linux-kernel - Re: [PATCH v8 2/2] lib: checksum: Use aligned accesses for ip_fast_csum and csum_ipv6

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Thu, 15 Feb 2024 13:56:21 -0500
From: John David Anglin <dave.anglin@...l.net>
To: Guenter Roeck <linux@...ck-us.net>, David Laight
 <David.Laight@...LAB.COM>, Charlie Jenkins <charlie@...osinc.com>,
 Palmer Dabbelt <palmer@...belt.com>,
 Andrew Morton <akpm@...ux-foundation.org>, Helge Deller <deller@....de>,
 "James E.J. Bottomley" <James.Bottomley@...senpartnership.com>,
 Parisc List <linux-parisc@...r.kernel.org>, Al Viro <viro@...iv.linux.org.uk>
Cc: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v8 2/2] lib: checksum: Use aligned accesses for
 ip_fast_csum and csum_ipv6_magic tests

On 2024-02-15 1:17 p.m., Guenter Roeck wrote:
> On 2/15/24 09:25, John David Anglin wrote:
> [ ... ]
>>>>> Source:
>>>>>
>>>>> static bool pc_is_kernel_fn(unsigned long pc, void *fn)
>>>>> {
>>>>>         return (unsigned long)dereference_kernel_function_descriptor(fn) == pc;
>>>> This looks wrong to me.  Function descriptors should always be 8-byte aligned.  I think this
>>>> routine should return false if fn isn't 8-byte aligned.
>>>
>>> Below you state "Code entry points only need 4-byte alignment."
>>>
>>> I think that contradicts each other. Also, the calling code is,
>>> for example,
>>>     pc_is_kernel_fn(pc, syscall_exit)
>>>
>>> I fail to see how this can be consolidated if it is ok
>>> that syscall_exit is 4-byte aligned but, at the same time,
>>> must be 8-byte aligned to be considered to be a kernel function.
>> In the above call, syscall_exit is treated as a function pointer. It points to an 8-byte aligned
>> function descriptor.  The descriptor holds the actual address of the function.  It only needs
>> 4-byte alignment.
>>
>> Descriptors need 8-byte alignment for efficiency on 64-bit parisc. The pc and gp are accessed
>> using ldd instructions.
>>
>
> Maybe code such as
>     pc_is_kernel_fn(pc, syscall_exit)
> is wrong because syscall_exit doesn't point to a function descriptor
> but to the actual address. The code and comments in arch/parisc/kernel/unwind.c
It depends on how syscall_exit is declared.    unwind.c lies the type of handle_interruption, etc:

         extern void * const handle_interruption;
         extern void * const ret_from_kernel_thread;
         extern void * const syscall_exit;
         extern void * const intr_return;
         extern void * const _switch_to_ret;
#ifdef CONFIG_IRQSTACKS
         extern void * const _call_on_stack;
#endif /* CONFIG_IRQSTACKS */

This should yield actual addresses.
> is for sure confusing because it talks about not using
> dereference_kernel_function_descriptor() to keep things simple but then calls
> dereference_kernel_function_descriptor() anyway. Maybe it should just be
>     if (pc == syscall_exit)
> instead.
Looks like.
>
> The entire code is really odd anyway.
>
>         ptr = dereference_kernel_function_descriptor(&handle_interruption);
>         if (pc_is_kernel_fn(pc, ptr)) {
>
> and then pc_is_kernel_fn() dereferences it again. Weird.
>
> It looks like commit 8e0ba125c2bf ("parisc/unwind: fix unwinder when
> CONFIG_64BIT is enabled") might have messed this up. No idea how to fix
> it properly, though.
This is Helge's code...  I'll let him fix it.

Dave

-- 
John David Anglin  dave.anglin@...l.net