Re: [PATCH v2] mfd: mt6370: Add the out-of-bound check to prevent the null pointer

From: ChiYuan Huang
Date: Fri Oct 21 2022 - 05:02:34 EST


Andy Shevchenko <andy.shevchenko@xxxxxxxxx> 於 2022年10月21日 週五 下午4:34寫道:
>
> On Fri, Oct 21, 2022 at 5:17 AM cy_huang <u0084500@xxxxxxxxx> wrote:
> >
> > From: ChiYuan Huang <cy_huang@xxxxxxxxxxx>
> >
> > This potential risk could happen at regmap_raw_read() or
> > regmap_raw_write() when accessing the over-bound register address.
> >
> > For testing, I try to reproduce it with a testing attribute file.
> > Below's the issue trace log.
>
> It looks like you randomly cut the trace.
> It's not what I meant and documentation suggests.
>
I checked the submitting-patch.rst.
To satisfy the requirement for 70-75 chars per line, I only keep the
important log.

May I ask what you mean for the 'trim'?
Is it 'Still keep 70-75 per line and cut the characters that's over
the limit to the next line"?

> > [41.314358] pc : i2c_smbus_xfer+0x58/0x120
> > [41.314371] lr : i2c_smbus_read_i2c_block_data+0x74/0xc0
> > [41.399677] Call trace:
> > [41.402153] i2c_smbus_xfer+0x58/0x120
> > [41.405956] i2c_smbus_read_i2c_block_data+0x74/0xc0
> > [41.410991] mt6370_regmap_read+0x40/0x60 [mt6370]
> > [41.415855] _regmap_raw_read+0xe4/0x278
> > [41.419834] regmap_raw_read+0xec/0x240
> > [41.423721] rg_bound_show+0xb0/0x120 [mt6370]
> > [41.428226] dev_attr_show+0x3c/0x80
> > [41.431851] sysfs_kf_seq_show+0xc4/0x150
> > [41.435916] kernfs_seq_show+0x48/0x60
> > [41.439718] seq_read_iter+0x11c/0x450
> > [41.443519] kernfs_fop_read_iter+0x124/0x1c0
> > [41.447937] vfs_read+0x1a8/0x288
> > [41.451296] ksys_read+0x74/0x100
> > [41.454654] __arm64_sys_read+0x24/0x30
> > [41.458541] invoke_syscall+0x54/0x118
> > [41.462344] el0_svc_common.constprop.4+0x94/0x128
> > [41.467202] do_el0_svc+0x3c/0xd0
> > [41.470562] el0_svc+0x20/0x60
> > [41.473658] el0t_64_sync_handler+0x94/0xb8
> > [41.477899] el0t_64_sync+0x15c/0x160
> > [41.481614] Code: 54000388 f9401262 aa1303e0 52800041 (f9400042)
> > [41.487793] ---[ end trace 0000000000000000 ]---
>
> ...
>
> > If there's still something improper, please kindly correct me.
>
> https://www.kernel.org/doc/html/latest/process/submitting-patches.html#backtraces-in-commit-messages
>
> --
> With Best Regards,
> Andy Shevchenko