Date:	Fri, 19 Feb 2016 10:48:30 -0500
From:	Murali Karicheri <m-karicheri2@...com>
To:	Arnd Bergmann <arnd@...db.de>
CC:	<linux-kernel@...r.kernel.org>, <netdev@...r.kernel.org>,
	<davem@...emloft.net>
Subject: Re: [PATCH v1 1/4] net: ti: netcp: restore get/set_pad_info()
 functionality

On 02/19/2016 09:41 AM, Arnd Bergmann wrote:
> On Thursday 18 February 2016 14:46:14 Murali Karicheri wrote:
>> From: Arnd Bergmann <arnd@...db.de>
>>
>> The commit 899077791403 ("netcp: try to reduce type confusion in
>> descriptors") introduces a regression in Kernel 4.5-rc1 and it breaks
>> get/set_pad_info() functionality.
>>
>> The TI NETCP driver uses pad0 and pad1 fields of knav_dma_desc to
>> store DMA/MEM buffer pointer and buffer size respectively. And in both
>> cases for Keystone 2 the pointer type size is 32 bit regardless of
>> LAPE enabled or not, because CONFIG_ARCH_DMA_ADDR_T_64BIT originally
>> is not expected to be defined.
>>
>> 			!LAPE	LPAE
>> sizeof(void*)		32bit	32bit
>> sizeof(dma_addr_t) 	32bit	32bit
>> sizeof(phys_addr_t) 	32bit	64bit
> 
> As this was never relevant or true, I don't think it needs to be
> mentioned here, it just confuses things. Please just assume that
> dma_addr_t can be 64-bit wide, but will only contain 32-bit
> numbers on keystone.
> 

I can remove this from the commit description and re-send.

>> Unfortunately, above commit changed buffer's pointers save/restore
>> code (get/set_pad_info()) and added intermediate conversation to u64
>> which works incorrectly on 32bit Keystone 2 and causes TI NETCP driver
>> crash in RX/TX path due to "Unable to handle kernel NULL pointer"
>> exception. This issue was reported and discussed in [1].
> 
> Have you been able to figure out why it actually broke? I'd still
> like to know.
> 
As Grygorii is out of office until Monday, I would like to step in.
I will take some time today to try review the reverted changes for
failure reason and get back. But as you have agreed in the discussion
at https://www.mail-archive.com/netdev@vger.kernel.org/msg96311.html
Can we fix the regression by applying this patch and rest of the series
if it looks good? If I need to separate this from rest of the series,
let me know and I can take care of that.

>> Hence, fix it by partially reverting above commit and restoring
>> get/set_pad_info() functionality as it was before.
>>
>> [1] https://www.mail-archive.com/netdev@vger.kernel.org/msg95361.html
>> Cc: Wingman Kwok <w-kwok2@...com>
>> Cc: Mugunthan V N <mugunthanvnm@...com>
>> CC: David Laight <David.Laight@...LAB.COM>
>> Reported-by: Franklin S Cooper Jr <fcooper@...com>
>> Signed-off-by: Arnd Bergmann <arnd@...db.de>
>> Signed-off-by: Grygorii Strashko <grygorii.strashko@...com>
>> Signed-off-by: Murali Karicheri <m-karicheri2@...com>
> 
> I don't think I sent this patch with a 'Signed-off-by', did I?
> (I could be misremembering that).
> 

I think you had agreed based on what I read at
https://www.mail-archive.com/netdev@vger.kernel.org/msg96311.html

reproduced below for your convenience.
=============================================================================

> What I could do now is update your/my patch as i mentioned in [1]
> and re-send it at the weekend (with your authorship and my signoff).
> Do you agree?
> 
> 
> [1] https://www.mail-archive.com/netdev@vger.kernel.org/msg95831.html

Yes, let's do that in the meantime. I can also make sure that that
the driver doesn't build on 64-bit, just in case.
=============================================================================

Hope I can keep your sign-off when I re-send this. Please confirm.

Murali

> 	Arnd
> 


-- 
Murali Karicheri
Linux Kernel, Keystone

Powered by blists - more mailing lists