2019-05-30 04:09:10

by Stephen Rothwell

[permalink] [raw]
Subject: linux-next: manual merge of the userns tree with the arc-current tree

Hi all,

Today's linux-next merge of the userns tree got a conflict in:

arch/arc/mm/fault.c

between commits:

a8c715b4dd73 ("ARC: mm: SIGSEGV userspace trying to access kernel virtual memory")
ea3885229b0f ("ARC: mm: do_page_fault refactor #5: scoot no_context to end")
acc639eca380 ("ARC: mm: do_page_fault refactor #6: error handlers to use same pattern")
0c85612550a4 ("ARC: mm: do_page_fault refactor #7: fold the various error handling")
c5d7f7610d88 ("ARC: mm: do_page_fault refactor #8: release mmap_sem sooner")

from the arc-current tree and commits:

351b6825b3a9 ("signal: Explicitly call force_sig_fault on current")
2e1661d26736 ("signal: Remove the task parameter from force_sig_fault")

from the userns tree.

I fixed it up (see below) and can carry the fix as necessary. This
is now fixed as far as linux-next is concerned, but any non trivial
conflicts should be mentioned to your upstream maintainer when your tree
is submitted for merging. You may also want to consider cooperating
with the maintainer of the conflicting tree to minimise any particularly
complex conflicts.

--
Cheers,
Stephen Rothwell

diff --cc arch/arc/mm/fault.c
index e93ea06c214c,5001f6418e92..000000000000
--- a/arch/arc/mm/fault.c
+++ b/arch/arc/mm/fault.c
@@@ -187,21 -228,14 +187,21 @@@ bad_area
return;
}

- goto no_context;
+ if (fault & VM_FAULT_SIGBUS) {
+ sig = SIGBUS;
+ si_code = BUS_ADRERR;
+ }
+ else {
+ sig = SIGSEGV;
+ }

-do_sigbus:
- up_read(&mm->mmap_sem);
+ tsk->thread.fault_address = address;
- force_sig_fault(sig, si_code, (void __user *)address, tsk);
++ force_sig_fault(sig, si_code, (void __user *)address);
+ return;

- if (!user_mode(regs))
- goto no_context;
+no_context:
+ if (fixup_exception(regs))
+ return;

- tsk->thread.fault_address = address;
- force_sig_fault(SIGBUS, BUS_ADRERR, (void __user *)address);
+ die("Oops", regs, address);
}


Attachments:
(No filename) (499.00 B)
OpenPGP digital signature

2019-05-30 17:13:56

by Vineet Gupta

[permalink] [raw]
Subject: Re: linux-next: manual merge of the userns tree with the arc-current tree

Hi Stephen,

On 5/29/19 8:17 PM, Stephen Rothwell wrote:
> Hi all,
>
> Today's linux-next merge of the userns tree got a conflict in:
>
> arch/arc/mm/fault.c
>
> between commits:
>
> a8c715b4dd73 ("ARC: mm: SIGSEGV userspace trying to access kernel virtual memory")
> ea3885229b0f ("ARC: mm: do_page_fault refactor #5: scoot no_context to end")
> acc639eca380 ("ARC: mm: do_page_fault refactor #6: error handlers to use same pattern")
> 0c85612550a4 ("ARC: mm: do_page_fault refactor #7: fold the various error handling")
> c5d7f7610d88 ("ARC: mm: do_page_fault refactor #8: release mmap_sem sooner")
>
> from the arc-current tree and commits:
>
> 351b6825b3a9 ("signal: Explicitly call force_sig_fault on current")
> 2e1661d26736 ("signal: Remove the task parameter from force_sig_fault")
>
> from the userns tree.
>
> I fixed it up (see below) and can carry the fix as necessary. This
> is now fixed as far as linux-next is concerned, but any non trivial
> conflicts should be mentioned to your upstream maintainer when your tree
> is submitted for merging. You may also want to consider cooperating
> with the maintainer of the conflicting tree to minimise any particularly
> complex conflicts.


Thx for this. Unfortunately I had to force push my for-next due to broken #7 and
#8 above. So you may have to do this once again.

-Vineet

2019-05-31 03:58:46

by Stephen Rothwell

[permalink] [raw]
Subject: Re: linux-next: manual merge of the userns tree with the arc-current tree

Hi Vineet,

On Thu, 30 May 2019 17:11:33 +0000 Vineet Gupta <[email protected]> wrote:
>
> Thx for this. Unfortunately I had to force push my for-next due to broken #7 and
> #8 above. So you may have to do this once again.

Thanks for the heads up, but "git rerere" seems to have still coped, so
its all good.

--
Cheers,
Stephen Rothwell


Attachments:
(No filename) (499.00 B)
OpenPGP digital signature

2019-07-08 23:48:15

by Stephen Rothwell

[permalink] [raw]
Subject: Re: linux-next: manual merge of the userns tree with the arc-current tree

Hi all,

On Thu, 30 May 2019 13:17:21 +1000 Stephen Rothwell <[email protected]> wrote:
>
> Today's linux-next merge of the userns tree got a conflict in:
>
> arch/arc/mm/fault.c
>
> between commits:
>
> a8c715b4dd73 ("ARC: mm: SIGSEGV userspace trying to access kernel virtual memory")
> ea3885229b0f ("ARC: mm: do_page_fault refactor #5: scoot no_context to end")
> acc639eca380 ("ARC: mm: do_page_fault refactor #6: error handlers to use same pattern")
> 0c85612550a4 ("ARC: mm: do_page_fault refactor #7: fold the various error handling")
> c5d7f7610d88 ("ARC: mm: do_page_fault refactor #8: release mmap_sem sooner")
>
> from the arc-current tree and commits:
>
> 351b6825b3a9 ("signal: Explicitly call force_sig_fault on current")
> 2e1661d26736 ("signal: Remove the task parameter from force_sig_fault")
>
> from the userns tree.
>
> I fixed it up (see below) and can carry the fix as necessary. This
> is now fixed as far as linux-next is concerned, but any non trivial
> conflicts should be mentioned to your upstream maintainer when your tree
> is submitted for merging. You may also want to consider cooperating
> with the maintainer of the conflicting tree to minimise any particularly
> complex conflicts.
>
> --
> Cheers,
> Stephen Rothwell
>
> diff --cc arch/arc/mm/fault.c
> index e93ea06c214c,5001f6418e92..000000000000
> --- a/arch/arc/mm/fault.c
> +++ b/arch/arc/mm/fault.c
> @@@ -187,21 -228,14 +187,21 @@@ bad_area
> return;
> }
>
> - goto no_context;
> + if (fault & VM_FAULT_SIGBUS) {
> + sig = SIGBUS;
> + si_code = BUS_ADRERR;
> + }
> + else {
> + sig = SIGSEGV;
> + }
>
> -do_sigbus:
> - up_read(&mm->mmap_sem);
> + tsk->thread.fault_address = address;
> - force_sig_fault(sig, si_code, (void __user *)address, tsk);
> ++ force_sig_fault(sig, si_code, (void __user *)address);
> + return;
>
> - if (!user_mode(regs))
> - goto no_context;
> +no_context:
> + if (fixup_exception(regs))
> + return;
>
> - tsk->thread.fault_address = address;
> - force_sig_fault(SIGBUS, BUS_ADRERR, (void __user *)address);
> + die("Oops", regs, address);
> }

I am still getting this conflict (the commit ids may have changed).
Just a reminder in case you think Linus may need to know.

--
Cheers,
Stephen Rothwell


Attachments:
(No filename) (499.00 B)
OpenPGP digital signature