netdev - Re: [PATCH] net/macb: fix ISR clear-on-write behavior only for some SoC

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:	Tue, 14 May 2013 13:38:50 +0200
From:	Michal Simek <monstr@...str.eu>
To:	Nicolas Ferre <nicolas.ferre@...el.com>
CC:	Jean-Christophe PLAGNIOL-VILLARD <plagnioj@...osoft.com>,
	michal.simek@...inx.com, s.trumtrar@...gutronix.de,
	hein_tibosch@...oo.es,
	Ludovic Desroches <ludovic.desroches@...el.com>,
	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
	netdev@...r.kernel.org, Anirudha Sarangi <anirudh@...inx.com>
Subject: Re: [PATCH] net/macb: fix ISR clear-on-write behavior only for some
 SoC

On 05/14/2013 11:16 AM, Nicolas Ferre wrote:
> On 13/05/2013 18:05, Jean-Christophe PLAGNIOL-VILLARD :
>>
>> On May 14, 2013, at 12:05 AM, Nicolas Ferre <nicolas.ferre@...el.com> wrote:
>>
>>> Commit 749a2b6 (net/macb: clear tx/rx completion flags in ISR)
>>> introduces clear-on-write on ISR register. This behavior is not always
>>> implemented when using Cadence MACB/GEM and is breaking other platforms.
>>> We are using a new Device Tree compatibility string and a capability
>>> property to actually activate this clear-on-write behavior on ISR.
>>>
>>> Reported-by: Hein Tibosch <hein_tibosch@...oo.es>
>>> Signed-off-by: Nicolas Ferre <nicolas.ferre@...el.com>
>>
>> can we detect it via the IP?
> 
> As said by Hein, we cannot use the IP revision number. *But* we may have the opportunity to read this integration configuration in the Design Configuration Register 1 (DCFG1 already used for determining data bus width).
> 
> So, Michal or Steffen, can you please tell me the value of:
> -> bit 23 at register address 0x280: mine is "1" which should mean "IRQ read clear", yours should be "0".

here is the whole reg map for zynq.
Reg 0x280 is undocumented in our TRM.
Please decode it not sure if bit 0 is LSB or MSB.

U-Boot-PetaLinux> md e000b000
e000b000: 0000001c 000e0013 00000006 00000000    ................
e000b010: 00180704 00000021 3ffba2e4 3ffbd32c    ....!......?,..?
e000b020: 00000003 00000000 00000000 00000000    ................
e000b030: 07ffffff 63c66f08 00000000 0000ffff    .....o.c........
e000b040: 000003ff 000003ff 00000000 00000000    ................
e000b050: 00000000 00000000 00000000 00000000    ................
e000b060: 00000000 00000000 00000000 00000000    ................
e000b070: 00000000 00000000 00000000 00000000    ................
e000b080: 00000000 00000000 00350a00 00001c48    ..........5.H...
e000b090: 00000000 00000000 00000000 00000000    ................
e000b0a0: 00000000 00000000 00000000 00000000    ................
e000b0b0: 00000000 00000000 00000000 00000000    ................
e000b0c0: 00000000 00000000 00000000 00000000    ................
e000b0d0: 00000000 00000000 00000000 00000000    ................
e000b0e0: 00000000 00000000 00000000 00000000    ................
e000b0f0: 00000000 00000000 00000000 00020118    ................
U-Boot-PetaLinux>
e000b100: 00014796 00000000 0000051e 00000001    .G..............
e000b110: 00000000 00000000 0000051d 00000001    ................
e000b120: 00000000 00000000 00000000 00000000    ................
e000b130: 00000000 00000000 00000000 00000000    ................
e000b140: 00000000 00000000 00000000 00000000    ................
e000b150: 001e58b6 00000000 00000526 00000000    .X......&.......
e000b160: 00000004 00000000 00000004 00000001    ................
e000b170: 00000000 00000004 00000000 0000051d    ................
e000b180: 00000000 00000000 00000000 00000000    ................
e000b190: 00000000 00000000 00000000 00000000    ................
e000b1a0: 00000038 00000000 00000000 00000000    8...............
e000b1b0: 00000000 00000000 00000000 00000000    ................
e000b1c0: 00000000 00000000 00000000 00000000    ................
e000b1d0: 00000000 00000000 00000000 00000000    ................
e000b1e0: 00000000 00000000 00000000 00000000    ................
e000b1f0: 00000000 00000000 00000000 00000000    ................
U-Boot-PetaLinux>
e000b200: 00000000 00000000 00000000 00000000    ................
e000b210: 00000000 00000000 00000000 00000000    ................
e000b220: 00000000 00000000 00000000 00000000    ................
e000b230: 00000000 00000000 00000000 00000000    ................
e000b240: 00000000 00000000 00000000 00000000    ................
e000b250: 00000000 00000000 00000000 00000000    ................
e000b260: 00000000 00000000 00000000 00000000    ................
e000b270: 00000000 00000000 00000000 00000000    ................
e000b280: 02500111 2ab13fff 00000000 00000000    ..P..?.*........
e000b290: 002f2145 00000200 00000000 00000000    E!/.............
e000b2a0: 00000000 00000000 00000000 00000000    ................
e000b2b0: 00000000 00000000 00000000 00000000    ................
e000b2c0: 00000000 00000000 00000000 00000000    ................
e000b2d0: 00000000 00000000 00000000 00000000    ................
e000b2e0: 00000000 00000000 00000000 00000000    ................
e000b2f0: 00000000 00000000 00000000 00000000    ................
U-Boot-PetaLinux>



> 
> Hein, in case of use of the MACB, we do not have this register included, so I will avoid to run the test when using MACB (we already have this information).
> 
> If it works, I plan to rewrite the patch but taking this information instead of the device tree compatibility string.

yep. Will be good to detect it instead of new compatible string.

Also is there an option to remove "CONFIG_ARCH_AT91"?

Thanks,
Michal

-- 
Michal Simek, Ing. (M.Eng), OpenPGP -> KeyID: FE3D1F91
w: www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel - Microblaze cpu - http://www.monstr.eu/fdt/
Maintainer of Linux kernel - Xilinx Zynq ARM architecture
Microblaze U-BOOT custodian and responsible for u-boot arm zynq platform



Download attachment "signature.asc" of type "application/pgp-signature" (264 bytes)