2016-03-11 06:31:44

by Dave Young

[permalink] [raw]
Subject: [PATCH 2/2] [PATCH 2/2] proc-vmcore: wrong data type casting fix

On i686 PAE enabled machine the contiguous physical area could be large
and it can cause triming down variables in below calculation in
read_vmcore() and mmap_vmcore():

tsz = min_t(size_t, m->offset + m->size - *fpos, buflen);

Then the real size passed down is not correct any more.
Suppose m->offset + m->size - *fpos being truncated to 0, buflen >0 then
we will get tsz = 0. It is of course not an expected result.

During out test there are two problems caused by it:
1) read_vmcore will refuse to continue so makedumpfile fails.
2) mmap_vmcore will trigger BUG_ON() in remap_pfn_range().

Use min_lt instead so that the variables are not truncated.

Signed-off-by: Baoquan He <[email protected]>
Signed-off-by: Dave Young <[email protected]>
---
fs/proc/vmcore.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

--- linux.orig/fs/proc/vmcore.c
+++ linux/fs/proc/vmcore.c
@@ -231,7 +231,8 @@ static ssize_t __read_vmcore(char *buffe

list_for_each_entry(m, &vmcore_list, list) {
if (*fpos < m->offset + m->size) {
- tsz = min_t(size_t, m->offset + m->size - *fpos, buflen);
+ tsz = (size_t)min_lt(m->offset + m->size - *fpos,
+ buflen);
start = m->paddr + *fpos - m->offset;
tmp = read_from_oldmem(buffer, tsz, &start, userbuf);
if (tmp < 0)
@@ -461,7 +462,7 @@ static int mmap_vmcore(struct file *file
if (start < m->offset + m->size) {
u64 paddr = 0;

- tsz = min_t(size_t, m->offset + m->size - start, size);
+ tsz = (size_t)min_lt(m->offset + m->size - start, size);
paddr = m->paddr + start - m->offset;
if (vmcore_remap_oldmem_pfn(vma, vma->vm_start + len,
paddr >> PAGE_SHIFT, tsz,



2016-03-11 07:15:43

by Minfei Huang

[permalink] [raw]
Subject: Re: [PATCH 2/2] [PATCH 2/2] proc-vmcore: wrong data type casting fix

On 03/11/16 at 02:21pm, [email protected] wrote:
> @@ -231,7 +231,8 @@ static ssize_t __read_vmcore(char *buffe
>
> list_for_each_entry(m, &vmcore_list, list) {
> if (*fpos < m->offset + m->size) {
> - tsz = min_t(size_t, m->offset + m->size - *fpos, buflen);
> + tsz = (size_t)min_lt(m->offset + m->size - *fpos,
> + buflen);
> start = m->paddr + *fpos - m->offset;
> tmp = read_from_oldmem(buffer, tsz, &start, userbuf);
> if (tmp < 0)
> @@ -461,7 +462,7 @@ static int mmap_vmcore(struct file *file
> if (start < m->offset + m->size) {
> u64 paddr = 0;
>
> - tsz = min_t(size_t, m->offset + m->size - start, size);
> + tsz = (size_t)min_lt(m->offset + m->size - start, size);
> paddr = m->paddr + start - m->offset;
> if (vmcore_remap_oldmem_pfn(vma, vma->vm_start + len,
> paddr >> PAGE_SHIFT, tsz,

Hi, Dave.

Seems the previous parameter is unsigned long long, and the later one is
size_t. The size of both these types doesn't change in running time, why
not use min_t(unsigned long long, a, b) instead?

Thanks
Minfei

2016-03-11 07:48:50

by Dave Young

[permalink] [raw]
Subject: Re: [PATCH 2/2] [PATCH 2/2] proc-vmcore: wrong data type casting fix

Hi, Minfei

On 03/11/16 at 03:19pm, Minfei Huang wrote:
> On 03/11/16 at 02:21pm, [email protected] wrote:
> > @@ -231,7 +231,8 @@ static ssize_t __read_vmcore(char *buffe
> >
> > list_for_each_entry(m, &vmcore_list, list) {
> > if (*fpos < m->offset + m->size) {
> > - tsz = min_t(size_t, m->offset + m->size - *fpos, buflen);
> > + tsz = (size_t)min_lt(m->offset + m->size - *fpos,
> > + buflen);
> > start = m->paddr + *fpos - m->offset;
> > tmp = read_from_oldmem(buffer, tsz, &start, userbuf);
> > if (tmp < 0)
> > @@ -461,7 +462,7 @@ static int mmap_vmcore(struct file *file
> > if (start < m->offset + m->size) {
> > u64 paddr = 0;
> >
> > - tsz = min_t(size_t, m->offset + m->size - start, size);
> > + tsz = (size_t)min_lt(m->offset + m->size - start, size);
> > paddr = m->paddr + start - m->offset;
> > if (vmcore_remap_oldmem_pfn(vma, vma->vm_start + len,
> > paddr >> PAGE_SHIFT, tsz,
>
> Hi, Dave.
>
> Seems the previous parameter is unsigned long long, and the later one is
> size_t. The size of both these types doesn't change in running time, why
> not use min_t(unsigned long long, a, b) instead?

Just want a common macro so it can benefit other users so that we can simply
use it without knowning type details and avoid such similar bugs also.

Thanks
Dave