linux-kernel - Re: [PATCH] bus: mhi: core: debugfs: Use correct format specifiers for addresses

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: <20200926052742.GB9302@linux>
Date:   Sat, 26 Sep 2020 10:57:42 +0530
From:   Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
To:     Jeffrey Hugo <jhugo@...eaurora.org>
Cc:     gregkh@...uxfoundation.org, hemantk@...eaurora.org,
        bbhatt@...eaurora.org, linux-arm-msm@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] bus: mhi: core: debugfs: Use correct format specifiers
 for addresses

On Fri, Sep 25, 2020 at 12:01:54PM -0600, Jeffrey Hugo wrote:
> On 9/25/2020 11:16 AM, Manivannan Sadhasivam wrote:
> > For exposing the addresses of read/write pointers and doorbell register,
> > let's use the correct format specifiers. This fixes the following issues
> > generated using W=1 build in ARM32 and reported by Kbuild bot:
> > 
> > All warnings (new ones prefixed by >>):
> > 
> > > > drivers/bus/mhi/core/debugfs.c:75:7: warning: format specifies type 'unsigned long long' but the argument has type 'dma_addr_t' (aka 'unsigned int') [-Wformat]
> >                                mhi_event->db_cfg.db_val);
> >                                ^~~~~~~~~~~~~~~~~~~~~~~~
> >     drivers/bus/mhi/core/debugfs.c:123:7: warning: format specifies type 'unsigned long long' but the argument has type 'dma_addr_t' (aka 'unsigned int') [-Wformat]
> >                                mhi_chan->db_cfg.db_val);
> >                                ^~~~~~~~~~~~~~~~~~~~~~~
> >     2 warnings generated.
> > 
> > drivers/bus/mhi/core/debugfs.c: In function ‘mhi_debugfs_events_show’:
> > drivers/bus/mhi/core/debugfs.c:74:51: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
> >     seq_printf(m, " local rp: 0x%llx db: 0x%pad\n", (u64)ring->rp,
> >                                                     ^
> > drivers/bus/mhi/core/debugfs.c: In function ‘mhi_debugfs_channels_show’:
> > drivers/bus/mhi/core/debugfs.c:122:7: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
> >         (u64)ring->rp, (u64)ring->wp,
> >         ^
> > drivers/bus/mhi/core/debugfs.c:122:22: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
> >         (u64)ring->rp, (u64)ring->wp,
> >                        ^
> > drivers/bus/mhi/core/debugfs.c:121:62: warning: format ‘%llx’ expects argument of type ‘long long unsigned int’, but argument 5 has type ‘dma_addr_t {aka unsigned int}’ [-Wformat=]
> >     seq_printf(m, " local rp: 0x%llx local wp: 0x%llx db: 0x%llx\n",
> >                                                             ~~~^
> >                                                             %x
> > drivers/bus/mhi/core/debugfs.c:123:7:
> >         mhi_chan->db_cfg.db_val);
> > 
> > Reported-by: kernel test robot <lkp@...el.com>
> > Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
> > ---
> > 
> > Greg: This fixes the issue seen while testing the char-misc/char-misc-testing
> > branch.
> > 
> >   drivers/bus/mhi/core/debugfs.c | 10 +++++-----
> >   1 file changed, 5 insertions(+), 5 deletions(-)
> > 
> > diff --git a/drivers/bus/mhi/core/debugfs.c b/drivers/bus/mhi/core/debugfs.c
> > index 53d05a8e168d..2536ff92b76f 100644
> > --- a/drivers/bus/mhi/core/debugfs.c
> > +++ b/drivers/bus/mhi/core/debugfs.c
> > @@ -71,8 +71,8 @@ static int mhi_debugfs_events_show(struct seq_file *m, void *d)
> >   		seq_printf(m, " rp: 0x%llx wp: 0x%llx", er_ctxt->rp,
> >   			   er_ctxt->wp);
> > -		seq_printf(m, " local rp: 0x%llx db: 0x%llx\n", (u64)ring->rp,
> > -			   mhi_event->db_cfg.db_val);
> > +		seq_printf(m, " local rp: 0x%px db: 0x%pad\n", ring->rp,
> > +			   &mhi_event->db_cfg.db_val);
> >   	}
> >   	return 0;
> > @@ -118,9 +118,9 @@ static int mhi_debugfs_channels_show(struct seq_file *m, void *d)
> >   		seq_printf(m, " base: 0x%llx len: 0x%llx wp: 0x%llx",
> >   			   chan_ctxt->rbase, chan_ctxt->rlen, chan_ctxt->wp);
> > -		seq_printf(m, " local rp: 0x%llx local wp: 0x%llx db: 0x%llx\n",
> > -			   (u64)ring->rp, (u64)ring->wp,
> > -			   mhi_chan->db_cfg.db_val);
> > +		seq_printf(m, " local rp: 0x%px local wp: 0x%px db: 0x%pad\n",
> > +			   ring->rp, ring->wp,
> > +			   &mhi_chan->db_cfg.db_val);
> >   	}
> >   	return 0;
> > 
> 
> Documentation/printk-formats.txt seems to point out that %px is "insecure"
> and thus perhaps not preferred.  Are we assuming that debugfs is only
> accessible by root, and thus the %px usage here is effectively the same as
> %pK?
> 

No, this debugfs entry can be read by non-root users also. But the idea here
is to effectively show the addresses to everyone so I don't think we need to
hide it. The term "insecure" applies to kernel log where exposing the address
doesn't make much sense (except for few obvious reasons).

Thanks,
Mani

> -- 
> Jeffrey Hugo
> Qualcomm Technologies, Inc. is a member of the
> Code Aurora Forum, a Linux Foundation Collaborative Project.