linux-hardening - Re: [PATCH 01/28] mfd: Add Microchip ZL3073x support

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]

Message-ID: <b54e4da8-20a5-4464-a4b7-f4d8f70af989@redhat.com>
Date: Mon, 14 Apr 2025 16:07:10 +0200
From: Ivan Vecera <ivecera@...hat.com>
To: Andy Shevchenko <andy@...nel.org>
Cc: netdev@...r.kernel.org, Michal Schmidt <mschmidt@...hat.com>,
 Vadim Fedorenko <vadim.fedorenko@...ux.dev>,
 Arkadiusz Kubalewski <arkadiusz.kubalewski@...el.com>,
 Jiri Pirko <jiri@...nulli.us>, Rob Herring <robh@...nel.org>,
 Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley
 <conor+dt@...nel.org>, Prathosh Satish <Prathosh.Satish@...rochip.com>,
 Lee Jones <lee@...nel.org>, Kees Cook <kees@...nel.org>,
 Andrew Morton <akpm@...ux-foundation.org>, devicetree@...r.kernel.org,
 linux-kernel@...r.kernel.org, linux-hardening@...r.kernel.org
Subject: Re: [PATCH 01/28] mfd: Add Microchip ZL3073x support

On 14. 04. 25 1:52 odp., Ivan Vecera wrote:
> 
> 
> On 14. 04. 25 1:39 odp., Ivan Vecera wrote:
>>
>>
>> On 14. 04. 25 8:36 dop., Andy Shevchenko wrote:
>>>> What is wrong here?
>>>>
>>>> I have a device that uses 7-bit addresses and have 16 register pages.
>>>> Each pages is from 0x00-0x7f and register 0x7f is used as page selector
>>>> where bits 0-3 select the page.
>>> The problem is that you overlap virtual page over the real one (the 
>>> main one).
>>>
>>> The drivers you mentioned in v2 discussions most likely are also buggy.
>>> As I implied in the above question the developers hardly get the 
>>> regmap ranges
>>> right. It took me quite a while to see the issue, so it's not 
>>> particularly your
>>> fault.
>> Hi Andy,
>>
>> thank you I see the point.
>>
>> Do you mean that the selector register should not be part of the range?
>>
>> If so, does it mean that I have to specify a range for each page? Like 
>> this:
>>
>>      {
>>          /* Page 0 */
>>          .range_min    = 0x000,
>>          .range_max    = 0x07e,
>>          .selector_reg    = ZL3073x_PAGE_SEL,
>>          .selector_mask    = GENMASK(3, 0),
>>          .selector_shift    = 0,
>>          .window_start    = 0,
>>          .window_len    = 0x7e,
>>      },
>>      {
>>          /* Page 1 */
>>          .range_min    = 0x080,
>>          .range_max    = 0x0fe,
>>          .selector_reg    = ZL3073x_PAGE_SEL,
>>          .selector_mask    = GENMASK(3, 0),
>>          .selector_shift    = 0,
>>          .window_start    = 0,
>>          .window_len    = 0x7e,
>>      },
>> ...

No, I will answer by myself... this is non-sense.... window_len has to 
be 0x80. But I probably know what do you mean...

regmap range should not overlap... so I should use something like:

       {
           /* original <0x000-0x77f> with offset of 0x100 to move
              the range outside of <0x00-0x7f> used by real one */
           .range_min = 0x100,
           .range_max = 0x87f,
           .selector_reg = 0x7f,
           .selector_mask = GENMASK(3, 0),
           .selector_shift = 0,
           .window_start = 0,
           .window_len = 0x80,
       },

With this I have to modify the driver to use this 0x100 offset. I mean 
the datasheet says that register BLAH is at 0x201-0x202. So in the 
driver I have to use 0x301-0x302.

Then the _regmap_select_page maps this 0x301 this way:
window_offset = (0x301 - range_min) % window_len;
window_page = (0x301 - range_min) / window len;
thus
window_offset = (0x301 - 0x100) % 0x80 = 0x001
window_page = (0x301 - 0x100) / 0x80 = 4

Long story short, I have to move virtual range outside real address 
range and apply this offset in the driver code.

Is this correct?

Thanks,
Ivan