sizeof(ptr) or sizeof(*ptr)?
Last week a bug was detected in n_tty.c where an array of char was replaced by a char pointer making a "(len > sizeof(buf))" condition test for len > 4 (or 8) bytes, instead of the original array size.
I decided to tweak sparse to give warnings on sizeof(pointer), so that I could check for other cases like this. The tweak was a very crude hack that I'm not proud of, and I am still trying to make it more reliable.
So far I found 2 interesting cases (in 2.6.11-rc5). I'm not sure they are bugs, but they sure look suspicious.
1: drivers/usb/storage/usb.c:788
/*
* Since this is a new device, we need to register a SCSI
* host definition with the higher SCSI layers.
*/
us->host = scsi_host_alloc(&usb_stor_host_template, sizeof(us));
if (!us->host) {
printk(KERN_WARNING USB_STORAGE
"Unable to allocate the scsi host\n");
return -EBUSY;
}"us" is a "struct us_data *". It seems pretty weird that we're allocating something the size of a pointer, and then waste a pointer to keep the address where it is allocated. So maybe this should be:
us->host = scsi_host_alloc(&usb_stor_host_template, sizeof(*us));
2: sound/core/control.c:936
ue = kcalloc(1, sizeof(struct user_element) + private_size + extra_size,
GFP_KERNEL);
if (ue == NULL)
return -ENOMEM;
ue->info = info;
ue->elem_data = (char *)ue + sizeof(ue);
ue->elem_data_size = private_size;
if (extra_size) {
ue->priv_data = (char *)ue + sizeof(ue) + private_size;
ue->priv_data_size = extra_size;
if (ue->info.type == SNDRV_CTL_ELEM_TYPE_ENUMERATED) {
if (copy_from_user(ue->priv_data, *(char __user
**)info.value.enumerated.name, extra_size))
return -EFAULT;
}
}If we're allocating "sizeof(struct user_element) + private_size + extra_size" it seems that in the instructions below we would be wanting to use that space, so both "sizeof(ue)" there should in fact be "sizeof(*ue)"
I'm not *really* sure that these are bugs, but they look suspicious. I'm CC'ing the maintainers of both these files so that they can check these out.
I'll probably bring more examples in the future, as the tool improves. Right now it gives about 550 false positives, so I think it is better to improve it before checking every case :)
--
Paulo Marques - www.grupopie.com
All that is necessary for the triumph of evil is that good men do nothing. Edmund Burke (1729 - 1797)
Quoted 20 lines · On Sun, Feb 27, 2005 at 08:25:04PM +0000, [email protected] wrote:
I decided to tweak sparse to give warnings on sizeof(pointer), so that I could check for other cases like this. The tweak was a very crude hack that I'm not proud of, and I am still trying to make it more reliable.
So far I found 2 interesting cases (in 2.6.11-rc5). I'm not sure they are bugs, but they sure look suspicious.
1:
drivers/usb/storage/usb.c:788/* * Since this is a new device, we need to register a SCSI * host definition with the higher SCSI layers. */ us->host = scsi_host_alloc(&usb_stor_host_template, sizeof(us)); if (!us->host) { printk(KERN_WARNING USB_STORAGE "Unable to allocate the scsi host\n"); return -EBUSY; }
Quoted 5 lines
"us" is a "struct
us_data*". It seems pretty weird that we're allocating something the size of a pointer, and then waste a pointer to keep the address where it is allocated. So maybe this should be:us->host = scsi_host_alloc(&usb_stor_host_template, sizeof(*us));
This is actually correct as-is. We're allocating a host, and asking for the sizeof(pointer) in the 'extra storage' section. We then store the pointer (not what it points to) in the extra storage section a few lines down.
Matt
Signature · 7 lines
Matthew Dharm Home: [email protected] Maintainer, Linux USB Mass Storage Driver S: Another stupid question? G: There's no such thing as a stupid question, only stupid people. -- Stef and Greg User Friendly, 7/15/1998
Quoting Matthew Dharm <[email protected]>:
Quoted 7 lines
[...]
us->host = scsi_host_alloc(&usb_stor_host_template, sizeof(*us));This is actually correct as-is. We're allocating a host, and asking for the sizeof(pointer) in the 'extra storage' section. We then store the pointer (not what it points to) in the extra storage section a few lines down.
Thanks for clarifying that. I guess the weekend effect got me, because at a certain point I was starting to read the scsi_host_alloc as if it were a kmalloc or something... :P
Anyway, after improving the tool and checking for false positives, there is only one more suspicious piece of code in drivers/acpi/video.c:561
status = acpi_video_device_lcd_query_levels(device, &obj);
if (obj && obj->type == ACPI_TYPE_PACKAGE && obj->package.count >= 2) {
int count = 0;
union acpi_object *o;
br = kmalloc(sizeof &br, GFP_KERNEL);
if (!br) {
printk(KERN_ERR "can't allocate memory\n");
} else {
memset(br, 0, sizeof &br);
br->levels = kmalloc(obj->package.count * sizeof &br->levels, GFP_KERNEL);
if (!br->levels)
goto out;"br" if of type "struct acpi_video_device_brightness *".
"sizeof &br" doesn't make much sense there (besides the unconventional use of sizeof without parenthesis) because the rest of the code seem to suggest that a whole structure should have been allocated. This is the last case I've seen, and I've added the maintainer to the cc list, so that he can check the code for correctness.
Well, sorry about bothering you with a false positive (that I should've recognised on my own, anyway), and thanks for the help.
--
Paulo Marques - www.grupopie.com
All that is necessary for the triumph of evil is that good men do nothing. Edmund Burke (1729 - 1797)
Review
Quoted 11 lines · "" <[email protected]> wrote:
Anyway, after improving the tool and checking for false positives, there is only
one more suspicious piece of code indrivers/acpi/video.c:561(quoted diff) +0
-0status = acpi_video_device_lcd_query_levels(device, &obj); if (obj && obj->type == ACPI_TYPE_PACKAGE && obj->package.count >= 2) { int count = 0; union acpi_object *o; br = kmalloc(sizeof &br, GFP_KERNEL);
yup, bug.
Review
Quoted 5 lines
(quoted diff) +0
-0if (!br) { printk(KERN_ERR "can't allocate memory\n"); } else { memset(br, 0, sizeof &br); br->levels = kmalloc(obj->package.count * sizeof &br->levels, GFP_KERNEL);
And another one, although it happens to work out OK.
I'll get these all fixed up, thanks.
Quoted 24 lines · Andrew Morton
Quoted 10 lines · "" <[email protected]> wrote:
Anyway, after improving the tool and checking for false positives, there is only one more suspicious piece of code in
drivers/acpi/video.c:561status = acpi_video_device_lcd_query_levels(device, &obj); if (obj && obj->type == ACPI_TYPE_PACKAGE && obj->package.count >= 2) { int count = 0; union acpi_object *o; br = kmalloc(sizeof &br, GFP_KERNEL);yup, bug.
Quoted 5 lines
if (!br) { printk(KERN_ERR "can't allocate memory\n"); } else { memset(br, 0, sizeof &br); br->levels = kmalloc(obj->package.count * sizeof &br->levels, GFP_KERNEL);
Quoted 3 lines
And another one, although it happens to work out OK.
I'll get these all fixed up, thanks.
I just checked the 2.6.11-mm1 release and this is only half-fixed there, and it is the worst of both halves: the kmalloc only mallocs the size of a pointer, but the memset is fixed, so it memset's the size of a structure (oops). This is partially my fault for not sending the patch in the first place, together with the bug report.
The attached patch against 2.6.11-mm1 should fix the kmalloc.
By the way, I haven't got any response from an alsa developer about the bug in sound/core/control.c, but this is already fixed in 2.6.11-mm1, along with several other changes to that file. So the status is: it was a bug, but it is already fixed :)
Signature · 4 lines
Paulo Marques - www.grupopie.com All that is necessary for the triumph of evil is that good men do nothing. Edmund Burke (1729 - 1797)
drivers/acpi/video.c renamed from drivers/acpi/video.c.orig +2 -2
@@ -564,11 +564,11 @@ acpi_video_device_find_cap (struct acpi_int count =0 ;union acpi_object *o; - br =kmalloc (sizeof &br,GFP_KERNEL ); + br =kmalloc (sizeof (*br),GFP_KERNEL );if (!br) {printk (KERN_ERR "can't allocate memory\n" ); }else { -memset (br,0 ,sizeof *br); +memset (br,0 ,sizeof (*br)); br->levels =kmalloc (obj->package.count *sizeof *(br->levels),GFP_KERNEL );if (!br->levels)