2021-04-21 13:08:48

by Greg KH

Subject: [PATCH 066/190] Revert "bpf: Remove unnecessary assertion on fp_old"

This reverts commit 5bf2fc1f9c88397b125d5ec5f65b1ed9300ba59d.

Commits from @umn.edu addresses have been found to be submitted in "bad
faith" to try to test the kernel community's ability to review "known
malicious" changes. The result of these submissions can be found in a
paper published at the 42nd IEEE Symposium on Security and Privacy
entitled, "Open Source Insecurity: Stealthily Introducing
Vulnerabilities via Hypocrite Commits" written by Qiushi Wu (University
of Minnesota) and Kangjie Lu (University of Minnesota).

Because of this, all submissions from this group must be reverted from
the kernel tree and will need to be re-reviewed again to determine if
they actually are a valid fix. Until that work is complete, remove this
change to ensure that no problems are being introduced into the
codebase.

Cc: Aditya Pakki <[email protected]>
Cc: Daniel Borkmann <[email protected]>
Cc: https
Signed-off-by: Greg Kroah-Hartman <[email protected]>
---
kernel/bpf/core.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index 75244ecb2389..da29211ea5d8 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -230,6 +230,8 @@ struct bpf_prog *bpf_prog_realloc(struct bpf_prog *fp_old, unsigned int size,
struct bpf_prog *fp;
u32 pages;

+ BUG_ON(fp_old == NULL);
+
size = round_up(size, PAGE_SIZE);
pages = size / PAGE_SIZE;
if (pages <= fp_old->pages)
--
2.31.1

2021-04-27 18:00:09

by Greg KH

Subject: Re: [PATCH 066/190] Revert "bpf: Remove unnecessary assertion on fp_old"

On Wed, Apr 21, 2021 at 02:59:01PM +0200, Greg Kroah-Hartman wrote:
> This reverts commit 5bf2fc1f9c88397b125d5ec5f65b1ed9300ba59d.
>
> Commits from @umn.edu addresses have been found to be submitted in "bad
> faith" to try to test the kernel community's ability to review "known
> malicious" changes. The result of these submissions can be found in a
> paper published at the 42nd IEEE Symposium on Security and Privacy
> entitled, "Open Source Insecurity: Stealthily Introducing
> Vulnerabilities via Hypocrite Commits" written by Qiushi Wu (University
> of Minnesota) and Kangjie Lu (University of Minnesota).
>
> Because of this, all submissions from this group must be reverted from
> the kernel tree and will need to be re-reviewed again to determine if
> they actually are a valid fix. Until that work is complete, remove this
> change to ensure that no problems are being introduced into the
> codebase.
>
> Cc: Aditya Pakki <[email protected]>
> Cc: Daniel Borkmann <[email protected]>
> Cc: https
> Signed-off-by: Greg Kroah-Hartman <[email protected]>
> ---
> kernel/bpf/core.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
> index 75244ecb2389..da29211ea5d8 100644
> --- a/kernel/bpf/core.c
> +++ b/kernel/bpf/core.c
> @@ -230,6 +230,8 @@ struct bpf_prog *bpf_prog_realloc(struct bpf_prog *fp_old, unsigned int size,
> struct bpf_prog *fp;
> u32 pages;
>
> + BUG_ON(fp_old == NULL);
> +
> size = round_up(size, PAGE_SIZE);
> pages = size / PAGE_SIZE;
> if (pages <= fp_old->pages)
> --
> 2.31.1
>

The original commit here is correct, I'll drop this revert.

thanks,

greg k-h

2021-04-27 18:07:18

by Alexei Starovoitov

Subject: Re: [PATCH 066/190] Revert "bpf: Remove unnecessary assertion on fp_old"

On Tue, Apr 27, 2021 at 10:59 AM Greg Kroah-Hartman
<[email protected]> wrote:
>
> On Wed, Apr 21, 2021 at 02:59:01PM +0200, Greg Kroah-Hartman wrote:
> > This reverts commit 5bf2fc1f9c88397b125d5ec5f65b1ed9300ba59d.
> >
> > Commits from @umn.edu addresses have been found to be submitted in "bad
> > faith" to try to test the kernel community's ability to review "known
> > malicious" changes. The result of these submissions can be found in a
> > paper published at the 42nd IEEE Symposium on Security and Privacy
> > entitled, "Open Source Insecurity: Stealthily Introducing
> > Vulnerabilities via Hypocrite Commits" written by Qiushi Wu (University
> > of Minnesota) and Kangjie Lu (University of Minnesota).
> >
> > Because of this, all submissions from this group must be reverted from
> > the kernel tree and will need to be re-reviewed again to determine if
> > they actually are a valid fix. Until that work is complete, remove this
> > change to ensure that no problems are being introduced into the
> > codebase.
> >
> > Cc: Aditya Pakki <[email protected]>
> > Cc: Daniel Borkmann <[email protected]>
> > Cc: https
> > Signed-off-by: Greg Kroah-Hartman <[email protected]>
> > ---
> > kernel/bpf/core.c | 2 ++
> > 1 file changed, 2 insertions(+)
> >
> > diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
> > index 75244ecb2389..da29211ea5d8 100644
> > --- a/kernel/bpf/core.c
> > +++ b/kernel/bpf/core.c
> > @@ -230,6 +230,8 @@ struct bpf_prog *bpf_prog_realloc(struct bpf_prog *fp_old, unsigned int size,
> > struct bpf_prog *fp;
> > u32 pages;
> >
> > + BUG_ON(fp_old == NULL);
> > +
> > size = round_up(size, PAGE_SIZE);
> > pages = size / PAGE_SIZE;
> > if (pages <= fp_old->pages)
> > --
> > 2.31.1
> >
>
> The original commit here is correct, I'll drop this revert.

Yes. No need to revert. The original commit removed BUG_ON and it's fine.
Thanks for checking.

2021-04-27 19:29:53

by Daniel Borkmann

Subject: Re: [PATCH 066/190] Revert "bpf: Remove unnecessary assertion on fp_old"

On 4/27/21 7:58 PM, Greg Kroah-Hartman wrote:
> On Wed, Apr 21, 2021 at 02:59:01PM +0200, Greg Kroah-Hartman wrote:
[...]
>> kernel/bpf/core.c | 2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
>> index 75244ecb2389..da29211ea5d8 100644
>> --- a/kernel/bpf/core.c
>> +++ b/kernel/bpf/core.c
>> @@ -230,6 +230,8 @@ struct bpf_prog *bpf_prog_realloc(struct bpf_prog *fp_old, unsigned int size,
>> struct bpf_prog *fp;
>> u32 pages;
>>
>> + BUG_ON(fp_old == NULL);
>> +
>> size = round_up(size, PAGE_SIZE);
>> pages = size / PAGE_SIZE;
>> if (pages <= fp_old->pages)
>
> The original commit here is correct, I'll drop this revert.

Ok, sounds good to me, thanks! (If there would ever be an incorrect fp_old with NULL, then at
latest we crash in fp_old->pages anyway, so the BUG_ON() was not needed.)

> thanks,
>
> greg k-h
>

2021-04-28 05:56:57

by Greg KH

Subject: Re: [PATCH 066/190] Revert "bpf: Remove unnecessary assertion on fp_old"

On Tue, Apr 27, 2021 at 09:28:31PM +0200, Daniel Borkmann wrote:
> On 4/27/21 7:58 PM, Greg Kroah-Hartman wrote:
> > On Wed, Apr 21, 2021 at 02:59:01PM +0200, Greg Kroah-Hartman wrote:
> [...]
> > > kernel/bpf/core.c | 2 ++
> > > 1 file changed, 2 insertions(+)
> > >
> > > diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
> > > index 75244ecb2389..da29211ea5d8 100644
> > > --- a/kernel/bpf/core.c
> > > +++ b/kernel/bpf/core.c
> > > @@ -230,6 +230,8 @@ struct bpf_prog *bpf_prog_realloc(struct bpf_prog *fp_old, unsigned int size,
> > > struct bpf_prog *fp;
> > > u32 pages;
> > > + BUG_ON(fp_old == NULL);
> > > +
> > > size = round_up(size, PAGE_SIZE);
> > > pages = size / PAGE_SIZE;
> > > if (pages <= fp_old->pages)
> >
> > The original commit here is correct, I'll drop this revert.
>
> Ok, sounds good to me, thanks! (If there would ever be an incorrect fp_old with NULL, then at
> latest we crash in fp_old->pages anyway, so the BUG_ON() was not needed.)

Thanks for the re-review, much appreciated.

greg k-h