Revert "bpf: program: Refuse non-O_RDWR flags in BPF_OBJ_GET"
Subsystems (3)
PATCH 1/1 Revert "bpf: program: Refuse non-O_RDWR flags in BPF_OBJ_GET"
From: Maciej Żenczykowski <[email protected]>
This reverts commit d37300ed182131f1757895a62e556332857417e5.
This breaks Android userspace which expects to be able to fetch programs with just read permissions.
--- kernel/bpf/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/bpf/inode.c +1 -1
@@ -543,7 +543,7 @@ int bpf_obj_get_user(const char __user *pathname, int flags)return PTR_ERR (raw);if (type ==BPF_TYPE_PROG ) - ret = (f_flags !=O_RDWR ) ? -EINVAL :bpf_prog_new_fd (raw); + ret =bpf_prog_new_fd (raw);else if (type ==BPF_TYPE_MAP ) ret =bpf_map_new_fd (raw, f_flags);else if (type ==BPF_TYPE_LINK )
Signature · 1 line
2.32.0.288.g62a8d224e6-goog
Quoted 5 lines · Maciej Żenczykowski
This reverts commit
d37300ed182131f1757895a62e556332857417e5.This breaks Android userspace which expects to be able to fetch programs with just read permissions.
Sorry about this! I'll defer to the maintainers what to do here. Reverting leaves us with a gaping hole for access control of pinned programs.
Signature · 4 lines
Lorenz Bauer | Systems Engineer 6th Floor, County Hall/The Riverside Building, SE1 7PB, UK www.cloudflare.com
Quoted 4 lines · Lorenz Bauer
Quoted 1 line · Maciej Żenczykowski
Quoted 8 lines
Quoted 4 lines
This reverts commit
d37300ed182131f1757895a62e556332857417e5.This breaks Android userspace which expects to be able to fetch programs with just read permissions.
Sorry about this! I'll defer to the maintainers what to do here. Reverting leaves us with a gaping hole for access control of pinned programs.
Not sure what hole you're referring to. Could you provide more details/explanation?
It seems perfectly reasonable to be able to get a program with just read privs. After all, you're not modifying it, just using it.
AFAIK there is no way to modify a program after it was loaded, has this changed? if so, the checks should be on the modifications not the fd fetch.
I guess one could argue fetching with write only privs doesn't make sense?
Anyway... userspace is broken... so revert is the answer.
In Android the process loading/pinning bpf maps/programs is a different process (the 'bpfloader') to the users (which are far less privileged)
A message that is not in the archive (sent off-list or not yet received): CANP3RGdrpb+KiD+a29zTSU3LKR8Qo6aFdo4QseRvPdNhZ_AOJw@mail.gmail.com
Quoted 11 lines · Maciej Żenczykowski
Quoted 9 lines · Lorenz Bauer
Quoted 5 lines · Maciej Żenczykowski
This reverts commit
d37300ed182131f1757895a62e556332857417e5.This breaks Android userspace which expects to be able to fetch programs with just read permissions.
Quoted 8 lines
Quoted 3 lines
Sorry about this! I'll defer to the maintainers what to do here. Reverting leaves us with a gaping hole for access control of pinned programs.
Not sure what hole you're referring to. Could you provide more details/explanation?
It seems perfectly reasonable to be able to get a program with just read privs. After all, you're not modifying it, just using it.
Agreed, if that was what the kernel is doing. What you get with BPF_F_RDONLY is a fully read-write fd, since the rest of the BPF subsystem doesn't check program fd flags. Hence my fix to only allow O_RDWR, which matches what the kernel actually does. Otherwise any user with read-only access can get a R/W fd.
Quoted 1 line
AFAIK there is no way to modify a program after it was loaded, has this changed?
You can't modify the program, but you can detach it, for example. Any program related bpf command that takes a program fd basically.
Quoted 1 line
if so, the checks should be on the modifications not the fd fetch.
True, unfortunately that code doesn't exist. It's also not straightforward to write and probably impossible to backport.
Quoted 6 lines
I guess one could argue fetching with write only privs doesn't make sense?
Anyway... userspace is broken... so revert is the answer.
In Android the process loading/pinning bpf maps/programs is a different process (the 'bpfloader') to the users (which are far less privileged)
If the revert happens you need to make sure that all of your pinned state is only readable by the bpfloader user. And everybody else, realistically.
Signature · 4 lines
Lorenz Bauer | Systems Engineer 6th Floor, County Hall/The Riverside Building, SE1 7PB, UK www.cloudflare.com
The patch this reverts breaks (Android) userspace, and I've even pointed out the specific code in question that it breaks. What happened to the policy of not breaking userspace?
Why are 'Changes Requested' (see
https://patchwork.kernel.org/project/netdevbpf/patch/[email protected]/
)
by whom? What changes? What do you expect me to do?
Why should I even care? Why should it even be me? I'm not the one that broke things.
Quoted 29 lines · Lorenz Bauer
Quoted 11 lines · Maciej Żenczykowski
Quoted 9 lines · Lorenz Bauer
Quoted 5 lines · Maciej Żenczykowski
This reverts commit
d37300ed182131f1757895a62e556332857417e5.This breaks Android userspace which expects to be able to fetch programs with just read permissions.
Quoted 8 lines
Quoted 3 lines
Sorry about this! I'll defer to the maintainers what to do here. Reverting leaves us with a gaping hole for access control of pinned programs.
Not sure what hole you're referring to. Could you provide more details/explanation?
It seems perfectly reasonable to be able to get a program with just read privs. After all, you're not modifying it, just using it.
Agreed, if that was what the kernel is doing. What you get with
BPF_F_RDONLYis a fully read-write fd, since the rest of the BPF subsystem doesn't check program fd flags. Hence my fix to only allowO_RDWR, which matches what the kernel actually does. Otherwise any user with read-only access can get a R/W fd.
Quoted 4 lines
Quoted 1 line
AFAIK there is no way to modify a program after it was loaded, has this changed?
You can't modify the program, but you can detach it, for example. Any program related bpf command that takes a program fd basically.
I fail to see how this is a problem, since it's not modifying the program, why should it need a rdwr file descriptor to do so?
AFAIK in many cases, you don't even need the bpf file descriptor at all (for example
you can detach a tc bpf program via removing the tc filter or tc qdisc or the network interface).
Do you perhaps mean to say you can unpin it? But, if so, then that's a problem in the unpin code...
Or it's even entirely unrelated, since deleting files does not need read or write access to the file, just the folder the file is in.
[maze@zeus ~]$ touch foo; chmod a-rwx foo; sudo chown root:root foo; ls -al foo ----------. 1 root root 0 Jun 21 14:26 foo
[maze@zeus ~]$ rm -f foo; ls -al foo
ls: cannot access 'foo': No such file or directory
Quoted 4 lines
Quoted 1 line
if so, the checks should be on the modifications not the fd fetch.
True, unfortunately that code doesn't exist. It's also not straightforward to write and probably impossible to backport.
Now you're suggesting you expect this broken patch (that I'm trying to revert) to make it into older LTS releases and break things out in the field???
Quoted 10 lines
Quoted 6 lines
I guess one could argue fetching with write only privs doesn't make sense?
Anyway... userspace is broken... so revert is the answer.
In Android the process loading/pinning bpf maps/programs is a different process (the 'bpfloader') to the users (which are far less privileged)
If the revert happens you need to make sure that all of your pinned state is only readable by the bpfloader user. And everybody else, realistically.
On Android selinux prevents anyone from doing untoward things, since they can't get the fds in the first place.
I *want* less privileged users (that can't load bpf programs, but have the selinux privs to get pinned program fds) to be able to attach/detach them (via xdp, tc or to cgroups). That's how stuff works right *now*.
Could I perhaps redesign the system to work around this? I don't know. Perhaps. Perhaps not.
I haven't given it that much thought - I'm still trying to fix (workaround) an hrtimer ncm performance regression that was introduced in 5.10.~24 LTS (there at least we can argue the old code was buggy, and the hrtimer implementation on the hardware in question outright terrible and super slow).
I'm guessing changes would be needed to the progRetrieve() function, and the tests, and the bpfloader, and the permissions embedded in the programs themselves, and possibly the iptables binary (and the netutils wrapper) since it uses xt_bpf, and possibly to binaries privileges and/or selinux capabilities/policies. Or maybe not. I simply don't know.
Well tested patches to make things work are welcome at the aosp project. A word of warning: I haven't checked / thought things through, but it may take 5+ years to roll them out.
At this point it's far too late to make such changes in Android 12/S.
Here's another example of fetching programs with BPF_F_RDONLY in iptables (yes, it was added by me, due to our use in Android):
https://android.googlesource.com/platform/external/iptables.git/+/refs/heads/master/extensions/libxt_bpf.c#64
The fact that this snuck into the 5.12 final release is not relevant (ie. this is a regression in 5.12 vs 5.11, and it is still broken in 5.13-rcX).
Please revert immediately. I've got better things to do. I shouldn't have to be thinking about this or arguing about this.
It already took me significantly more than a day simply to track this down (arguably due to miscommunications with Greg, who'd earlier actually found this in 5.12, but misunderstood the problem, but still...).
Thanks,
Maciej
Quoted 7 lines · Maciej Żenczykowski
Please revert immediately. I've got better things to do. I shouldn't have to be thinking about this or arguing about this.
It already took me significantly more than a day simply to track this down (arguably due to miscommunications with Greg, who'd earlier actually found this in 5.12, but misunderstood the problem, but still...).
You're barking up the wrong tree. I don't object to reverting the patch, you asked me for context and I gave it to you.
Best
Lorenz
Signature · 4 lines
Lorenz Bauer | Systems Engineer 6th Floor, County Hall/The Riverside Building, SE1 7PB, UK www.cloudflare.com
Daniel Borkmann
maintainerQuoted 10 lines · Lorenz Bauer
Quoted 7 lines · Maciej Żenczykowski
Please revert immediately. I've got better things to do. I shouldn't have to be thinking about this or arguing about this.
It already took me significantly more than a day simply to track this down (arguably due to miscommunications with Greg, who'd earlier actually found this in 5.12, but misunderstood the problem, but still...).
Quoted 2 lines
You're barking up the wrong tree. I don't object to reverting the patch, you asked me for context and I gave it to you.
+1, this kind of barking was unnecessary and inappropriate.
I revamped the commit message a bit to have some more context for future reference when we need to get back on this.
Anyway, applied.
Quoted 9 lines
Quoted 2 lines
You're barking up the wrong tree. I don't object to reverting the patch, you asked me for context and I gave it to you.
+1, this kind of barking was unnecessary and inappropriate.
I revamped the commit message a bit to have some more context for future reference when we need to get back on this.
Anyway, applied.
Thank you for applying the change.
Sorry for the tone, but work has been particularly hectic these past few quarters, with a constant onslaught of neverending deadlines...
The current kernel patch management system provides virtually no feedback about (rejected) patches.
There is no email when a patch is rejected (and often even when applied) and there is no notification of who did it or why. It's very much like dealing with a faceless robot.
I'm stuck periodically refreshing a browser tab with
https://patchwork.kernel.org/project/netdevbpf/list/ and waiting on patch state to change from new to accepted or 'changes requested'. In this particular case, there was no email feedback
from any maintainer, except for the patch going into 'changes requested' state on patchworks, and the email thread (which only had me and Lorenz participating) also didn't appear to have any reason in it either....
As such, it was extremely unclear what was being asked of me. It felt like I was simply being ignored.
- Maciej
Quoted 9 lines · Maciej Żenczykowski
From: Maciej Żenczykowski <[email protected]>
This reverts commit
d37300ed182131f1757895a62e556332857417e5.This breaks Android userspace which expects to be able to fetch programs with just read permissions.
Quoted 8 lines
Cc Alexei Starovoitov <[email protected]>Cc Andrii Nakryiko <[email protected]>Cc Daniel Borkmann <[email protected]>Cc Greg Kroah-Hartman <[email protected]>Cc Lorenz Bauer <[email protected]>Fixes d37300ed1821 ("bpf: program: Refuse non-O_RDWR flags in BPF_OBJ_GET")Signed-off-by Maciej Żenczykowski <[email protected]>---
Quoted 9 lines · Maciej Żenczykowski
From: Maciej Żenczykowski <[email protected]>
This reverts commit
d37300ed182131f1757895a62e556332857417e5.This breaks Android userspace which expects to be able to fetch programs with just read permissions.
As a follow up, what does Android expect to be able to do with this read only FD?
Lorenz
Signature · 4 lines
Lorenz Bauer | Systems Engineer 6th Floor, County Hall/The Riverside Building, SE1 7PB, UK www.cloudflare.com
Reposting since apparently my reply only went to Lorenz.
Quoted 8 lines · Lorenz Bauer
Quoted 5 lines · Maciej Żenczykowski
From: Maciej Żenczykowski <[email protected]>
This reverts commit
d37300ed182131f1757895a62e556332857417e5.
Quoted 7 lines
Quoted 4 lines
This breaks Android userspace which expects to be able to fetch programs with just read permissions.
As a follow up, what does Android expect to be able to do with this read only FD?
I'm not actually sure of all the use cases, but at a bare minimum:
We use it for iptables xt_bpf, and to attach to cgroup net hooks and tc bpf hooks.
There's also some still incomplete support for xdp.
There's also non-networking stuff like gpu memory tracking and tracepoints that I know very little about - probably something perf related.
So I think the answer is that mostly we expect to be able to attach it to places (iptables/cgroup/tc/xdp/....others...??)