jessie

# qemu-binfmt-conf.sh --qemu-path / --systemd ppc --credential yes \
--persistent no --preserve-argv0 yes
# systemctl restart systemd-binfmt.service
# cat /proc/sys/fs/binfmt_misc/qemu-ppc
enabled
interpreter //qemu-ppc
flags: POC
offset 0
magic 7f454c4601020100000000000000000000020014
mask ffffffffffffff00fffffffffffffffffffeffff
# chroot /chroot/powerpc/jessie sh -c 'echo $0'
sh

# qemu-binfmt-conf.sh --qemu-path / --systemd ppc --credential yes \
--persistent no --preserve-argv0 no
# systemctl restart systemd-binfmt.service
# cat /proc/sys/fs/binfmt_misc/qemu-ppc
enabled
interpreter //qemu-ppc
flags: OC
offset 0
magic 7f454c4601020100000000000000000000020014
mask ffffffffffffff00fffffffffffffffffffeffff
# chroot /chroot/powerpc/jessie sh -c 'echo $0'
/bin/sh

v4: Update to merge with linux-next: 20200305.
v3: mix my patch with one from YunQiang Su and my comments on it
introduce a new flag in the uabi for the AT_FLAGS
v2: only pass special flags (remove Magic and Enabled flags)

---
 fs/binfmt_elf.c              | 6 ++++--
 fs/binfmt_elf_fdpic.c        | 5 ++++-
 fs/binfmt_misc.c             | 4 +++-
 include/linux/binfmts.h      | 4 ++++
 include/uapi/linux/binfmts.h | 4 ++++
 5 files changed, 19 insertions(+), 4 deletions(-)
fs/binfmt_elf.c +4 -2
index 13f25e241ac4..e9f5d135b847 100644
@@ -178,7 +178,7 @@ create_elf_tables(struct linux_binprm *bprm, const struct elfhdr *exec,
 	const char *k_base_platform = ELF_BASE_PLATFORM;
 	unsigned char k_rand_bytes[16];
 	int items;
-	elf_addr_t *elf_info;
+	elf_addr_t *elf_info, flags = 0;
 	int ei_index;
 	const struct cred *cred = current_cred();
 	struct vm_area_struct *vma;
@@ -253,7 +253,9 @@ create_elf_tables(struct linux_binprm *bprm, const struct elfhdr *exec,
 	NEW_AUX_ENT(AT_PHENT, sizeof(struct elf_phdr));
 	NEW_AUX_ENT(AT_PHNUM, exec->e_phnum);
 	NEW_AUX_ENT(AT_BASE, interp_load_addr);
-	NEW_AUX_ENT(AT_FLAGS, 0);
+	if (bprm->interp_flags & BINPRM_FLAGS_PRESERVE_ARGV0)
+		flags |= AT_FLAGS_PRESERVE_ARGV0;
+	NEW_AUX_ENT(AT_FLAGS, flags);
 	NEW_AUX_ENT(AT_ENTRY, e_entry);
 	NEW_AUX_ENT(AT_UID, from_kuid_munged(cred->user_ns, cred->uid));
 	NEW_AUX_ENT(AT_EUID, from_kuid_munged(cred->user_ns, cred->euid));
fs/binfmt_elf_fdpic.c +4 -1
index 240f66663543..abb90d82aa58 100644
@@ -507,6 +507,7 @@ static int create_elf_fdpic_tables(struct linux_binprm *bprm,
 	char __user *u_platform, *u_base_platform, *p;
 	int loop;
 	int nr;	/* reset for each csp adjustment */
+	unsigned long flags = 0;
 
 #ifdef CONFIG_MMU
 	/* In some cases (e.g. Hyper-Threading), we want to avoid L1 evictions
@@ -647,7 +648,9 @@ static int create_elf_fdpic_tables(struct linux_binprm *bprm,
 	NEW_AUX_ENT(AT_PHENT,	sizeof(struct elf_phdr));
 	NEW_AUX_ENT(AT_PHNUM,	exec_params->hdr.e_phnum);
 	NEW_AUX_ENT(AT_BASE,	interp_params->elfhdr_addr);
-	NEW_AUX_ENT(AT_FLAGS,	0);
+	if (bprm->interp_flags & BINPRM_FLAGS_PRESERVE_ARGV0)
+		flags |= AT_FLAGS_PRESERVE_ARGV0;
+	NEW_AUX_ENT(AT_FLAGS,	flags);
 	NEW_AUX_ENT(AT_ENTRY,	exec_params->entry_addr);
 	NEW_AUX_ENT(AT_UID,	(elf_addr_t) from_kuid_munged(cred->user_ns, cred->uid));
 	NEW_AUX_ENT(AT_EUID,	(elf_addr_t) from_kuid_munged(cred->user_ns, cred->euid));
fs/binfmt_misc.c +3 -1
index cdb45829354d..b9acdd26a654 100644
@@ -154,7 +154,9 @@ static int load_misc_binary(struct linux_binprm *bprm)
 	if (bprm->interp_flags & BINPRM_FLAGS_PATH_INACCESSIBLE)
 		goto ret;
 
-	if (!(fmt->flags & MISC_FMT_PRESERVE_ARGV0)) {
+	if (fmt->flags & MISC_FMT_PRESERVE_ARGV0) {
+		bprm->interp_flags |= BINPRM_FLAGS_PRESERVE_ARGV0;
+	} else {
 		retval = remove_arg_zero(bprm);
 		if (retval)
 			goto ret;
include/linux/binfmts.h +4 -0
index b40fc633f3be..265b80d5fd6f 100644
@@ -78,6 +78,10 @@ struct linux_binprm {
 #define BINPRM_FLAGS_PATH_INACCESSIBLE_BIT 2
 #define BINPRM_FLAGS_PATH_INACCESSIBLE (1 << BINPRM_FLAGS_PATH_INACCESSIBLE_BIT)
 
+/* if preserve the argv0 for the interpreter  */
+#define BINPRM_FLAGS_PRESERVE_ARGV0_BIT 3
+#define BINPRM_FLAGS_PRESERVE_ARGV0 (1 << BINPRM_FLAGS_PRESERVE_ARGV0_BIT)
+
 /* Function parameter for binfmt->coredump */
 struct coredump_params {
 	const kernel_siginfo_t *siginfo;
include/uapi/linux/binfmts.h +4 -0
index 689025d9c185..a70747416130 100644
@@ -18,4 +18,8 @@ struct pt_regs;
 /* sizeof(linux_binprm->buf) */
 #define BINPRM_BUF_SIZE 256
 
+/* if preserve the argv0 for the interpreter  */
+#define AT_FLAGS_PRESERVE_ARGV0_BIT 0
+#define AT_FLAGS_PRESERVE_ARGV0 (1 << AT_FLAGS_PRESERVE_ARGV0_BIT)
+
 #endif /* _UAPI_LINUX_BINFMTS_H */
Signature · 1 line
2.25.1
#1 lore raw

* YunQiang Su:

fs/binfmt_elf.c create_elf_tables(struct linux_binprm *bprm, const struct elfhdr *exec, · line 258

Quoted 3 lines
fs/binfmt_elf.c +3 -0
+	if (bprm->interp_flags & BINPRM_FLAGS_PRESERVE_ARGV0)
+		flags |= AT_FLAGS_PRESERVE_ARGV0;
+	NEW_AUX_ENT(AT_FLAGS, flags);

Is it necessary to reuse AT_FLAGS? I think it's cleaner to define a separate AT_ tag dedicated to binfmt_misc.

#2 lore raw

Le 06/03/2020 à 09:13, Florian Weimer a écrit :

Quoted 9 lines

* YunQiang Su:

fs/binfmt_elf.c create_elf_tables(struct linux_binprm *bprm, const struct elfhdr *exec, · line 258

Quoted 3 lines
fs/binfmt_elf.c +3 -0
+	if (bprm->interp_flags & BINPRM_FLAGS_PRESERVE_ARGV0)
+		flags |= AT_FLAGS_PRESERVE_ARGV0;
+	NEW_AUX_ENT(AT_FLAGS, flags);

Is it necessary to reuse AT_FLAGS? I think it's cleaner to define a separate AT_ tag dedicated to binfmt_misc.

Not necessary, but it seemed simpler and cleaner to re-use a flag that is marked as unused and with a name matching the new role. It avoids to patch other packages (like glibc) to add it as it is already defined.

Thanks,
Laurent

#3 lore raw

* Laurent Vivier:

Quoted 10 lines

Le 06/03/2020 à 09:13, Florian Weimer a écrit :

Quoted 8 lines

* YunQiang Su:

fs/binfmt_elf.c create_elf_tables(struct linux_binprm *bprm, const struct elfhdr *exec, · line 258

Quoted 3 lines
fs/binfmt_elf.c +3 -0
+	if (bprm->interp_flags & BINPRM_FLAGS_PRESERVE_ARGV0)
+		flags |= AT_FLAGS_PRESERVE_ARGV0;
+	NEW_AUX_ENT(AT_FLAGS, flags);

Is it necessary to reuse AT_FLAGS? I think it's cleaner to define a separate AT_ tag dedicated to binfmt_misc.

Quoted 3 lines

Not necessary, but it seemed simpler and cleaner to re-use a flag that is marked as unused and with a name matching the new role. It avoids to patch other packages (like glibc) to add it as it is already defined.

You still need to define AT_FLAGS_PRESERVE_ARGV0. At that point, you might as well define AT_BINFMT and AT_BINFMT_PRESERVE_ARGV0.

#4 lore raw

Le 06/03/2020 à 09:37, Florian Weimer a écrit :

Quoted 12 lines

* Laurent Vivier:

Quoted 10 lines

Le 06/03/2020 à 09:13, Florian Weimer a écrit :

Quoted 8 lines

* YunQiang Su:

fs/binfmt_elf.c create_elf_tables(struct linux_binprm *bprm, const struct elfhdr *exec, · line 258

Quoted 3 lines
fs/binfmt_elf.c +3 -0
+	if (bprm->interp_flags & BINPRM_FLAGS_PRESERVE_ARGV0)
+		flags |= AT_FLAGS_PRESERVE_ARGV0;
+	NEW_AUX_ENT(AT_FLAGS, flags);

Is it necessary to reuse AT_FLAGS? I think it's cleaner to define a separate AT_ tag dedicated to binfmt_misc.

Quoted 7 lines
Quoted 3 lines

Not necessary, but it seemed simpler and cleaner to re-use a flag that is marked as unused and with a name matching the new role. It avoids to patch other packages (like glibc) to add it as it is already defined.

You still need to define AT_FLAGS_PRESERVE_ARGV0. At that point, you might as well define AT_BINFMT and AT_BINFMT_PRESERVE_ARGV0.

Yes, you're right.

But is there any reason to not reuse AT_FLAGS?

Thanks,
Laurent

#5 lore raw
Quoted 18 lines · Laurent Vivier <[email protected]> 于2020年3月6日周五…
Quoted 16 lines · Florian Weimer

* Laurent Vivier:

Quoted 13 lines · Florian Weimer
Quoted 8 lines

* YunQiang Su:

fs/binfmt_elf.c create_elf_tables(struct linux_binprm *bprm, const struct elfhdr *exec, · line 258

Quoted 3 lines
fs/binfmt_elf.c +3 -0
+  if (bprm->interp_flags & BINPRM_FLAGS_PRESERVE_ARGV0)
+          flags |= AT_FLAGS_PRESERVE_ARGV0;
+  NEW_AUX_ENT(AT_FLAGS, flags);

Is it necessary to reuse AT_FLAGS? I think it's cleaner to define a separate AT_ tag dedicated to binfmt_misc.

Not necessary, but it seemed simpler and cleaner to re-use a flag that is marked as unused and with a name matching the new role. It avoids to patch other packages (like glibc) to add it as it is already defined.

Quoted 7 lines
Quoted 3 lines

You still need to define AT_FLAGS_PRESERVE_ARGV0. At that point, you might as well define AT_BINFMT and AT_BINFMT_PRESERVE_ARGV0.

Yes, you're right.

But is there any reason to not reuse AT_FLAGS?

AT_* only has 32 slot and now. I was afraid that maybe we shouldn't take one.
/* AT_* values 18 through 22 are reserved */
27,28,29,30 are not used now.
Which should we use?

Quoted 3 lines

Thanks,
Laurent

#6 lore raw

* YunQiang Su:

Quoted 4 lines

AT_* only has 32 slot and now. I was afraid that maybe we shouldn't take one.
/* AT_* values 18 through 22 are reserved */
27,28,29,30 are not used now.
Which should we use?

Where does this limit of 32 tags come from? I don't see it from a userspace perspective.

#7 lore raw
Quoted 10 lines · Florian Weimer <[email protected]> 于2020年3月6日周五 下午7:42写道:

* YunQiang Su:

Quoted 4 lines

AT_* only has 32 slot and now. I was afraid that maybe we shouldn't take one.
/* AT_* values 18 through 22 are reserved */
27,28,29,30 are not used now.
Which should we use?

Where does this limit of 32 tags come from? I don't see it from a userspace perspective.

Sorry it is my mistake: In linux/auxvec.h, I saw

#define AT_RANDOM 25 /* address of 16 random bytes */
#define AT_HWCAP2 26 /* extension of AT_HWCAP */

#define AT_EXECFN 31 /* filename of program */

The number jump to 31 from 26.

It is my fault: in x86_64-linux-gnu/bits/auxv.h, the max number is 47 now.

#8 lore raw

* YunQiang Su:

Quoted 14 lines
Quoted 10 lines · Florian Weimer <[email protected]> 于2020年3月6日周五 下午7:42写道:

* YunQiang Su:

Quoted 4 lines

AT_* only has 32 slot and now. I was afraid that maybe we shouldn't take one.
/* AT_* values 18 through 22 are reserved */
27,28,29,30 are not used now.
Which should we use?

Where does this limit of 32 tags come from? I don't see it from a userspace perspective.

Sorry it is my mistake: In linux/auxvec.h, I saw

Quoted 8 lines

#define AT_RANDOM 25 /* address of 16 random bytes */
#define AT_HWCAP2 26 /* extension of AT_HWCAP */

#define AT_EXECFN 31 /* filename of program */

The number jump to 31 from 26.

It is my fault: in x86_64-linux-gnu/bits/auxv.h, the max number is 47 now.

So AT_* tags aren't a scarce resource after all?

#9 lore raw
In reply to YunQiang Su

Le 06/03/2020 à 12:48, YunQiang Su a écrit :

Quoted 14 lines
Quoted 10 lines · Florian Weimer <[email protected]> 于2020年3月6日周五…

* YunQiang Su:

Quoted 4 lines

AT_* only has 32 slot and now. I was afraid that maybe we shouldn't take one.
/* AT_* values 18 through 22 are reserved */
27,28,29,30 are not used now.
Which should we use?

Where does this limit of 32 tags come from? I don't see it from a userspace perspective.

Sorry it is my mistake: In linux/auxvec.h, I saw

Quoted 9 lines

#define AT_RANDOM 25 /* address of 16 random bytes */
#define AT_HWCAP2 26 /* extension of AT_HWCAP */

#define AT_EXECFN 31 /* filename of program */

The number jump to 31 from 26.

It is my fault: in x86_64-linux-gnu/bits/auxv.h, the max number is 47 now.

Numbers starting from 32 are arch specific. 18 to 22 are also reserved (and used by some archs). (linux/arch/*/include/uapi/asm/auxvec.h)

So there is only 4 entries available (27 to 30) (linux/include/uapi/linux/auxvec.h)
Do we want to waste one more entry whereas we can use an unused one?

Thanks,
Laurent

#10 lore raw
In reply to YunQiang Su

Le 06/03/2020 à 09:09, YunQiang Su a écrit :

Quoted 7 lines

From: Laurent Vivier <[email protected]>

It can be useful to the interpreter to know which flags are in use.

For instance, knowing if the preserve-argv[0] is in use would allow to skip the pathname argument.

Quoted 6 lines

This patch uses an unused auxiliary vector, AT_FLAGS, to add a flag to inform interpreter if the preserve-argv[0] is enabled.

---

Any advice to have this merged?

Do I really need to use another AT_ entry?

Thanks,
Laurent