Re: [PATCH 4.4 00/37] 4.4.110-stable review

[Mike Galbraith]

On Thu, 2018-01-04 at 15:45 -0800, Guenter Roeck wrote:
> 
> The crash part of this problem may be solved with the following patch
> (thanks to Hugh for the hint). There is still another problem, though -
> with this patch applied, the qemu session aborts with "VCPU Shutdown
> request", whatever that means.

The crash part is not fixed by your patch here, w/wo I get this, and it
is PTI, as virgin 109 boots/works with identical everything else.  My
shiny new PTI equipped enterprise 4.4 RT kernels also boot/work fine,
which seems a bit odd.. and not particularly comforting.

[    1.244354] Freeing unused kernel memory: 1192K
[    1.245278] Write protecting the kernel read-only data: 10240k
[    1.247626] Freeing unused kernel memory: 1152K
[    1.251318] Freeing unused kernel memory: 1476K
[    1.253393] init[1]: segfault at ffffffffff5ff100 ip 00007fffb7ffac6e sp 00007fffb7fa07d8 error 5
[    1.254629] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
[    1.254629] 
[    1.256202] CPU: 4 PID: 1 Comm: init Not tainted 4.4.110-rc1-smp #4
[    1.257169] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.0.0-prebuilt.qemu-project.org 04/01/2014
[    1.258563]  0000000000000000 ffffffff8125a9c0 ffffffff817de7c8 ffff880197e83cf0
[    1.260850]  ffffffff8112bb2d ffffffff00000010 ffff880197e83d00 ffff880197e83ca0
[    1.263091]  ffffffff81c3cf30 000000000000000b ffff880197e90010 0000000000000000
[    1.264580] Call Trace:
[    1.265617]  [<ffffffff8125a9c0>] ? dump_stack+0x5c/0x7c
[    1.266671]  [<ffffffff8112bb2d>] ? panic+0xc8/0x20f
[    1.267799]  [<ffffffff81060af0>] ? do_exit+0xa50/0xa50
[    1.268971]  [<ffffffff810618e9>] ? do_group_exit+0x39/0xa0
[    1.270281]  [<ffffffff8106c8a0>] ? get_signal+0x1d0/0x600
[    1.271347]  [<ffffffff810041e3>] ? do_signal+0x23/0x5b0
[    1.272259]  [<ffffffff8106ade9>] ? __send_signal+0x179/0x460
[    1.273235]  [<ffffffff8104b88f>] ? force_sig_info_fault+0x5f/0x70
[    1.274258]  [<ffffffff8104bf6c>] ? __bad_area_nosemaphore+0x1cc/0x200
[    1.275268]  [<ffffffff8105a052>] ? exit_to_usermode_loop+0x54/0x95
[    1.276262]  [<ffffffff81001961>] ? prepare_exit_to_usermode+0x31/0x40
[    1.277266]  [<ffffffff814d9dbe>] ? retint_user+0x8/0x2c
[    1.278274] Dumping ftrace buffer:
[    1.279011]    (ftrace buffer empty)
[    1.279728] Kernel Offset: disabled
[    1.280432] Rebooting in 60 seconds..

virsh # exit
 
> 
> Guenter
> 
> ---
> From: Guenter Roeck <groeck@xxxxxxxxxxxx>
> Date: Thu, 4 Jan 2018 13:41:55 -0800
> Subject: [PATCH 2/2] WIP: kaiser: Set _PAGE_NX only if supported
> 
> Change-Id: Ie6ab566c1d725b24c4b3aa80a47c3ff3a5feddb9
> Signed-off-by: Guenter Roeck <groeck@xxxxxxxxxxxx>
> ---
>  arch/x86/mm/kaiser.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/mm/kaiser.c b/arch/x86/mm/kaiser.c
> index 7d2f7eb6857f..e4706273d4a1 100644
> --- a/arch/x86/mm/kaiser.c
> +++ b/arch/x86/mm/kaiser.c
> @@ -421,7 +421,8 @@ pgd_t kaiser_set_shadow_pgd(pgd_t *pgdp, pgd_t pgd)
>  			 * get out to userspace running on the kernel CR3,
>  			 * userspace will crash instead of running.
>  			 */
> -			pgd.pgd |= _PAGE_NX;
> +			if (__supported_pte_mask & _PAGE_NX)
> +				pgd.pgd |= _PAGE_NX;
>  		}
>  	} else if (!pgd.pgd) {
>  		/*