linux-mtd.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
@ 2022-08-03 15:53 Yifei Liu
  2022-08-21 18:21 ` Joakim Tjernlund
  0 siblings, 1 reply; 12+ messages in thread
From: Yifei Liu @ 2022-08-03 15:53 UTC (permalink / raw)
  Cc: yifeliu, ezk, madkar, David Woodhouse, Richard Weinberger,
	Matthew Wilcox (Oracle),
	Kyeong Yoo, linux-mtd, linux-kernel

Bug description and fix:

1. Write data to a file, say all 1s from offset 0 to 16.

2. Truncate the file to a smaller size, say 8 bytes.

3. Write new bytes (say 2s) from an offset past the original size of the
file, say at offset 20, for 4 bytes.  This is supposed to create a "hole"
in the file, meaning that the bytes from offset 8 (where it was truncated
above) up to the new write at offset 20, should all be 0s (zeros).

4. Flush all caches using "echo 3 > /proc/sys/vm/drop_caches" (or unmount
and remount) the f/s.

5. Check the content of the file.  It is wrong.  The 1s that used to be
between bytes 9 and 16, before the truncation, have REAPPEARED (they should
be 0s).

We wrote a script and helper C program to reproduce the bug
(reproduce_jffs2_write_begin_issue.sh, write_file.c, and Makefile).  We can
make them available to anyone.

The above example is shown when writing a small file within the same first
page.  But the bug happens for larger files, as long as steps 1, 2, and 3
above all happen within the same page.

The problem was traced to the jffs2_write_begin code, where it goes into an
'if' statement intended to handle writes past the current EOF (i.e., writes
that may create a hole).  The code computes a 'pageofs' that is the floor
of the write position (pos), aligned to the page size boundary.  In other
words, 'pageofs' will never be larger than 'pos'.  The code then sets the
internal jffs2_raw_inode->isize to the size of max(current inode size,
pageofs) but that is wrong: the new file size should be the 'pos', which is
larger than both the current inode size and pageofs.

Similarly, the code incorrectly sets the internal jffs2_raw_inode->dsize to
the difference between the pageofs minus current inode size; instead it
should be the current pos minus the current inode size.  Finally,
inode->i_size was also set incorrectly.

The patch below fixes this bug.  The bug was discovered using a new tool
for finding f/s bugs using model checking, called MCFS (Model Checking File
Systems).

Signed-off-by: Yifei Liu <yifeliu@cs.stonybrook.edu>
Signed-off-by: Erez Zadok <ezk@cs.stonybrook.edu>
Signed-off-by: Manish Adkar <madkar@cs.stonybrook.edu>
---
 fs/jffs2/file.c | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
index ba86acbe12d3..0479096b96e4 100644
--- a/fs/jffs2/file.c
+++ b/fs/jffs2/file.c
@@ -137,19 +137,18 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
 	struct jffs2_inode_info *f = JFFS2_INODE_INFO(inode);
 	struct jffs2_sb_info *c = JFFS2_SB_INFO(inode->i_sb);
 	pgoff_t index = pos >> PAGE_SHIFT;
-	uint32_t pageofs = index << PAGE_SHIFT;
 	int ret = 0;
 
 	jffs2_dbg(1, "%s()\n", __func__);
 
-	if (pageofs > inode->i_size) {
-		/* Make new hole frag from old EOF to new page */
+	if (pos > inode->i_size) {
+		/* Make new hole frag from old EOF to new position */
 		struct jffs2_raw_inode ri;
 		struct jffs2_full_dnode *fn;
 		uint32_t alloc_len;
 
-		jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new page\n",
-			  (unsigned int)inode->i_size, pageofs);
+		jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new position\n",
+			  (unsigned int)inode->i_size, (uint32_t)pos);
 
 		ret = jffs2_reserve_space(c, sizeof(ri), &alloc_len,
 					  ALLOC_NORMAL, JFFS2_SUMMARY_INODE_SIZE);
@@ -169,10 +168,10 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
 		ri.mode = cpu_to_jemode(inode->i_mode);
 		ri.uid = cpu_to_je16(i_uid_read(inode));
 		ri.gid = cpu_to_je16(i_gid_read(inode));
-		ri.isize = cpu_to_je32(max((uint32_t)inode->i_size, pageofs));
+		ri.isize = cpu_to_je32((uint32_t)pos);
 		ri.atime = ri.ctime = ri.mtime = cpu_to_je32(JFFS2_NOW());
 		ri.offset = cpu_to_je32(inode->i_size);
-		ri.dsize = cpu_to_je32(pageofs - inode->i_size);
+		ri.dsize = cpu_to_je32((uint32_t)pos - inode->i_size);
 		ri.csize = cpu_to_je32(0);
 		ri.compr = JFFS2_COMPR_ZERO;
 		ri.node_crc = cpu_to_je32(crc32(0, &ri, sizeof(ri)-8));
@@ -202,7 +201,7 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
 			goto out_err;
 		}
 		jffs2_complete_reservation(c);
-		inode->i_size = pageofs;
+		inode->i_size = pos;
 		mutex_unlock(&f->sem);
 	}
 
-- 
2.25.1


______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply related	[flat|nested] 12+ messages in thread

* Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
  2022-08-03 15:53 [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin Yifei Liu
@ 2022-08-21 18:21 ` Joakim Tjernlund
  2022-08-22  6:25   ` Richard Weinberger
  0 siblings, 1 reply; 12+ messages in thread
From: Joakim Tjernlund @ 2022-08-21 18:21 UTC (permalink / raw)
  To: Yifei Liu
  Cc: ezk, madkar, David Woodhouse, Richard Weinberger,
	Matthew Wilcox (Oracle),
	Kyeong Yoo, linux-mtd, linux-kernel

What happened with this patch? Looks like a important fix but I don't see it applied ?

  Jocke
________________________________________
From: linux-mtd <linux-mtd-bounces@lists.infradead.org> on behalf of Yifei Liu <yifeliu@cs.stonybrook.edu>
Sent: 03 August 2022 17:53
Cc: yifeliu@cs.stonybrook.edu; ezk@cs.stonybrook.edu; madkar@cs.stonybrook.edu; David Woodhouse; Richard Weinberger; Matthew Wilcox (Oracle); Kyeong Yoo; linux-mtd@lists.infradead.org; linux-kernel@vger.kernel.org
Subject: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin

Bug description and fix:

1. Write data to a file, say all 1s from offset 0 to 16.

2. Truncate the file to a smaller size, say 8 bytes.

3. Write new bytes (say 2s) from an offset past the original size of the
file, say at offset 20, for 4 bytes.  This is supposed to create a "hole"
in the file, meaning that the bytes from offset 8 (where it was truncated
above) up to the new write at offset 20, should all be 0s (zeros).

4. Flush all caches using "echo 3 > /proc/sys/vm/drop_caches" (or unmount
and remount) the f/s.

5. Check the content of the file.  It is wrong.  The 1s that used to be
between bytes 9 and 16, before the truncation, have REAPPEARED (they should
be 0s).

We wrote a script and helper C program to reproduce the bug
(reproduce_jffs2_write_begin_issue.sh, write_file.c, and Makefile).  We can
make them available to anyone.

The above example is shown when writing a small file within the same first
page.  But the bug happens for larger files, as long as steps 1, 2, and 3
above all happen within the same page.

The problem was traced to the jffs2_write_begin code, where it goes into an
'if' statement intended to handle writes past the current EOF (i.e., writes
that may create a hole).  The code computes a 'pageofs' that is the floor
of the write position (pos), aligned to the page size boundary.  In other
words, 'pageofs' will never be larger than 'pos'.  The code then sets the
internal jffs2_raw_inode->isize to the size of max(current inode size,
pageofs) but that is wrong: the new file size should be the 'pos', which is
larger than both the current inode size and pageofs.

Similarly, the code incorrectly sets the internal jffs2_raw_inode->dsize to
the difference between the pageofs minus current inode size; instead it
should be the current pos minus the current inode size.  Finally,
inode->i_size was also set incorrectly.

The patch below fixes this bug.  The bug was discovered using a new tool
for finding f/s bugs using model checking, called MCFS (Model Checking File
Systems).

Signed-off-by: Yifei Liu <yifeliu@cs.stonybrook.edu>
Signed-off-by: Erez Zadok <ezk@cs.stonybrook.edu>
Signed-off-by: Manish Adkar <madkar@cs.stonybrook.edu>
---
 fs/jffs2/file.c | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
index ba86acbe12d3..0479096b96e4 100644
--- a/fs/jffs2/file.c
+++ b/fs/jffs2/file.c
@@ -137,19 +137,18 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
        struct jffs2_inode_info *f = JFFS2_INODE_INFO(inode);
        struct jffs2_sb_info *c = JFFS2_SB_INFO(inode->i_sb);
        pgoff_t index = pos >> PAGE_SHIFT;
-       uint32_t pageofs = index << PAGE_SHIFT;
        int ret = 0;

        jffs2_dbg(1, "%s()\n", __func__);

-       if (pageofs > inode->i_size) {
-               /* Make new hole frag from old EOF to new page */
+       if (pos > inode->i_size) {
+               /* Make new hole frag from old EOF to new position */
                struct jffs2_raw_inode ri;
                struct jffs2_full_dnode *fn;
                uint32_t alloc_len;

-               jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new page\n",
-                         (unsigned int)inode->i_size, pageofs);
+               jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new position\n",
+                         (unsigned int)inode->i_size, (uint32_t)pos);

                ret = jffs2_reserve_space(c, sizeof(ri), &alloc_len,
                                          ALLOC_NORMAL, JFFS2_SUMMARY_INODE_SIZE);
@@ -169,10 +168,10 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
                ri.mode = cpu_to_jemode(inode->i_mode);
                ri.uid = cpu_to_je16(i_uid_read(inode));
                ri.gid = cpu_to_je16(i_gid_read(inode));
-               ri.isize = cpu_to_je32(max((uint32_t)inode->i_size, pageofs));
+               ri.isize = cpu_to_je32((uint32_t)pos);
                ri.atime = ri.ctime = ri.mtime = cpu_to_je32(JFFS2_NOW());
                ri.offset = cpu_to_je32(inode->i_size);
-               ri.dsize = cpu_to_je32(pageofs - inode->i_size);
+               ri.dsize = cpu_to_je32((uint32_t)pos - inode->i_size);
                ri.csize = cpu_to_je32(0);
                ri.compr = JFFS2_COMPR_ZERO;
                ri.node_crc = cpu_to_je32(crc32(0, &ri, sizeof(ri)-8));
@@ -202,7 +201,7 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
                        goto out_err;
                }
                jffs2_complete_reservation(c);
-               inode->i_size = pageofs;
+               inode->i_size = pos;
                mutex_unlock(&f->sem);
        }

--
2.25.1


______________________________________________________
Linux MTD discussion mailing list
https://nam11.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.infradead.org%2Fmailman%2Flistinfo%2Flinux-mtd%2F&amp;data=05%7C01%7Cjoakim.tjernlund%40infinera.com%7Cb544edff033b48e8098708da7568ad4f%7C285643de5f5b4b03a1530ae2dc8aaf77%7C1%7C0%7C637951389689871127%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&amp;sdata=NLesT9SYnUR8pO%2BDq0YW3hoGUbTdlmHQO3cVtqS6H%2Fo%3D&amp;reserved=0

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply related	[flat|nested] 12+ messages in thread

* Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
  2022-08-21 18:21 ` Joakim Tjernlund
@ 2022-08-22  6:25   ` Richard Weinberger
  2022-09-04 19:39     ` Yifei Liu
  0 siblings, 1 reply; 12+ messages in thread
From: Richard Weinberger @ 2022-08-22  6:25 UTC (permalink / raw)
  To: Joakim Tjernlund
  Cc: Yifei Liu, ezk, madkar, David Woodhouse, Matthew Wilcox,
	Kyeong Yoo, linux-mtd, linux-kernel

----- Ursprüngliche Mail -----
> Von: "Joakim Tjernlund" <Joakim.Tjernlund@infinera.com>
> An: "Yifei Liu" <yifeliu@cs.stonybrook.edu>
> CC: ezk@cs.stonybrook.edu, madkar@cs.stonybrook.edu, "David Woodhouse" <dwmw2@infradead.org>, "richard"
> <richard@nod.at>, "Matthew Wilcox" <willy@infradead.org>, "Kyeong Yoo" <kyeong.yoo@alliedtelesis.co.nz>, "linux-mtd"
> <linux-mtd@lists.infradead.org>, "linux-kernel" <linux-kernel@vger.kernel.org>
> Gesendet: Sonntag, 21. August 2022 20:21:04
> Betreff: Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin

> What happened with this patch? Looks like a important fix but I don't see it
> applied ?

It will be part of the next fixes PR after I had a chance to review it.

Thanks,
//richard

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
  2022-08-22  6:25   ` Richard Weinberger
@ 2022-09-04 19:39     ` Yifei Liu
  2022-09-21 21:30       ` Yifei Liu
  0 siblings, 1 reply; 12+ messages in thread
From: Yifei Liu @ 2022-09-04 19:39 UTC (permalink / raw)
  To: Richard Weinberger
  Cc: Joakim Tjernlund, ezk, madkar, David Woodhouse, Matthew Wilcox,
	Kyeong Yoo, linux-mtd, linux-kernel

A gentle reminder.

Best Regards,
Yifei

Best Regards,
Yifei


On Mon, Aug 22, 2022 at 2:26 AM Richard Weinberger <richard@nod.at> wrote:
>
> ----- Ursprüngliche Mail -----
> > Von: "Joakim Tjernlund" <Joakim.Tjernlund@infinera.com>
> > An: "Yifei Liu" <yifeliu@cs.stonybrook.edu>
> > CC: ezk@cs.stonybrook.edu, madkar@cs.stonybrook.edu, "David Woodhouse" <dwmw2@infradead.org>, "richard"
> > <richard@nod.at>, "Matthew Wilcox" <willy@infradead.org>, "Kyeong Yoo" <kyeong.yoo@alliedtelesis.co.nz>, "linux-mtd"
> > <linux-mtd@lists.infradead.org>, "linux-kernel" <linux-kernel@vger.kernel.org>
> > Gesendet: Sonntag, 21. August 2022 20:21:04
> > Betreff: Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
>
> > What happened with this patch? Looks like a important fix but I don't see it
> > applied ?
>
> It will be part of the next fixes PR after I had a chance to review it.
>
> Thanks,
> //richard

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
  2022-09-04 19:39     ` Yifei Liu
@ 2022-09-21 21:30       ` Yifei Liu
  0 siblings, 0 replies; 12+ messages in thread
From: Yifei Liu @ 2022-09-21 21:30 UTC (permalink / raw)
  To: Richard Weinberger
  Cc: Joakim Tjernlund, ezk, madkar, David Woodhouse, Matthew Wilcox,
	Kyeong Yoo, linux-mtd, linux-kernel, stable, linux-fsdevel

A reminder about a JFFS2 patch submitted months ago.

Best Regards,
Yifei


On Sun, Sep 4, 2022 at 3:39 PM Yifei Liu <yifeliu@cs.stonybrook.edu> wrote:
>
> A gentle reminder.
>
> Best Regards,
> Yifei
>
> Best Regards,
> Yifei
>
>
> On Mon, Aug 22, 2022 at 2:26 AM Richard Weinberger <richard@nod.at> wrote:
> >
> > ----- Ursprüngliche Mail -----
> > > Von: "Joakim Tjernlund" <Joakim.Tjernlund@infinera.com>
> > > An: "Yifei Liu" <yifeliu@cs.stonybrook.edu>
> > > CC: ezk@cs.stonybrook.edu, madkar@cs.stonybrook.edu, "David Woodhouse" <dwmw2@infradead.org>, "richard"
> > > <richard@nod.at>, "Matthew Wilcox" <willy@infradead.org>, "Kyeong Yoo" <kyeong.yoo@alliedtelesis.co.nz>, "linux-mtd"
> > > <linux-mtd@lists.infradead.org>, "linux-kernel" <linux-kernel@vger.kernel.org>
> > > Gesendet: Sonntag, 21. August 2022 20:21:04
> > > Betreff: Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
> >
> > > What happened with this patch? Looks like a important fix but I don't see it
> > > applied ?
> >
> > It will be part of the next fixes PR after I had a chance to review it.
> >
> > Thanks,
> > //richard

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
  2022-07-25 22:08   ` Vanessa Page
@ 2022-07-26  0:11     ` Vanessa Page
  0 siblings, 0 replies; 12+ messages in thread
From: Vanessa Page @ 2022-07-26  0:11 UTC (permalink / raw)
  To: Yifei Liu
  Cc: ezk, madkar, David Woodhouse, Richard Weinberger,
	Matthew Wilcox (Oracle),
	Kyeong Yoo, linux-mtd, linux-kernel

Stop fucking emailing me you suck fuck. You want a password so you can hack my shit you broke motherfucker

People are not THAT stupid. 

> On Jul 25, 2022, at 6:11 PM, Vanessa Page <Vebpe@outlook.com> wrote:
> 
> You lame ass stalking motherfucker. You are committing a crime and I have reported you to the police. 
> 
> Koop😂😂😑😑😑😎😄😄😴😴👀👀💁🙈🙈🙈🙈💁👀👀😑💁😁😳😎😴😑👀😑😄🤣🤣😀😃😑😴😄🤥🤫😥😰🥵🥵😡😡🤬🤯😭😭😟😟🥳🤩🥸🥸🥸🙁😩🥺☹️😣😔😔😔😖😣😢😡🥶😰😰😥😥😶‍🌫️🤬🤬🤬🤯😶‍🌫️🥶😰😰🤫🤫😓🤯🤯🤯
> Ooiggvklpoure🤬🤬🤬🤬🤬🤬🤬🤬🤬🤬🤯🤯🤯🤯🤯🤯🤯🤯🤯🤫😓🤬🤯🤯🤬🤬🙁😒😍😒😒😎🥸🤩🥳😟😖🥺😩😩🤩🥳🥳😟😖😣🙁😩☹️😟😫😤😭🤬🤩🤩🥸😎🙁😩😡😵😪😧🤥🤥😯😵🤮😬🤥😓🤗🤗😶😲🤤🤐🤒🤕😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫🤑😼😽☠️💀💀👻👻👻👹👹👺👽👽💀💀👻💀💀💀☠️☠️☠️☠️
> 
>> On Jul 25, 2022, at 1:04 AM, Vanessa Page <Vebpe@outlook.com> wrote:
>> 
>> 🥰🥰🥰🥰🥰🥰🥰🥰🥰😍👌😍👌😍😍😍😍😍☺️☺️☺️☺️☺️💕💕😚😚😚😚🥰😚😍😍😍😚😚😍☺️☺️☺️😍💕😚🥰🥰😍☺️☺️😚🥰😍☺️😍🥰😍☺️☺️💕🥰🥰😍☺️☺️😊
>> 
>> Sent from my iPhon✌️
>> 
>>>> On Jul 25, 2022, at 1:01 AM, Yifei Liu <yifeliu@cs.stonybrook.edu> wrote:
>>> 
>>> Bug description and fix:
>>> 
>>> 1. Write data to a file, say all 1s from offset 0 to 16.
>>> 
>>> 2. Truncate the file to a smaller size, say 8 bytes.
>>> 
>>> 3. Write new bytes (say 2s) from an offset past the original size of the
>>> file, say at offset 20, for 4 bytes.  This is supposed to create a "hole"
>>> in the file, meaning that the bytes from offset 8 (where it was truncated
>>> above) up to the new write at offset 20, should all be 0s (zeros).
>>> 
>>> 4. flush all caches using "echo 3 > /proc/sys/vm/drop_caches" (or unmount
>>> and remount) the f/s.
>>> 
>>> 5. Check the content of the file.  It is wrong.  The 1s that used to be
>>> between bytes 9 and 16, before the truncation, have REAPPEARED (they should
>>> be 0s).
>>> 
>>> We wrote a script and helper C program to reproduce the bug
>>> (reproduce_jffs2_write_begin_issue.sh, write_file.c, and Makefile).  We can
>>> make them available to anyone.
>>> 
>>> The above example is shown when writing a small file within the same first
>>> page.  But the bug happens for larger files, as long as steps 1, 2, and 3
>>> above all happen within the same page.
>>> 
>>> The problem was traced to the jffs2_write_begin code, where it goes into an
>>> 'if' statement intended to handle writes past the current EOF (i.e., writes
>>> that may create a hole).  The code computes a 'pageofs' that is the floor
>>> of the write position (pos), aligned to the page size boundary.  In other
>>> words, 'pageofs' will never be larger than 'pos'.  The code then sets the
>>> internal jffs2_raw_inode->isize to the size of max(current inode size,
>>> pageofs) but that is wrong: the new file size should be the 'pos', which is
>>> larger than both the current inode size and pageofs.
>>> 
>>> Similarly, the code incorrectly sets the internal jffs2_raw_inode->dsize to
>>> the difference between the pageofs minus current inode size; instead it
>>> should be the current pos minus the current inode size.  Finally,
>>> inode->i_size was also set incorrectly.
>>> 
>>> The patch below fixes this bug.  The bug was discovered using a new tool
>>> for finding f/s bugs using model checking, called MCFS (Model Checking File
>>> Systems).
>>> 
>>> Signed-off-by: Yifei Liu <yifeliu@cs.stonybrook.edu>
>>> Signed-off-by: Erez Zadok <ezk@cs.stonybrook.edu>
>>> Signed-off-by: Manish Adkar <madkar@cs.stonybrook.edu>
>>> ---
>>> fs/jffs2/file.c | 15 +++++++--------
>>> 1 file changed, 7 insertions(+), 8 deletions(-)
>>> 
>>> diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
>>> index ba86acbe12d3..0479096b96e4 100644
>>> --- a/fs/jffs2/file.c
>>> +++ b/fs/jffs2/file.c
>>> @@ -137,19 +137,18 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>>  struct jffs2_inode_info *f = JFFS2_INODE_INFO(inode);
>>>  struct jffs2_sb_info *c = JFFS2_SB_INFO(inode->i_sb);
>>>  pgoff_t index = pos >> PAGE_SHIFT;
>>> -    uint32_t pageofs = index << PAGE_SHIFT;
>>>  int ret = 0;
>>> 
>>>  jffs2_dbg(1, "%s()\n", __func__);
>>> 
>>> -    if (pageofs > inode->i_size) {
>>> -        /* Make new hole frag from old EOF to new page */
>>> +    if (pos > inode->i_size) {
>>> +        /* Make new hole frag from old EOF to new position */
>>>      struct jffs2_raw_inode ri;
>>>      struct jffs2_full_dnode *fn;
>>>      uint32_t alloc_len;
>>> 
>>> -        jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new page\n",
>>> -              (unsigned int)inode->i_size, pageofs);
>>> +        jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new position\n",
>>> +              (unsigned int)inode->i_size, (uint32_t)pos);
>>> 
>>>      ret = jffs2_reserve_space(c, sizeof(ri), &alloc_len,
>>>                    ALLOC_NORMAL, JFFS2_SUMMARY_INODE_SIZE);
>>> @@ -169,10 +168,10 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>>      ri.mode = cpu_to_jemode(inode->i_mode);
>>>      ri.uid = cpu_to_je16(i_uid_read(inode));
>>>      ri.gid = cpu_to_je16(i_gid_read(inode));
>>> -        ri.isize = cpu_to_je32(max((uint32_t)inode->i_size, pageofs));
>>> +        ri.isize = cpu_to_je32((uint32_t)pos);
>>>      ri.atime = ri.ctime = ri.mtime = cpu_to_je32(JFFS2_NOW());
>>>      ri.offset = cpu_to_je32(inode->i_size);
>>> -        ri.dsize = cpu_to_je32(pageofs - inode->i_size);
>>> +        ri.dsize = cpu_to_je32((uint32_t)pos - inode->i_size);
>>>      ri.csize = cpu_to_je32(0);
>>>      ri.compr = JFFS2_COMPR_ZERO;
>>>      ri.node_crc = cpu_to_je32(crc32(0, &ri, sizeof(ri)-8));
>>> @@ -202,7 +201,7 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>>          goto out_err;
>>>      }
>>>      jffs2_complete_reservation(c);
>>> -        inode->i_size = pageofs;
>>> +        inode->i_size = pos;
>>>      mutex_unlock(&f->sem);
>>>  }
>>> 
>>> -- 
>>> 2.25.1
>>> 
>>> 
>>> ______________________________________________________
>>> Linux MTD discussion mailing list
>>> http://lists.infradead.org/mailman/listinfo/linux-mtd/
______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
  2022-07-25  5:03 ` Vanessa Page
  2022-07-25 21:34   ` Vanessa Page
  2022-07-25 22:08   ` Vanessa Page
@ 2022-07-25 22:08   ` Vanessa Page
  2022-07-26  0:11     ` Vanessa Page
  2 siblings, 1 reply; 12+ messages in thread
From: Vanessa Page @ 2022-07-25 22:08 UTC (permalink / raw)
  To: Yifei Liu
  Cc: ezk, madkar, David Woodhouse, Richard Weinberger,
	Matthew Wilcox (Oracle),
	Kyeong Yoo, linux-mtd, linux-kernel

You lame ass stalking motherfucker. You are committing a crime and I have reported you to the police. 

Koop😂😂😑😑😑😎😄😄😴😴👀👀💁🙈🙈🙈🙈💁👀👀😑💁😁😳😎😴😑👀😑😄🤣🤣😀😃😑😴😄🤥🤫😥😰🥵🥵😡😡🤬🤯😭😭😟😟🥳🤩🥸🥸🥸🙁😩🥺☹️😣😔😔😔😖😣😢😡🥶😰😰😥😥😶‍🌫️🤬🤬🤬🤯😶‍🌫️🥶😰😰🤫🤫😓🤯🤯🤯
Ooiggvklpoure🤬🤬🤬🤬🤬🤬🤬🤬🤬🤬🤯🤯🤯🤯🤯🤯🤯🤯🤯🤫😓🤬🤯🤯🤬🤬🙁😒😍😒😒😎🥸🤩🥳😟😖🥺😩😩🤩🥳🥳😟😖😣🙁😩☹️😟😫😤😭🤬🤩🤩🥸😎🙁😩😡😵😪😧🤥🤥😯😵🤮😬🤥😓🤗🤗😶😲🤤🤐🤒🤕😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫🤑😼😽☠️💀💀👻👻👻👹👹👺👽👽💀💀👻💀💀💀☠️☠️☠️☠️

> On Jul 25, 2022, at 1:04 AM, Vanessa Page <Vebpe@outlook.com> wrote:
> 
> 🥰🥰🥰🥰🥰🥰🥰🥰🥰😍👌😍👌😍😍😍😍😍☺️☺️☺️☺️☺️💕💕😚😚😚😚🥰😚😍😍😍😚😚😍☺️☺️☺️😍💕😚🥰🥰😍☺️☺️😚🥰😍☺️😍🥰😍☺️☺️💕🥰🥰😍☺️☺️😊
> 
> Sent from my iPhon✌️
> 
>> On Jul 25, 2022, at 1:01 AM, Yifei Liu <yifeliu@cs.stonybrook.edu> wrote:
>> 
>> Bug description and fix:
>> 
>> 1. Write data to a file, say all 1s from offset 0 to 16.
>> 
>> 2. Truncate the file to a smaller size, say 8 bytes.
>> 
>> 3. Write new bytes (say 2s) from an offset past the original size of the
>> file, say at offset 20, for 4 bytes.  This is supposed to create a "hole"
>> in the file, meaning that the bytes from offset 8 (where it was truncated
>> above) up to the new write at offset 20, should all be 0s (zeros).
>> 
>> 4. flush all caches using "echo 3 > /proc/sys/vm/drop_caches" (or unmount
>> and remount) the f/s.
>> 
>> 5. Check the content of the file.  It is wrong.  The 1s that used to be
>> between bytes 9 and 16, before the truncation, have REAPPEARED (they should
>> be 0s).
>> 
>> We wrote a script and helper C program to reproduce the bug
>> (reproduce_jffs2_write_begin_issue.sh, write_file.c, and Makefile).  We can
>> make them available to anyone.
>> 
>> The above example is shown when writing a small file within the same first
>> page.  But the bug happens for larger files, as long as steps 1, 2, and 3
>> above all happen within the same page.
>> 
>> The problem was traced to the jffs2_write_begin code, where it goes into an
>> 'if' statement intended to handle writes past the current EOF (i.e., writes
>> that may create a hole).  The code computes a 'pageofs' that is the floor
>> of the write position (pos), aligned to the page size boundary.  In other
>> words, 'pageofs' will never be larger than 'pos'.  The code then sets the
>> internal jffs2_raw_inode->isize to the size of max(current inode size,
>> pageofs) but that is wrong: the new file size should be the 'pos', which is
>> larger than both the current inode size and pageofs.
>> 
>> Similarly, the code incorrectly sets the internal jffs2_raw_inode->dsize to
>> the difference between the pageofs minus current inode size; instead it
>> should be the current pos minus the current inode size.  Finally,
>> inode->i_size was also set incorrectly.
>> 
>> The patch below fixes this bug.  The bug was discovered using a new tool
>> for finding f/s bugs using model checking, called MCFS (Model Checking File
>> Systems).
>> 
>> Signed-off-by: Yifei Liu <yifeliu@cs.stonybrook.edu>
>> Signed-off-by: Erez Zadok <ezk@cs.stonybrook.edu>
>> Signed-off-by: Manish Adkar <madkar@cs.stonybrook.edu>
>> ---
>> fs/jffs2/file.c | 15 +++++++--------
>> 1 file changed, 7 insertions(+), 8 deletions(-)
>> 
>> diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
>> index ba86acbe12d3..0479096b96e4 100644
>> --- a/fs/jffs2/file.c
>> +++ b/fs/jffs2/file.c
>> @@ -137,19 +137,18 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>   struct jffs2_inode_info *f = JFFS2_INODE_INFO(inode);
>>   struct jffs2_sb_info *c = JFFS2_SB_INFO(inode->i_sb);
>>   pgoff_t index = pos >> PAGE_SHIFT;
>> -    uint32_t pageofs = index << PAGE_SHIFT;
>>   int ret = 0;
>> 
>>   jffs2_dbg(1, "%s()\n", __func__);
>> 
>> -    if (pageofs > inode->i_size) {
>> -        /* Make new hole frag from old EOF to new page */
>> +    if (pos > inode->i_size) {
>> +        /* Make new hole frag from old EOF to new position */
>>       struct jffs2_raw_inode ri;
>>       struct jffs2_full_dnode *fn;
>>       uint32_t alloc_len;
>> 
>> -        jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new page\n",
>> -              (unsigned int)inode->i_size, pageofs);
>> +        jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new position\n",
>> +              (unsigned int)inode->i_size, (uint32_t)pos);
>> 
>>       ret = jffs2_reserve_space(c, sizeof(ri), &alloc_len,
>>                     ALLOC_NORMAL, JFFS2_SUMMARY_INODE_SIZE);
>> @@ -169,10 +168,10 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>       ri.mode = cpu_to_jemode(inode->i_mode);
>>       ri.uid = cpu_to_je16(i_uid_read(inode));
>>       ri.gid = cpu_to_je16(i_gid_read(inode));
>> -        ri.isize = cpu_to_je32(max((uint32_t)inode->i_size, pageofs));
>> +        ri.isize = cpu_to_je32((uint32_t)pos);
>>       ri.atime = ri.ctime = ri.mtime = cpu_to_je32(JFFS2_NOW());
>>       ri.offset = cpu_to_je32(inode->i_size);
>> -        ri.dsize = cpu_to_je32(pageofs - inode->i_size);
>> +        ri.dsize = cpu_to_je32((uint32_t)pos - inode->i_size);
>>       ri.csize = cpu_to_je32(0);
>>       ri.compr = JFFS2_COMPR_ZERO;
>>       ri.node_crc = cpu_to_je32(crc32(0, &ri, sizeof(ri)-8));
>> @@ -202,7 +201,7 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>           goto out_err;
>>       }
>>       jffs2_complete_reservation(c);
>> -        inode->i_size = pageofs;
>> +        inode->i_size = pos;
>>       mutex_unlock(&f->sem);
>>   }
>> 
>> -- 
>> 2.25.1
>> 
>> 
>> ______________________________________________________
>> Linux MTD discussion mailing list
>> http://lists.infradead.org/mailman/listinfo/linux-mtd/
______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
  2022-07-25  5:03 ` Vanessa Page
  2022-07-25 21:34   ` Vanessa Page
@ 2022-07-25 22:08   ` Vanessa Page
  2022-07-25 22:08   ` Vanessa Page
  2 siblings, 0 replies; 12+ messages in thread
From: Vanessa Page @ 2022-07-25 22:08 UTC (permalink / raw)
  To: Yifei Liu
  Cc: ezk, madkar, David Woodhouse, Richard Weinberger,
	Matthew Wilcox (Oracle),
	Kyeong Yoo, linux-mtd, linux-kernel

Koop😂😂😑😑😑😎😄😄😴😴👀👀💁🙈🙈🙈🙈💁👀👀😑💁😁😳😎😴😑👀😑😄🤣🤣😀😃😑😴😄🤥🤫😥😰🥵🥵😡😡🤬🤯😭😭😟😟🥳🤩🥸🥸🥸🙁😩🥺☹️😣😔😔😔😖😣😢😡🥶😰😰😥😥😶‍🌫️🤬🤬🤬🤯😶‍🌫️🥶😰😰🤫🤫😓🤯🤯🤯
Ooiggvklpoure🤬🤬🤬🤬🤬🤬🤬🤬🤬🤬🤯🤯🤯🤯🤯🤯🤯🤯🤯🤫😓🤬🤯🤯🤬🤬🙁😒😍😒😒😎🥸🤩🥳😟😖🥺😩😩🤩🥳🥳😟😖😣🙁😩☹️😟😫😤😭🤬🤩🤩🥸😎🙁😩😡😵😪😧🤥🤥😯😵🤮😬🤥😓🤗🤗😶😲🤤🤐🤒🤕😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫🤑😼😽☠️💀💀👻👻👻👹👹👺👽👽💀💀👻💀💀💀☠️☠️☠️☠️

Sent from my iPhone

> On Jul 25, 2022, at 1:04 AM, Vanessa Page <Vebpe@outlook.com> wrote:
> 
> 🥰🥰🥰🥰🥰🥰🥰🥰🥰😍👌😍👌😍😍😍😍😍☺️☺️☺️☺️☺️💕💕😚😚😚😚🥰😚😍😍😍😚😚😍☺️☺️☺️😍💕😚🥰🥰😍☺️☺️😚🥰😍☺️😍🥰😍☺️☺️💕🥰🥰😍☺️☺️😊
> 
> Sent from my iPhon✌️
> 
>> On Jul 25, 2022, at 1:01 AM, Yifei Liu <yifeliu@cs.stonybrook.edu> wrote:
>> 
>> Bug description and fix:
>> 
>> 1. Write data to a file, say all 1s from offset 0 to 16.
>> 
>> 2. Truncate the file to a smaller size, say 8 bytes.
>> 
>> 3. Write new bytes (say 2s) from an offset past the original size of the
>> file, say at offset 20, for 4 bytes.  This is supposed to create a "hole"
>> in the file, meaning that the bytes from offset 8 (where it was truncated
>> above) up to the new write at offset 20, should all be 0s (zeros).
>> 
>> 4. flush all caches using "echo 3 > /proc/sys/vm/drop_caches" (or unmount
>> and remount) the f/s.
>> 
>> 5. Check the content of the file.  It is wrong.  The 1s that used to be
>> between bytes 9 and 16, before the truncation, have REAPPEARED (they should
>> be 0s).
>> 
>> We wrote a script and helper C program to reproduce the bug
>> (reproduce_jffs2_write_begin_issue.sh, write_file.c, and Makefile).  We can
>> make them available to anyone.
>> 
>> The above example is shown when writing a small file within the same first
>> page.  But the bug happens for larger files, as long as steps 1, 2, and 3
>> above all happen within the same page.
>> 
>> The problem was traced to the jffs2_write_begin code, where it goes into an
>> 'if' statement intended to handle writes past the current EOF (i.e., writes
>> that may create a hole).  The code computes a 'pageofs' that is the floor
>> of the write position (pos), aligned to the page size boundary.  In other
>> words, 'pageofs' will never be larger than 'pos'.  The code then sets the
>> internal jffs2_raw_inode->isize to the size of max(current inode size,
>> pageofs) but that is wrong: the new file size should be the 'pos', which is
>> larger than both the current inode size and pageofs.
>> 
>> Similarly, the code incorrectly sets the internal jffs2_raw_inode->dsize to
>> the difference between the pageofs minus current inode size; instead it
>> should be the current pos minus the current inode size.  Finally,
>> inode->i_size was also set incorrectly.
>> 
>> The patch below fixes this bug.  The bug was discovered using a new tool
>> for finding f/s bugs using model checking, called MCFS (Model Checking File
>> Systems).
>> 
>> Signed-off-by: Yifei Liu <yifeliu@cs.stonybrook.edu>
>> Signed-off-by: Erez Zadok <ezk@cs.stonybrook.edu>
>> Signed-off-by: Manish Adkar <madkar@cs.stonybrook.edu>
>> ---
>> fs/jffs2/file.c | 15 +++++++--------
>> 1 file changed, 7 insertions(+), 8 deletions(-)
>> 
>> diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
>> index ba86acbe12d3..0479096b96e4 100644
>> --- a/fs/jffs2/file.c
>> +++ b/fs/jffs2/file.c
>> @@ -137,19 +137,18 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>   struct jffs2_inode_info *f = JFFS2_INODE_INFO(inode);
>>   struct jffs2_sb_info *c = JFFS2_SB_INFO(inode->i_sb);
>>   pgoff_t index = pos >> PAGE_SHIFT;
>> -    uint32_t pageofs = index << PAGE_SHIFT;
>>   int ret = 0;
>> 
>>   jffs2_dbg(1, "%s()\n", __func__);
>> 
>> -    if (pageofs > inode->i_size) {
>> -        /* Make new hole frag from old EOF to new page */
>> +    if (pos > inode->i_size) {
>> +        /* Make new hole frag from old EOF to new position */
>>       struct jffs2_raw_inode ri;
>>       struct jffs2_full_dnode *fn;
>>       uint32_t alloc_len;
>> 
>> -        jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new page\n",
>> -              (unsigned int)inode->i_size, pageofs);
>> +        jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new position\n",
>> +              (unsigned int)inode->i_size, (uint32_t)pos);
>> 
>>       ret = jffs2_reserve_space(c, sizeof(ri), &alloc_len,
>>                     ALLOC_NORMAL, JFFS2_SUMMARY_INODE_SIZE);
>> @@ -169,10 +168,10 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>       ri.mode = cpu_to_jemode(inode->i_mode);
>>       ri.uid = cpu_to_je16(i_uid_read(inode));
>>       ri.gid = cpu_to_je16(i_gid_read(inode));
>> -        ri.isize = cpu_to_je32(max((uint32_t)inode->i_size, pageofs));
>> +        ri.isize = cpu_to_je32((uint32_t)pos);
>>       ri.atime = ri.ctime = ri.mtime = cpu_to_je32(JFFS2_NOW());
>>       ri.offset = cpu_to_je32(inode->i_size);
>> -        ri.dsize = cpu_to_je32(pageofs - inode->i_size);
>> +        ri.dsize = cpu_to_je32((uint32_t)pos - inode->i_size);
>>       ri.csize = cpu_to_je32(0);
>>       ri.compr = JFFS2_COMPR_ZERO;
>>       ri.node_crc = cpu_to_je32(crc32(0, &ri, sizeof(ri)-8));
>> @@ -202,7 +201,7 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>           goto out_err;
>>       }
>>       jffs2_complete_reservation(c);
>> -        inode->i_size = pageofs;
>> +        inode->i_size = pos;
>>       mutex_unlock(&f->sem);
>>   }
>> 
>> -- 
>> 2.25.1
>> 
>> 
>> ______________________________________________________
>> Linux MTD discussion mailing list
>> http://lists.infradead.org/mailman/listinfo/linux-mtd/
______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
  2022-07-25 21:34   ` Vanessa Page
@ 2022-07-25 22:07     ` Vanessa Page
  0 siblings, 0 replies; 12+ messages in thread
From: Vanessa Page @ 2022-07-25 22:07 UTC (permalink / raw)
  To: Yifei Liu
  Cc: ezk, madkar, David Woodhouse, Richard Weinberger,
	Matthew Wilcox (Oracle),
	Kyeong Yoo, linux-mtd, linux-kernel

Koop😂😂😑😑😑😎😄😄😴😴👀👀💁🙈🙈🙈🙈💁👀👀😑💁😁😳😎😴😑👀😑😄🤣🤣😀😃😑😴😄🤥🤫😥😰🥵🥵😡😡🤬🤯😭😭😟😟🥳🤩🥸🥸🥸🙁😩🥺☹️😣😔😔😔😖😣😢😡🥶😰😰😥😥😶‍🌫️🤬🤬🤬🤯😶‍🌫️🥶😰😰🤫🤫😓🤯🤯🤯
Ooiggvklpoure🤬🤬🤬🤬🤬🤬🤬🤬🤬🤬🤯🤯🤯🤯🤯🤯🤯🤯🤯🤫😓🤬🤯🤯🤬🤬🙁😒😍😒😒😎🥸🤩🥳😟😖🥺😩😩🤩🥳🥳😟😖😣🙁😩☹️😟😫😤😭🤬🤩🤩🥸😎🙁😩😡😵😪😧🤥🤥😯😵🤮😬🤥😓🤗🤗😶😲🤤🤐🤒🤕😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫🤑😼😽☠️💀💀👻👻👻👹👹👺👽👽💀💀👻💀💀💀☠️☠️☠️☠️

Sent from my iPhone

> On Jul 25, 2022, at 5:37 PM, Vanessa Page <Vebpe@outlook.com> wrote:
> 
> Koop😂😂😑😑😑😎😄😄😴😴👀👀💁🙈🙈🙈🙈💁👀👀😑💁😁😳😎😴😑👀😑😄🤣🤣😀😃😑😴😄🤥🤫😥😰🥵🥵😡😡🤬🤯😭😭😟😟🥳🤩🥸🥸🥸🙁😩🥺☹️😣😔😔😔😖😣😢😡🥶😰😰😥😥😶‍🌫️🤬🤬🤬🤯😶‍🌫️🥶😰😰🤫🤫😓🤯🤯🤯
> Ooiggvklpoure🤬🤬🤬🤬🤬🤬🤬🤬🤬🤬🤯🤯🤯🤯🤯🤯🤯🤯🤯🤫😓🤬🤯🤯🤬🤬🙁😒😍😒😒😎🥸🤩🥳😟😖🥺😩😩🤩🥳🥳😟😖😣🙁😩☹️😟😫😤😭🤬🤩🤩🥸😎🙁😩😡😵😪😧🤥🤥😯😵🤮😬🤥😓🤗🤗😶😲🤤🤐🤒🤕😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫🤑😼😽☠️💀💀👻👻👻👹👹👺👽👽💀💀👻💀💀💀☠️☠️☠️☠️
______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
  2022-07-25  5:03 ` Vanessa Page
@ 2022-07-25 21:34   ` Vanessa Page
  2022-07-25 22:07     ` Vanessa Page
  2022-07-25 22:08   ` Vanessa Page
  2022-07-25 22:08   ` Vanessa Page
  2 siblings, 1 reply; 12+ messages in thread
From: Vanessa Page @ 2022-07-25 21:34 UTC (permalink / raw)
  To: Yifei Liu
  Cc: ezk, madkar, David Woodhouse, Richard Weinberger,
	Matthew Wilcox (Oracle),
	Kyeong Yoo, linux-mtd, linux-kernel

Koop😂😂😑😑😑😎😄😄😴😴👀👀💁🙈🙈🙈🙈💁👀👀😑💁😁😳😎😴😑👀😑😄🤣🤣😀😃😑😴😄🤥🤫😥😰🥵🥵😡😡🤬🤯😭😭😟😟🥳🤩🥸🥸🥸🙁😩🥺☹️😣😔😔😔😖😣😢😡🥶😰😰😥😥😶‍🌫️🤬🤬🤬🤯😶‍🌫️🥶😰😰🤫🤫😓🤯🤯🤯
Ooiggvklpoure🤬🤬🤬🤬🤬🤬🤬🤬🤬🤬🤯🤯🤯🤯🤯🤯🤯🤯🤯🤫😓🤬🤯🤯🤬🤬🙁😒😍😒😒😎🥸🤩🥳😟😖🥺😩😩🤩🥳🥳😟😖😣🙁😩☹️😟😫😤😭🤬🤩🤩🥸😎🙁😩😡😵😪😧🤥🤥😯😵🤮😬🤥😓🤗🤗😶😲🤤🤐🤒🤕😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫😵‍💫🤑😼😽☠️💀💀👻👻👻👹👹👺👽👽💀💀👻💀💀💀☠️☠️☠️☠️

> On Jul 25, 2022, at 1:06 AM, Vanessa Page <Vebpe@outlook.com> wrote:
> 
> 🥰🥰🥰🥰🥰🥰🥰🥰🥰😍👌😍👌😍😍😍😍😍☺️☺️☺️☺️☺️💕💕😚😚😚😚🥰😚😍😍😍😚😚😍☺️☺️☺️😍💕😚🥰🥰😍☺️☺️😚🥰😍☺️😍🥰😍☺️☺️💕🥰🥰😍☺️☺️😊
> 
> Sent from my iPhon✌️
> 
>> On Jul 25, 2022, at 1:01 AM, Yifei Liu <yifeliu@cs.stonybrook.edu> wrote:
>> 
>> Bug description and fix:
>> 
>> 1. Write data to a file, say all 1s from offset 0 to 16.
>> 
>> 2. Truncate the file to a smaller size, say 8 bytes.
>> 
>> 3. Write new bytes (say 2s) from an offset past the original size of the
>> file, say at offset 20, for 4 bytes.  This is supposed to create a "hole"
>> in the file, meaning that the bytes from offset 8 (where it was truncated
>> above) up to the new write at offset 20, should all be 0s (zeros).
>> 
>> 4. flush all caches using "echo 3 > /proc/sys/vm/drop_caches" (or unmount
>> and remount) the f/s.
>> 
>> 5. Check the content of the file.  It is wrong.  The 1s that used to be
>> between bytes 9 and 16, before the truncation, have REAPPEARED (they should
>> be 0s).
>> 
>> We wrote a script and helper C program to reproduce the bug
>> (reproduce_jffs2_write_begin_issue.sh, write_file.c, and Makefile).  We can
>> make them available to anyone.
>> 
>> The above example is shown when writing a small file within the same first
>> page.  But the bug happens for larger files, as long as steps 1, 2, and 3
>> above all happen within the same page.
>> 
>> The problem was traced to the jffs2_write_begin code, where it goes into an
>> 'if' statement intended to handle writes past the current EOF (i.e., writes
>> that may create a hole).  The code computes a 'pageofs' that is the floor
>> of the write position (pos), aligned to the page size boundary.  In other
>> words, 'pageofs' will never be larger than 'pos'.  The code then sets the
>> internal jffs2_raw_inode->isize to the size of max(current inode size,
>> pageofs) but that is wrong: the new file size should be the 'pos', which is
>> larger than both the current inode size and pageofs.
>> 
>> Similarly, the code incorrectly sets the internal jffs2_raw_inode->dsize to
>> the difference between the pageofs minus current inode size; instead it
>> should be the current pos minus the current inode size.  Finally,
>> inode->i_size was also set incorrectly.
>> 
>> The patch below fixes this bug.  The bug was discovered using a new tool
>> for finding f/s bugs using model checking, called MCFS (Model Checking File
>> Systems).
>> 
>> Signed-off-by: Yifei Liu <yifeliu@cs.stonybrook.edu>
>> Signed-off-by: Erez Zadok <ezk@cs.stonybrook.edu>
>> Signed-off-by: Manish Adkar <madkar@cs.stonybrook.edu>
>> ---
>> fs/jffs2/file.c | 15 +++++++--------
>> 1 file changed, 7 insertions(+), 8 deletions(-)
>> 
>> diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
>> index ba86acbe12d3..0479096b96e4 100644
>> --- a/fs/jffs2/file.c
>> +++ b/fs/jffs2/file.c
>> @@ -137,19 +137,18 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>   struct jffs2_inode_info *f = JFFS2_INODE_INFO(inode);
>>   struct jffs2_sb_info *c = JFFS2_SB_INFO(inode->i_sb);
>>   pgoff_t index = pos >> PAGE_SHIFT;
>> -    uint32_t pageofs = index << PAGE_SHIFT;
>>   int ret = 0;
>> 
>>   jffs2_dbg(1, "%s()\n", __func__);
>> 
>> -    if (pageofs > inode->i_size) {
>> -        /* Make new hole frag from old EOF to new page */
>> +    if (pos > inode->i_size) {
>> +        /* Make new hole frag from old EOF to new position */
>>       struct jffs2_raw_inode ri;
>>       struct jffs2_full_dnode *fn;
>>       uint32_t alloc_len;
>> 
>> -        jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new page\n",
>> -              (unsigned int)inode->i_size, pageofs);
>> +        jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new position\n",
>> +              (unsigned int)inode->i_size, (uint32_t)pos);
>> 
>>       ret = jffs2_reserve_space(c, sizeof(ri), &alloc_len,
>>                     ALLOC_NORMAL, JFFS2_SUMMARY_INODE_SIZE);
>> @@ -169,10 +168,10 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>       ri.mode = cpu_to_jemode(inode->i_mode);
>>       ri.uid = cpu_to_je16(i_uid_read(inode));
>>       ri.gid = cpu_to_je16(i_gid_read(inode));
>> -        ri.isize = cpu_to_je32(max((uint32_t)inode->i_size, pageofs));
>> +        ri.isize = cpu_to_je32((uint32_t)pos);
>>       ri.atime = ri.ctime = ri.mtime = cpu_to_je32(JFFS2_NOW());
>>       ri.offset = cpu_to_je32(inode->i_size);
>> -        ri.dsize = cpu_to_je32(pageofs - inode->i_size);
>> +        ri.dsize = cpu_to_je32((uint32_t)pos - inode->i_size);
>>       ri.csize = cpu_to_je32(0);
>>       ri.compr = JFFS2_COMPR_ZERO;
>>       ri.node_crc = cpu_to_je32(crc32(0, &ri, sizeof(ri)-8));
>> @@ -202,7 +201,7 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>>           goto out_err;
>>       }
>>       jffs2_complete_reservation(c);
>> -        inode->i_size = pageofs;
>> +        inode->i_size = pos;
>>       mutex_unlock(&f->sem);
>>   }
>> 
>> -- 
>> 2.25.1
>> 
>> 
>> ______________________________________________________
>> Linux MTD discussion mailing list
>> http://lists.infradead.org/mailman/listinfo/linux-mtd/
______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
  2022-07-25  4:58 Yifei Liu
@ 2022-07-25  5:03 ` Vanessa Page
  2022-07-25 21:34   ` Vanessa Page
                     ` (2 more replies)
  0 siblings, 3 replies; 12+ messages in thread
From: Vanessa Page @ 2022-07-25  5:03 UTC (permalink / raw)
  To: Yifei Liu
  Cc: ezk, madkar, David Woodhouse, Richard Weinberger,
	Matthew Wilcox (Oracle),
	Kyeong Yoo, linux-mtd, linux-kernel

🥰🥰🥰🥰🥰🥰🥰🥰🥰😍👌😍👌😍😍😍😍😍☺️☺️☺️☺️☺️💕💕😚😚😚😚🥰😚😍😍😍😚😚😍☺️☺️☺️😍💕😚🥰🥰😍☺️☺️😚🥰😍☺️😍🥰😍☺️☺️💕🥰🥰😍☺️☺️😊

Sent from my iPhon✌️

> On Jul 25, 2022, at 1:01 AM, Yifei Liu <yifeliu@cs.stonybrook.edu> wrote:
> 
> Bug description and fix:
> 
> 1. Write data to a file, say all 1s from offset 0 to 16.
> 
> 2. Truncate the file to a smaller size, say 8 bytes.
> 
> 3. Write new bytes (say 2s) from an offset past the original size of the
> file, say at offset 20, for 4 bytes.  This is supposed to create a "hole"
> in the file, meaning that the bytes from offset 8 (where it was truncated
> above) up to the new write at offset 20, should all be 0s (zeros).
> 
> 4. flush all caches using "echo 3 > /proc/sys/vm/drop_caches" (or unmount
> and remount) the f/s.
> 
> 5. Check the content of the file.  It is wrong.  The 1s that used to be
> between bytes 9 and 16, before the truncation, have REAPPEARED (they should
> be 0s).
> 
> We wrote a script and helper C program to reproduce the bug
> (reproduce_jffs2_write_begin_issue.sh, write_file.c, and Makefile).  We can
> make them available to anyone.
> 
> The above example is shown when writing a small file within the same first
> page.  But the bug happens for larger files, as long as steps 1, 2, and 3
> above all happen within the same page.
> 
> The problem was traced to the jffs2_write_begin code, where it goes into an
> 'if' statement intended to handle writes past the current EOF (i.e., writes
> that may create a hole).  The code computes a 'pageofs' that is the floor
> of the write position (pos), aligned to the page size boundary.  In other
> words, 'pageofs' will never be larger than 'pos'.  The code then sets the
> internal jffs2_raw_inode->isize to the size of max(current inode size,
> pageofs) but that is wrong: the new file size should be the 'pos', which is
> larger than both the current inode size and pageofs.
> 
> Similarly, the code incorrectly sets the internal jffs2_raw_inode->dsize to
> the difference between the pageofs minus current inode size; instead it
> should be the current pos minus the current inode size.  Finally,
> inode->i_size was also set incorrectly.
> 
> The patch below fixes this bug.  The bug was discovered using a new tool
> for finding f/s bugs using model checking, called MCFS (Model Checking File
> Systems).
> 
> Signed-off-by: Yifei Liu <yifeliu@cs.stonybrook.edu>
> Signed-off-by: Erez Zadok <ezk@cs.stonybrook.edu>
> Signed-off-by: Manish Adkar <madkar@cs.stonybrook.edu>
> ---
> fs/jffs2/file.c | 15 +++++++--------
> 1 file changed, 7 insertions(+), 8 deletions(-)
> 
> diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
> index ba86acbe12d3..0479096b96e4 100644
> --- a/fs/jffs2/file.c
> +++ b/fs/jffs2/file.c
> @@ -137,19 +137,18 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>    struct jffs2_inode_info *f = JFFS2_INODE_INFO(inode);
>    struct jffs2_sb_info *c = JFFS2_SB_INFO(inode->i_sb);
>    pgoff_t index = pos >> PAGE_SHIFT;
> -    uint32_t pageofs = index << PAGE_SHIFT;
>    int ret = 0;
> 
>    jffs2_dbg(1, "%s()\n", __func__);
> 
> -    if (pageofs > inode->i_size) {
> -        /* Make new hole frag from old EOF to new page */
> +    if (pos > inode->i_size) {
> +        /* Make new hole frag from old EOF to new position */
>        struct jffs2_raw_inode ri;
>        struct jffs2_full_dnode *fn;
>        uint32_t alloc_len;
> 
> -        jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new page\n",
> -              (unsigned int)inode->i_size, pageofs);
> +        jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new position\n",
> +              (unsigned int)inode->i_size, (uint32_t)pos);
> 
>        ret = jffs2_reserve_space(c, sizeof(ri), &alloc_len,
>                      ALLOC_NORMAL, JFFS2_SUMMARY_INODE_SIZE);
> @@ -169,10 +168,10 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>        ri.mode = cpu_to_jemode(inode->i_mode);
>        ri.uid = cpu_to_je16(i_uid_read(inode));
>        ri.gid = cpu_to_je16(i_gid_read(inode));
> -        ri.isize = cpu_to_je32(max((uint32_t)inode->i_size, pageofs));
> +        ri.isize = cpu_to_je32((uint32_t)pos);
>        ri.atime = ri.ctime = ri.mtime = cpu_to_je32(JFFS2_NOW());
>        ri.offset = cpu_to_je32(inode->i_size);
> -        ri.dsize = cpu_to_je32(pageofs - inode->i_size);
> +        ri.dsize = cpu_to_je32((uint32_t)pos - inode->i_size);
>        ri.csize = cpu_to_je32(0);
>        ri.compr = JFFS2_COMPR_ZERO;
>        ri.node_crc = cpu_to_je32(crc32(0, &ri, sizeof(ri)-8));
> @@ -202,7 +201,7 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
>            goto out_err;
>        }
>        jffs2_complete_reservation(c);
> -        inode->i_size = pageofs;
> +        inode->i_size = pos;
>        mutex_unlock(&f->sem);
>    }
> 
> -- 
> 2.25.1
> 
> 
> ______________________________________________________
> Linux MTD discussion mailing list
> http://lists.infradead.org/mailman/listinfo/linux-mtd/
______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply	[flat|nested] 12+ messages in thread

* [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin
@ 2022-07-25  4:58 Yifei Liu
  2022-07-25  5:03 ` Vanessa Page
  0 siblings, 1 reply; 12+ messages in thread
From: Yifei Liu @ 2022-07-25  4:58 UTC (permalink / raw)
  Cc: yifeliu, ezk, madkar, David Woodhouse, Richard Weinberger,
	Matthew Wilcox (Oracle),
	Kyeong Yoo, linux-mtd, linux-kernel

Bug description and fix:

1. Write data to a file, say all 1s from offset 0 to 16.

2. Truncate the file to a smaller size, say 8 bytes.

3. Write new bytes (say 2s) from an offset past the original size of the
file, say at offset 20, for 4 bytes.  This is supposed to create a "hole"
in the file, meaning that the bytes from offset 8 (where it was truncated
above) up to the new write at offset 20, should all be 0s (zeros).

4. flush all caches using "echo 3 > /proc/sys/vm/drop_caches" (or unmount
and remount) the f/s.

5. Check the content of the file.  It is wrong.  The 1s that used to be
between bytes 9 and 16, before the truncation, have REAPPEARED (they should
be 0s).

We wrote a script and helper C program to reproduce the bug
(reproduce_jffs2_write_begin_issue.sh, write_file.c, and Makefile).  We can
make them available to anyone.

The above example is shown when writing a small file within the same first
page.  But the bug happens for larger files, as long as steps 1, 2, and 3
above all happen within the same page.

The problem was traced to the jffs2_write_begin code, where it goes into an
'if' statement intended to handle writes past the current EOF (i.e., writes
that may create a hole).  The code computes a 'pageofs' that is the floor
of the write position (pos), aligned to the page size boundary.  In other
words, 'pageofs' will never be larger than 'pos'.  The code then sets the
internal jffs2_raw_inode->isize to the size of max(current inode size,
pageofs) but that is wrong: the new file size should be the 'pos', which is
larger than both the current inode size and pageofs.

Similarly, the code incorrectly sets the internal jffs2_raw_inode->dsize to
the difference between the pageofs minus current inode size; instead it
should be the current pos minus the current inode size.  Finally,
inode->i_size was also set incorrectly.

The patch below fixes this bug.  The bug was discovered using a new tool
for finding f/s bugs using model checking, called MCFS (Model Checking File
Systems).

Signed-off-by: Yifei Liu <yifeliu@cs.stonybrook.edu>
Signed-off-by: Erez Zadok <ezk@cs.stonybrook.edu>
Signed-off-by: Manish Adkar <madkar@cs.stonybrook.edu>
---
 fs/jffs2/file.c | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
index ba86acbe12d3..0479096b96e4 100644
--- a/fs/jffs2/file.c
+++ b/fs/jffs2/file.c
@@ -137,19 +137,18 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
 	struct jffs2_inode_info *f = JFFS2_INODE_INFO(inode);
 	struct jffs2_sb_info *c = JFFS2_SB_INFO(inode->i_sb);
 	pgoff_t index = pos >> PAGE_SHIFT;
-	uint32_t pageofs = index << PAGE_SHIFT;
 	int ret = 0;
 
 	jffs2_dbg(1, "%s()\n", __func__);
 
-	if (pageofs > inode->i_size) {
-		/* Make new hole frag from old EOF to new page */
+	if (pos > inode->i_size) {
+		/* Make new hole frag from old EOF to new position */
 		struct jffs2_raw_inode ri;
 		struct jffs2_full_dnode *fn;
 		uint32_t alloc_len;
 
-		jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new page\n",
-			  (unsigned int)inode->i_size, pageofs);
+		jffs2_dbg(1, "Writing new hole frag 0x%x-0x%x between current EOF and new position\n",
+			  (unsigned int)inode->i_size, (uint32_t)pos);
 
 		ret = jffs2_reserve_space(c, sizeof(ri), &alloc_len,
 					  ALLOC_NORMAL, JFFS2_SUMMARY_INODE_SIZE);
@@ -169,10 +168,10 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
 		ri.mode = cpu_to_jemode(inode->i_mode);
 		ri.uid = cpu_to_je16(i_uid_read(inode));
 		ri.gid = cpu_to_je16(i_gid_read(inode));
-		ri.isize = cpu_to_je32(max((uint32_t)inode->i_size, pageofs));
+		ri.isize = cpu_to_je32((uint32_t)pos);
 		ri.atime = ri.ctime = ri.mtime = cpu_to_je32(JFFS2_NOW());
 		ri.offset = cpu_to_je32(inode->i_size);
-		ri.dsize = cpu_to_je32(pageofs - inode->i_size);
+		ri.dsize = cpu_to_je32((uint32_t)pos - inode->i_size);
 		ri.csize = cpu_to_je32(0);
 		ri.compr = JFFS2_COMPR_ZERO;
 		ri.node_crc = cpu_to_je32(crc32(0, &ri, sizeof(ri)-8));
@@ -202,7 +201,7 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
 			goto out_err;
 		}
 		jffs2_complete_reservation(c);
-		inode->i_size = pageofs;
+		inode->i_size = pos;
 		mutex_unlock(&f->sem);
 	}
 
-- 
2.25.1


______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

^ permalink raw reply related	[flat|nested] 12+ messages in thread

end of thread, other threads:[~2022-09-21 21:30 UTC | newest]

Thread overview: 12+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-08-03 15:53 [PATCH] jffs2: correct logic when creating a hole in jffs2_write_begin Yifei Liu
2022-08-21 18:21 ` Joakim Tjernlund
2022-08-22  6:25   ` Richard Weinberger
2022-09-04 19:39     ` Yifei Liu
2022-09-21 21:30       ` Yifei Liu
  -- strict thread matches above, loose matches on Subject: below --
2022-07-25  4:58 Yifei Liu
2022-07-25  5:03 ` Vanessa Page
2022-07-25 21:34   ` Vanessa Page
2022-07-25 22:07     ` Vanessa Page
2022-07-25 22:08   ` Vanessa Page
2022-07-25 22:08   ` Vanessa Page
2022-07-26  0:11     ` Vanessa Page

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).