Skip to contentSkip to conversation

char: hpet: Use flexible-array member

PATCH In review Gustavo A. R. Silva 2 participants 4 messages Jan 21, 2020 – Jan 23, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 char: hpet: Use flexible-array member

Old code in the kernel uses 1-byte and 0-byte arrays to indicate the presence of a "variable length array":

struct something {
    int length;
    u8 data[1];
};

struct something *instance;

instance = kmalloc(sizeof(*instance) + size, GFP_KERNEL); instance->length = size;
memcpy(instance->data, source, size);

There is also 0-byte arrays. Both cases pose confusion for things like sizeof(), CONFIG_FORTIFY_SOURCE, etc.[1] Instead, the preferred mechanism to declare variable-length types such as the one above is a flexible array member[2] which need to be the last member of a structure and empty-sized:

struct something {
        int stuff;
        u8 data[];
};

Also, by making use of the mechanism above, we will get a compiler warning in case the flexible array does not occur last in the structure, which will help us prevent some kind of undefined behavior bugs from being unadvertenly introduced[3] to the codebase from now on.

[1] https://github.com/KSPP/linux/issues/21
[2] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

---
 drivers/char/hpet.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/char/hpet.c +1 -1
index 9ac6671bb514..aed2c45f7968 100644
@@ -110,7 +110,7 @@ struct hpets {
 	unsigned long hp_delta;
 	unsigned int hp_ntimer;
 	unsigned int hp_which;
-	struct hpet_dev hp_dev[1];
+	struct hpet_dev hp_dev[];
 };
 
 static struct hpets *hpets;
Signature · 1 line
2.23.0
#1 lore raw
Quoted 43 lines · Gustavo A. R. Silva

Old code in the kernel uses 1-byte and 0-byte arrays to indicate the presence of a "variable length array":

struct something {
    int length;
    u8 data[1];
};

struct something *instance;

instance = kmalloc(sizeof(*instance) + size, GFP_KERNEL); instance->length = size;
memcpy(instance->data, source, size);

There is also 0-byte arrays. Both cases pose confusion for things like sizeof(), CONFIG_FORTIFY_SOURCE, etc.[1] Instead, the preferred mechanism to declare variable-length types such as the one above is a flexible array member[2] which need to be the last member of a structure and empty-sized:

struct something {
        int stuff;
        u8 data[];
};

Also, by making use of the mechanism above, we will get a compiler warning in case the flexible array does not occur last in the structure, which will help us prevent some kind of undefined behavior bugs from being unadvertenly introduced[3] to the codebase from now on.

[1] https://github.com/KSPP/linux/issues/21
[2] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

---
 drivers/char/hpet.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/char/hpet.c +0 -0
index 9ac6671bb514..aed2c45f7968 100644

drivers/char/hpet.c struct hpets { · line 113

Quoted 5 lines
drivers/char/hpet.c +1 -1
@@ -110,7 +110,7 @@ struct hpets {
 	unsigned long hp_delta;
 	unsigned int hp_ntimer;
 	unsigned int hp_which;
-	struct hpet_dev hp_dev[1];
+	struct hpet_dev hp_dev[];

Are you sure the allocation size is the same again? Much like the n_hdlc patch was, I think you need to adjust the variable size here. Maybe, it's a bit of a pain to figure out at a quick glance, I just want to make sure you at least do look at that :)

thanks,

greg k-h

#2 lore raw
Quoted 39 lines · Greg Kroah-Hartman
Quoted 38 lines · Gustavo A. R. Silva

Old code in the kernel uses 1-byte and 0-byte arrays to indicate the presence of a "variable length array":

struct something {
    int length;
    u8 data[1];
};

struct something *instance;

instance = kmalloc(sizeof(*instance) + size, GFP_KERNEL); instance->length = size;
memcpy(instance->data, source, size);

There is also 0-byte arrays. Both cases pose confusion for things like sizeof(), CONFIG_FORTIFY_SOURCE, etc.[1] Instead, the preferred mechanism to declare variable-length types such as the one above is a flexible array member[2] which need to be the last member of a structure and empty-sized:

struct something {
        int stuff;
        u8 data[];
};

Also, by making use of the mechanism above, we will get a compiler warning in case the flexible array does not occur last in the structure, which will help us prevent some kind of undefined behavior bugs from being unadvertenly introduced[3] to the codebase from now on.

[1] https://github.com/KSPP/linux/issues/21
[2] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

---
 drivers/char/hpet.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 16 lines

drivers/char/hpet.c struct hpets { · line 113

Quoted 10 lines
drivers/char/hpet.c +1 -1
index 9ac6671bb514..aed2c45f7968 100644
@@ -110,7 +110,7 @@ struct hpets {
 	unsigned long hp_delta;
 	unsigned int hp_ntimer;
 	unsigned int hp_which;
-	struct hpet_dev hp_dev[1];
+	struct hpet_dev hp_dev[];

Are you sure the allocation size is the same again? Much like the n_hdlc patch was, I think you need to adjust the variable size here. Maybe, it's a bit of a pain to figure out at a quick glance, I just want to make sure you at least do look at that :)

Yep. The allocation thing was already handled almost a year ago by the following patch, and it didn't require to increase the size at that time:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=401c9bd10beef4b030eb9e34d16b5341dc6c683b

Thanks
--
Gustavo

#3 lore raw
Quoted 59 lines · Gustavo A. R. Silva
Quoted 39 lines · Greg Kroah-Hartman
Quoted 38 lines · Gustavo A. R. Silva

Old code in the kernel uses 1-byte and 0-byte arrays to indicate the presence of a "variable length array":

struct something {
    int length;
    u8 data[1];
};

struct something *instance;

instance = kmalloc(sizeof(*instance) + size, GFP_KERNEL); instance->length = size;
memcpy(instance->data, source, size);

There is also 0-byte arrays. Both cases pose confusion for things like sizeof(), CONFIG_FORTIFY_SOURCE, etc.[1] Instead, the preferred mechanism to declare variable-length types such as the one above is a flexible array member[2] which need to be the last member of a structure and empty-sized:

struct something {
        int stuff;
        u8 data[];
};

Also, by making use of the mechanism above, we will get a compiler warning in case the flexible array does not occur last in the structure, which will help us prevent some kind of undefined behavior bugs from being unadvertenly introduced[3] to the codebase from now on.

[1] https://github.com/KSPP/linux/issues/21
[2] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

---
 drivers/char/hpet.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 16 lines

drivers/char/hpet.c struct hpets { · line 113

Quoted 10 lines
drivers/char/hpet.c +1 -1
index 9ac6671bb514..aed2c45f7968 100644
@@ -110,7 +110,7 @@ struct hpets {
 	unsigned long hp_delta;
 	unsigned int hp_ntimer;
 	unsigned int hp_which;
-	struct hpet_dev hp_dev[1];
+	struct hpet_dev hp_dev[];

Are you sure the allocation size is the same again? Much like the n_hdlc patch was, I think you need to adjust the variable size here. Maybe, it's a bit of a pain to figure out at a quick glance, I just want to make sure you at least do look at that :)

Quoted 4 lines

Yep. The allocation thing was already handled almost a year ago by the following patch, and it didn't require to increase the size at that time:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=401c9bd10beef4b030eb9e34d16b5341dc6c683b

Great, thanks for verifying, I'll go queue this up now.

greg k-h