Re: Potential info leak: Kernel pointer leak?

[Vinod Koul]

On 22-01-19, 21:25, Fuqian Huang wrote:
> Hi, recently I came across some code and it seems to be able to leak
> kernel address?
> Is the following code cause info leak in the Linux kernel?
> The callback function address is printed to debugfs.
> The local user could know the kernel object address, and is able to
> bypass kASLR.
> linux-4.14.90
> drivers/dma/qcom/hidma_dbg.c:46
> function - hidma_ll_chstats

Doesnt %p not print kernel addresses anymore, see
Documentation/core-api/printk-formats.rst

> 
> The hidma_ll_chstats function in drivers/dma/qcom/hidma_dbg.c in the
> Linux kernel 4.14.90 allows local users to obtain sensitive address
> information by reading "callback=" lines in a debugfs file.
> 
> Similar to CVE-2018-7754

-- 
~Vinod