Re: [PATCH] fbdev: defio: fix the pagelist corruption

[Paul Menzel <pmenzel@molgen.mpg.de>]

Dear Chuansheng,


Am 31.03.22 um 02:06 schrieb Liu, Chuansheng:

>> -----Original Message-----
>> From: Paul Menzel <pmenzel@molgen.mpg.de>
>> Sent: Thursday, March 31, 2022 12:47 AM

[…]

>> Am 29.03.22 um 01:58 schrieb Liu, Chuansheng:
>>
>>>> -----Original Message-----
>>>> From: Paul Menzel
>>>> Sent: Monday, March 28, 2022 2:15 PM
>>
>>>> Am 28.03.22 um 02:58 schrieb Liu, Chuansheng:
>>>>
>>>>>> -----Original Message-----
>>>>
>>>>>> Sent: Saturday, March 26, 2022 4:11 PM
>>>>
>>>>>> Am 17.03.22 um 06:46 schrieb Chuansheng Liu:
>>>>>>> Easily hit the below list corruption:
>>>>>>> ==
>>>>>>> list_add corruption. prev->next should be next (ffffffffc0ceb090), but
>>>>>>> was ffffec604507edc8. (prev=ffffec604507edc8).
>>>>>>> WARNING: CPU: 65 PID: 3959 at lib/list_debug.c:26
>>>>>>> __list_add_valid+0x53/0x80
>>>>>>> CPU: 65 PID: 3959 Comm: fbdev Tainted: G     U
>>>>>>> RIP: 0010:__list_add_valid+0x53/0x80
>>>>>>> Call Trace:
>>>>>>>      <TASK>
>>>>>>>      fb_deferred_io_mkwrite+0xea/0x150
>>>>>>>      do_page_mkwrite+0x57/0xc0
>>>>>>>      do_wp_page+0x278/0x2f0
>>>>>>>      __handle_mm_fault+0xdc2/0x1590
>>>>>>>      handle_mm_fault+0xdd/0x2c0
>>>>>>>      do_user_addr_fault+0x1d3/0x650
>>>>>>>      exc_page_fault+0x77/0x180
>>>>>>>      ? asm_exc_page_fault+0x8/0x30
>>>>>>>      asm_exc_page_fault+0x1e/0x30
>>>>>>> RIP: 0033:0x7fd98fc8fad1
>>>>>>> ==
>>>>>>>
>>>>>>> Figure out the race happens when one process is adding &page->lru into
>>>>>>> the pagelist tail in fb_deferred_io_mkwrite(), another process is
>>>>>>> re-initializing the same &page->lru in fb_deferred_io_fault(), which is
>>>>>>> not protected by the lock.
>>>>>>>
>>>>>>> This fix is to init all the page lists one time during initialization,
>>>>>>> it not only fixes the list corruption, but also avoids INIT_LIST_HEAD()
>>>>>>> redundantly.
>>>>>>>
>>>>>>> Fixes: 105a940416fc ("fbdev/defio: Early-out if page is already enlisted")
>>>>>>> Cc: Thomas Zimmermann <tzimmermann@suse.de>
>>>>>>> Signed-off-by: Chuansheng Liu <chuansheng.liu@intel.com>
>>>>>>> ---
>>>>>>>      drivers/video/fbdev/core/fb_defio.c | 9 ++++++++-
>>>>>>>      1 file changed, 8 insertions(+), 1 deletion(-)
>>>>>>>
>>>>>>> diff --git a/drivers/video/fbdev/core/fb_defio.c b/drivers/video/fbdev/core/fb_defio.c
>>>>>>> index 98b0f23bf5e2..eafb66ca4f28 100644
>>>>>>> --- a/drivers/video/fbdev/core/fb_defio.c
>>>>>>> +++ b/drivers/video/fbdev/core/fb_defio.c
>>>>>>> @@ -59,7 +59,6 @@ static vm_fault_t fb_deferred_io_fault(struct vm_fault *vmf)
>>>>>>>      		printk(KERN_ERR "no mapping available\n");
>>>>>>>
>>>>>>>      	BUG_ON(!page->mapping);
>>>>>>> -	INIT_LIST_HEAD(&page->lru);
>>>>>>>      	page->index = vmf->pgoff;
>>>>>>>
>>>>>>>      	vmf->page = page;
>>>>>>> @@ -220,6 +219,8 @@ static void fb_deferred_io_work(struct work_struct *work)
>>>>>>>      void fb_deferred_io_init(struct fb_info *info)
>>>>>>>      {
>>>>>>>      	struct fb_deferred_io *fbdefio = info->fbdefio;
>>>>>>> +	struct page *page;
>>>>>>> +	int i;
>>>>>>>
>>>>>>>      	BUG_ON(!fbdefio);
>>>>>>>      	mutex_init(&fbdefio->lock);
>>>>>>> @@ -227,6 +228,12 @@ void fb_deferred_io_init(struct fb_info *info)
>>>>>>>      	INIT_LIST_HEAD(&fbdefio->pagelist);
>>>>>>>      	if (fbdefio->delay == 0) /* set a default of 1 s */
>>>>>>>      		fbdefio->delay = HZ;
>>>>>>> +
>>>>>>> +	/* initialize all the page lists one time */
>>>>>>> +	for (i = 0; i < info->fix.smem_len; i += PAGE_SIZE) {
>>>>>>> +		page = fb_deferred_io_page(info, i);
>>>>>>> +		INIT_LIST_HEAD(&page->lru);
>>>>>>> +	}
>>>>>>>      }
>>>>>>>      EXPORT_SYMBOL_GPL(fb_deferred_io_init);
>>>>>>>
>>>>>> Applying your patch on top of current Linus’ master branch, tty0 is
>>>>>> unusable and looks frozen. Sometimes network card still works, sometimes
>>>>>> not.
>>>>>
>>>>> I don't see how the patch would cause below BUG call stack, need some time to
>>>>> debug. Just few comments:
>>>>> 1. Will the system work well without this patch?
>>>>
>>>> Yes, the framebuffer works well without the patch.
>>>>
>>>>> 2. When you are sure the patch causes the regression you saw, please get free
>>>> to submit one reverted patch, thanks : )
>>>>
>>>> I think you for patch wasn’t submitted yet – at least not pulled by Linus.
>>> The patch has been in drm-tip, could you have a try with the latest drm-tip to see
>>> if the Framebuffer works well, in that case, we could revert it in drm-tip then.
>>
>> With drm-tip (drm-tip: 2022y-03m-29d-13h-14m-35s UTC integration
>> manifest) everything works fine. (I had to disable amdgpu driver, as it
>> failed to build.) Is anyone able to explain that?
> 
> My patch is for fixing another patch which is in the drm-tip at least,

The referenced commit 105a940416fc in the Fixes tag is also in Linus’ 
master branch.

> so I assume applying my patch into Linus tree directly is not
> completely proper. That's my intention of asking your help for
> retesting drm-tip.
If there were such a relation, that would need to be documented in the 
commit message.

> You mean everything working fine means another issue you hit is also
> gone?
No, I just mean the hang when applying your patch.

Anyway, after figuring out, that drm-tip, is actually not behind Linus’ 
master branch, I tried to figure out the differences, and it turns out 
it’s also related to commit fac54e2bfb5b (x86/Kconfig: Select 
HAVE_ARCH_HUGE_VMALLOC with HAVE_ARCH_HUGE_VMAP) [1], which is in Linus’ 
master branch, but not drm-tip. Note, I am using a 32-bit user space and 
a 64-bit Linux kernel. Reverting commit fac54e2bfb5b, and having your 
patch a applied, the hang is gone.

I am adding the people involved in the other discussion to make them 
aware of this failure case.


Kind regards,

Paul


[1]: https://linux-regtracking.leemhuis.info/regzbot/mainline/