linux-kernel - Re: [PATCH 3/8] dmaengine: ste_dma40: Actually write the runtime configuration to registers

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:	Mon, 15 Apr 2013 12:59:06 +0100
From:	Lee Jones <lee.jones@...aro.org>
To:	Rabin Vincent <rabin@....in>
Cc:	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
	arnd@...db.de, linus.walleij@...ricsson.com
Subject: Re: [PATCH 3/8] dmaengine: ste_dma40: Actually write the runtime
 configuration to registers

On Mon, 15 Apr 2013, Rabin Vincent wrote:

> 2013/4/15 Lee Jones <lee.jones@...aro.org>:
> > On Fri, 12 Apr 2013, Rabin Vincent wrote:
> >> 2013/4/9 Lee Jones <lee.jones@...aro.org>:
> >> > Someone has spent a fair amount of effort writing a runtime configuration
> >> > changing algorithm for DMA clients. However, the config appears to never
> >> > actually make it to hardware. In order for the configuration to take hold
> >> > we need to issue a d40_config_write(), as this is the routine which writes
> >> > it into the hardware's registers.
> >>
> >> No, it's not.  This function is only for initial configuration which
> >> should only be written when the channel is allocated.  In fact, by
> >> calling it here in runtime_config, you are introducing a serious bug:
> >> other logical channels on the same physical channel will stop because of
> >> the SSLNK/SDLNK of the physical channel being zeroed.
> >>
> >> The runtime config already makes it the hardware in the existing code,
> >> via d40_*_cfg().
> >
> > Sorry Rabin, but the only place I can see the config being written is
> > in d40_config_write().
> >
> > Can you paste the line of code in d40_*_cfg() which actually writes
> > the config to hardware please? I don't see it.
> 
> It's not that simple.  There are some pointers passed to d40_*_cfg() and
> that function writes the configuration to the variables those pointers
> point to (d40c->log_def.lcsp1, d40c->src_def_cfg, etc.).  Please read
> the code to see how those variables end up being used later when the
> LLIs are prepared for the HW.

I have read the code, which is why I know that the config only gets
written in d40_config_write(). :)

So the configuration which gets set in the runtime_config routine
doesn't ever make it to hardware - hence this patch.

Unless I'm missing something?

-- 
Lee Jones
Linaro ST-Ericsson Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/