2020-09-23 13:57:57

by Vadym Kochan

[permalink] [raw]
Subject: [PATCH v2] nvmem: core: fix possibly memleak when use nvmem_cell_info_to_nvmem_cell()

Fix missing 'kfree_const(cell->name)' when call to
nvmem_cell_info_to_nvmem_cell() in several places:

* after nvmem_cell_info_to_nvmem_cell() failed during
nvmem_add_cells()

* during nvmem_device_cell_{read,write}. This is fixed by simply
re-using info->name instead of duplicating it:

cell->name = info->name

Because cell->name is not used except for error message printing in case
of un-aligned access, the new __nvmem_cell_info_to_nvmem_cell() helper
was introduced.

Fixes: e2a5402ec7c6 ("nvmem: Add nvmem_device based consumer apis.")
Signed-off-by: Vadym Kochan <[email protected]>
---
v2:
* remove not needed 'kfree_const(cell->name)' after nvmem_cell_info_to_nvmem_cell()
failed.

drivers/nvmem/core.c | 35 ++++++++++++++++++++++++++---------
1 file changed, 26 insertions(+), 9 deletions(-)

diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
index 6cd3edb2eaf6..e6d1bc414faf 100644
--- a/drivers/nvmem/core.c
+++ b/drivers/nvmem/core.c
@@ -361,16 +361,15 @@ static void nvmem_cell_add(struct nvmem_cell *cell)
blocking_notifier_call_chain(&nvmem_notifier, NVMEM_CELL_ADD, cell);
}

-static int nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
- const struct nvmem_cell_info *info,
- struct nvmem_cell *cell)
+static int
+__nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
+ const struct nvmem_cell_info *info,
+ struct nvmem_cell *cell)
{
cell->nvmem = nvmem;
cell->offset = info->offset;
cell->bytes = info->bytes;
- cell->name = kstrdup_const(info->name, GFP_KERNEL);
- if (!cell->name)
- return -ENOMEM;
+ cell->name = info->name;

cell->bit_offset = info->bit_offset;
cell->nbits = info->nbits;
@@ -382,13 +381,31 @@ static int nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
if (!IS_ALIGNED(cell->offset, nvmem->stride)) {
dev_err(&nvmem->dev,
"cell %s unaligned to nvmem stride %d\n",
- cell->name, nvmem->stride);
+ cell->name ?: "<unknown>", nvmem->stride);
return -EINVAL;
}

return 0;
}

+static int
+nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
+ const struct nvmem_cell_info *info,
+ struct nvmem_cell *cell)
+{
+ int err;
+
+ err = __nvmem_cell_info_to_nvmem_cell(nvmem, info, cell);
+ if (err)
+ return err;
+
+ cell->name = kstrdup_const(info->name, GFP_KERNEL);
+ if (!cell->name)
+ return -ENOMEM;
+
+ return 0;
+}
+
/**
* nvmem_add_cells() - Add cell information to an nvmem device
*
@@ -1460,7 +1477,7 @@ ssize_t nvmem_device_cell_read(struct nvmem_device *nvmem,
if (!nvmem)
return -EINVAL;

- rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
+ rc = __nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
if (rc)
return rc;

@@ -1490,7 +1507,7 @@ int nvmem_device_cell_write(struct nvmem_device *nvmem,
if (!nvmem)
return -EINVAL;

- rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
+ rc = __nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
if (rc)
return rc;

--
2.17.1


2020-09-23 14:14:51

by Srinivas Kandagatla

[permalink] [raw]
Subject: Re: [PATCH v2] nvmem: core: fix possibly memleak when use nvmem_cell_info_to_nvmem_cell()



On 23/09/2020 14:53, Vadym Kochan wrote:
> Fix missing 'kfree_const(cell->name)' when call to
> nvmem_cell_info_to_nvmem_cell() in several places:
>
> * after nvmem_cell_info_to_nvmem_cell() failed during
> nvmem_add_cells()
>
> * during nvmem_device_cell_{read,write}. This is fixed by simply
> re-using info->name instead of duplicating it:
>
> cell->name = info->name
>
> Because cell->name is not used except for error message printing in case
> of un-aligned access, the new __nvmem_cell_info_to_nvmem_cell() helper
> was introduced.
>
> Fixes: e2a5402ec7c6 ("nvmem: Add nvmem_device based consumer apis.")
> Signed-off-by: Vadym Kochan <[email protected]>
> ---
> v2:
> * remove not needed 'kfree_const(cell->name)' after nvmem_cell_info_to_nvmem_cell()
> failed.
>
> drivers/nvmem/core.c | 35 ++++++++++++++++++++++++++---------
> 1 file changed, 26 insertions(+), 9 deletions(-)



Really :-)


Below change should just fix this the reported issue!
------------------------>cut<---------------------------

diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
index 6cd3edb2eaf6..9fb9112fe75d 100644
--- a/drivers/nvmem/core.c
+++ b/drivers/nvmem/core.c
@@ -383,6 +383,7 @@ static int nvmem_cell_info_to_nvmem_cell(struct
nvmem_device *nvmem,
dev_err(&nvmem->dev,
"cell %s unaligned to nvmem stride %d\n",
cell->name, nvmem->stride);
+ kfree_const(cell->name);
return -EINVAL;
}

------------------------>cut<---------------------------

I don't see a point in the way your patch try to fix this!!


--srini

>
> diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
> index 6cd3edb2eaf6..e6d1bc414faf 100644
> --- a/drivers/nvmem/core.c
> +++ b/drivers/nvmem/core.c
> @@ -361,16 +361,15 @@ static void nvmem_cell_add(struct nvmem_cell *cell)
> blocking_notifier_call_chain(&nvmem_notifier, NVMEM_CELL_ADD, cell);
> }
>
> -static int nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> - const struct nvmem_cell_info *info,
> - struct nvmem_cell *cell)
> +static int
> +__nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> + const struct nvmem_cell_info *info,
> + struct nvmem_cell *cell)
> {
> cell->nvmem = nvmem;
> cell->offset = info->offset;
> cell->bytes = info->bytes;
> - cell->name = kstrdup_const(info->name, GFP_KERNEL);
> - if (!cell->name)
> - return -ENOMEM;
> + cell->name = info->name;
>
> cell->bit_offset = info->bit_offset;
> cell->nbits = info->nbits;
> @@ -382,13 +381,31 @@ static int nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> if (!IS_ALIGNED(cell->offset, nvmem->stride)) {
> dev_err(&nvmem->dev,
> "cell %s unaligned to nvmem stride %d\n",
> - cell->name, nvmem->stride);
> + cell->name ?: "<unknown>", nvmem->stride);
> return -EINVAL;
> }
>
> return 0;
> }
>
> +static int
> +nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> + const struct nvmem_cell_info *info,
> + struct nvmem_cell *cell)
> +{
> + int err;
> +
> + err = __nvmem_cell_info_to_nvmem_cell(nvmem, info, cell);
> + if (err)
> + return err;
> +
> + cell->name = kstrdup_const(info->name, GFP_KERNEL);
> + if (!cell->name)
> + return -ENOMEM;
> +
> + return 0;
> +}
> +
> /**
> * nvmem_add_cells() - Add cell information to an nvmem device
> *
> @@ -1460,7 +1477,7 @@ ssize_t nvmem_device_cell_read(struct nvmem_device *nvmem,
> if (!nvmem)
> return -EINVAL;
>
> - rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> + rc = __nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> if (rc)
> return rc;
>
> @@ -1490,7 +1507,7 @@ int nvmem_device_cell_write(struct nvmem_device *nvmem,
> if (!nvmem)
> return -EINVAL;
>
> - rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> + rc = __nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> if (rc)
> return rc;
>
>

2020-09-23 14:18:08

by Vadym Kochan

[permalink] [raw]
Subject: Re: [PATCH v2] nvmem: core: fix possibly memleak when use nvmem_cell_info_to_nvmem_cell()

On Wed, Sep 23, 2020 at 03:10:36PM +0100, Srinivas Kandagatla wrote:
>
>
> On 23/09/2020 14:53, Vadym Kochan wrote:
> > Fix missing 'kfree_const(cell->name)' when call to
> > nvmem_cell_info_to_nvmem_cell() in several places:
> >
> > * after nvmem_cell_info_to_nvmem_cell() failed during
> > nvmem_add_cells()
> >
> > * during nvmem_device_cell_{read,write}. This is fixed by simply
> > re-using info->name instead of duplicating it:
> >
> > cell->name = info->name
> >
> > Because cell->name is not used except for error message printing in case
> > of un-aligned access, the new __nvmem_cell_info_to_nvmem_cell() helper
> > was introduced.
> >
> > Fixes: e2a5402ec7c6 ("nvmem: Add nvmem_device based consumer apis.")
> > Signed-off-by: Vadym Kochan <[email protected]>
> > ---
> > v2:
> > * remove not needed 'kfree_const(cell->name)' after nvmem_cell_info_to_nvmem_cell()
> > failed.
> >
> > drivers/nvmem/core.c | 35 ++++++++++++++++++++++++++---------
> > 1 file changed, 26 insertions(+), 9 deletions(-)
>
>
>
> Really :-)
>
But what about nvmem_device_cell_{read,write} case ?
In my understanding the cell is allocated on the stack but kstrdup() is
not freed in the end, or I missed something ?

>
> Below change should just fix this the reported issue!
> ------------------------>cut<---------------------------
>
> diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
> index 6cd3edb2eaf6..9fb9112fe75d 100644
> --- a/drivers/nvmem/core.c
> +++ b/drivers/nvmem/core.c
> @@ -383,6 +383,7 @@ static int nvmem_cell_info_to_nvmem_cell(struct
> nvmem_device *nvmem,
> dev_err(&nvmem->dev,
> "cell %s unaligned to nvmem stride %d\n",
> cell->name, nvmem->stride);
> + kfree_const(cell->name);
> return -EINVAL;
> }
>
> ------------------------>cut<---------------------------
>
> I don't see a point in the way your patch try to fix this!!
>
>
> --srini
>
> >
> > diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
> > index 6cd3edb2eaf6..e6d1bc414faf 100644
> > --- a/drivers/nvmem/core.c
> > +++ b/drivers/nvmem/core.c
> > @@ -361,16 +361,15 @@ static void nvmem_cell_add(struct nvmem_cell *cell)
> > blocking_notifier_call_chain(&nvmem_notifier, NVMEM_CELL_ADD, cell);
> > }
> > -static int nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> > - const struct nvmem_cell_info *info,
> > - struct nvmem_cell *cell)
> > +static int
> > +__nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> > + const struct nvmem_cell_info *info,
> > + struct nvmem_cell *cell)
> > {
> > cell->nvmem = nvmem;
> > cell->offset = info->offset;
> > cell->bytes = info->bytes;
> > - cell->name = kstrdup_const(info->name, GFP_KERNEL);
> > - if (!cell->name)
> > - return -ENOMEM;
> > + cell->name = info->name;
> > cell->bit_offset = info->bit_offset;
> > cell->nbits = info->nbits;
> > @@ -382,13 +381,31 @@ static int nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> > if (!IS_ALIGNED(cell->offset, nvmem->stride)) {
> > dev_err(&nvmem->dev,
> > "cell %s unaligned to nvmem stride %d\n",
> > - cell->name, nvmem->stride);
> > + cell->name ?: "<unknown>", nvmem->stride);
> > return -EINVAL;
> > }
> > return 0;
> > }
> > +static int
> > +nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> > + const struct nvmem_cell_info *info,
> > + struct nvmem_cell *cell)
> > +{
> > + int err;
> > +
> > + err = __nvmem_cell_info_to_nvmem_cell(nvmem, info, cell);
> > + if (err)
> > + return err;
> > +
> > + cell->name = kstrdup_const(info->name, GFP_KERNEL);
> > + if (!cell->name)
> > + return -ENOMEM;
> > +
> > + return 0;
> > +}
> > +
> > /**
> > * nvmem_add_cells() - Add cell information to an nvmem device
> > *
> > @@ -1460,7 +1477,7 @@ ssize_t nvmem_device_cell_read(struct nvmem_device *nvmem,
> > if (!nvmem)
> > return -EINVAL;
> > - rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> > + rc = __nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> > if (rc)
> > return rc;
> > @@ -1490,7 +1507,7 @@ int nvmem_device_cell_write(struct nvmem_device *nvmem,
> > if (!nvmem)
> > return -EINVAL;
> > - rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> > + rc = __nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> > if (rc)
> > return rc;
> >

2020-09-23 14:49:53

by Srinivas Kandagatla

[permalink] [raw]
Subject: Re: [PATCH v2] nvmem: core: fix possibly memleak when use nvmem_cell_info_to_nvmem_cell()



On 23/09/2020 15:13, Vadym Kochan wrote:
> On Wed, Sep 23, 2020 at 03:10:36PM +0100, Srinivas Kandagatla wrote:
>>
>>
>> On 23/09/2020 14:53, Vadym Kochan wrote:
>>> Fix missing 'kfree_const(cell->name)' when call to
>>> nvmem_cell_info_to_nvmem_cell() in several places:
>>>
>>> * after nvmem_cell_info_to_nvmem_cell() failed during
>>> nvmem_add_cells()
>>>
>>> * during nvmem_device_cell_{read,write}. This is fixed by simply
>>> re-using info->name instead of duplicating it:
>>>
>>> cell->name = info->name
>>>
>>> Because cell->name is not used except for error message printing in case
>>> of un-aligned access, the new __nvmem_cell_info_to_nvmem_cell() helper
>>> was introduced.
>>>
>>> Fixes: e2a5402ec7c6 ("nvmem: Add nvmem_device based consumer apis.")
>>> Signed-off-by: Vadym Kochan <[email protected]>
>>> ---
>>> v2:
>>> * remove not needed 'kfree_const(cell->name)' after nvmem_cell_info_to_nvmem_cell()
>>> failed.
>>>
>>> drivers/nvmem/core.c | 35 ++++++++++++++++++++++++++---------
>>> 1 file changed, 26 insertions(+), 9 deletions(-)
>>
>>
>>
>> Really :-)
>>
> But what about nvmem_device_cell_{read,write} case ?
> In my understanding the cell is allocated on the stack but kstrdup() is
You are right!

That is the second issue where the caller outside would fail after
successful call to nvmem_cell_info_to_nvmem_cell() .

Probably we cam free it in failure cases!
something like:

------------------------>cut<---------------------------
diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
index 6cd3edb2eaf6..fb1e756adcee 100644
--- a/drivers/nvmem/core.c
+++ b/drivers/nvmem/core.c
@@ -383,6 +383,7 @@ static int nvmem_cell_info_to_nvmem_cell(struct
nvmem_device *nvmem,
dev_err(&nvmem->dev,
"cell %s unaligned to nvmem stride %d\n",
cell->name, nvmem->stride);
+ kfree_const(cell->name);
return -EINVAL;
}

@@ -1465,8 +1466,10 @@ ssize_t nvmem_device_cell_read(struct
nvmem_device *nvmem,
return rc;

rc = __nvmem_cell_read(nvmem, &cell, buf, &len);
- if (rc)
+ if (rc) {
+ kfree_const(cell->name);
return rc;
+ }

return len;
}
@@ -1494,7 +1497,11 @@ int nvmem_device_cell_write(struct nvmem_device
*nvmem,
if (rc)
return rc;

- return nvmem_cell_write(&cell, buf, cell.bytes);
+ rc = nvmem_cell_write(&cell, buf, cell.bytes);
+ if (rc)
+ kfree_const(cell->name);
+
+ return rc;
}
EXPORT_SYMBOL_GPL(nvmem_device_cell_write);
------------------------>cut<---------------------------

--srini

> not freed in the end, or I missed something ?
>
>>
>> Below change should just fix this the reported issue!
>> ------------------------>cut<---------------------------
>>
>> diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
>> index 6cd3edb2eaf6..9fb9112fe75d 100644
>> --- a/drivers/nvmem/core.c
>> +++ b/drivers/nvmem/core.c
>> @@ -383,6 +383,7 @@ static int nvmem_cell_info_to_nvmem_cell(struct
>> nvmem_device *nvmem,
>> dev_err(&nvmem->dev,
>> "cell %s unaligned to nvmem stride %d\n",
>> cell->name, nvmem->stride);
>> + kfree_const(cell->name);
>> return -EINVAL;
>> }
>>
>> ------------------------>cut<---------------------------
>>
>> I don't see a point in the way your patch try to fix this!!
>>
>>
>> --srini
>>
>>>
>>> diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
>>> index 6cd3edb2eaf6..e6d1bc414faf 100644
>>> --- a/drivers/nvmem/core.c
>>> +++ b/drivers/nvmem/core.c
>>> @@ -361,16 +361,15 @@ static void nvmem_cell_add(struct nvmem_cell *cell)
>>> blocking_notifier_call_chain(&nvmem_notifier, NVMEM_CELL_ADD, cell);
>>> }
>>> -static int nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
>>> - const struct nvmem_cell_info *info,
>>> - struct nvmem_cell *cell)
>>> +static int
>>> +__nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
>>> + const struct nvmem_cell_info *info,
>>> + struct nvmem_cell *cell)
>>> {
>>> cell->nvmem = nvmem;
>>> cell->offset = info->offset;
>>> cell->bytes = info->bytes;
>>> - cell->name = kstrdup_const(info->name, GFP_KERNEL);
>>> - if (!cell->name)
>>> - return -ENOMEM;
>>> + cell->name = info->name;
>>> cell->bit_offset = info->bit_offset;
>>> cell->nbits = info->nbits;
>>> @@ -382,13 +381,31 @@ static int nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
>>> if (!IS_ALIGNED(cell->offset, nvmem->stride)) {
>>> dev_err(&nvmem->dev,
>>> "cell %s unaligned to nvmem stride %d\n",
>>> - cell->name, nvmem->stride);
>>> + cell->name ?: "<unknown>", nvmem->stride);
>>> return -EINVAL;
>>> }
>>> return 0;
>>> }
>>> +static int
>>> +nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
>>> + const struct nvmem_cell_info *info,
>>> + struct nvmem_cell *cell)
>>> +{
>>> + int err;
>>> +
>>> + err = __nvmem_cell_info_to_nvmem_cell(nvmem, info, cell);
>>> + if (err)
>>> + return err;
>>> +
>>> + cell->name = kstrdup_const(info->name, GFP_KERNEL);
>>> + if (!cell->name)
>>> + return -ENOMEM;
>>> +
>>> + return 0;
>>> +}
>>> +
>>> /**
>>> * nvmem_add_cells() - Add cell information to an nvmem device
>>> *
>>> @@ -1460,7 +1477,7 @@ ssize_t nvmem_device_cell_read(struct nvmem_device *nvmem,
>>> if (!nvmem)
>>> return -EINVAL;
>>> - rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
>>> + rc = __nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
>>> if (rc)
>>> return rc;
>>> @@ -1490,7 +1507,7 @@ int nvmem_device_cell_write(struct nvmem_device *nvmem,
>>> if (!nvmem)
>>> return -EINVAL;
>>> - rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
>>> + rc = __nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
>>> if (rc)
>>> return rc;
>>>

2020-09-23 14:53:55

by Vadym Kochan

[permalink] [raw]
Subject: Re: [PATCH v2] nvmem: core: fix possibly memleak when use nvmem_cell_info_to_nvmem_cell()

On Wed, Sep 23, 2020 at 03:47:14PM +0100, Srinivas Kandagatla wrote:
>
>
> On 23/09/2020 15:13, Vadym Kochan wrote:
> > On Wed, Sep 23, 2020 at 03:10:36PM +0100, Srinivas Kandagatla wrote:
> > >
> > >
> > > On 23/09/2020 14:53, Vadym Kochan wrote:
> > > > Fix missing 'kfree_const(cell->name)' when call to
> > > > nvmem_cell_info_to_nvmem_cell() in several places:
> > > >
> > > > * after nvmem_cell_info_to_nvmem_cell() failed during
> > > > nvmem_add_cells()
> > > >
> > > > * during nvmem_device_cell_{read,write}. This is fixed by simply
> > > > re-using info->name instead of duplicating it:
> > > >
> > > > cell->name = info->name
> > > >
> > > > Because cell->name is not used except for error message printing in case
> > > > of un-aligned access, the new __nvmem_cell_info_to_nvmem_cell() helper
> > > > was introduced.
> > > >
> > > > Fixes: e2a5402ec7c6 ("nvmem: Add nvmem_device based consumer apis.")
> > > > Signed-off-by: Vadym Kochan <[email protected]>
> > > > ---
> > > > v2:
> > > > * remove not needed 'kfree_const(cell->name)' after nvmem_cell_info_to_nvmem_cell()
> > > > failed.
> > > >
> > > > drivers/nvmem/core.c | 35 ++++++++++++++++++++++++++---------
> > > > 1 file changed, 26 insertions(+), 9 deletions(-)
> > >
> > >
> > >
> > > Really :-)
> > >
> > But what about nvmem_device_cell_{read,write} case ?
> > In my understanding the cell is allocated on the stack but kstrdup() is
> You are right!
>
> That is the second issue where the caller outside would fail after
> successful call to nvmem_cell_info_to_nvmem_cell() .
>
> Probably we cam free it in failure cases!
> something like:
>
> ------------------------>cut<---------------------------
> diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
> index 6cd3edb2eaf6..fb1e756adcee 100644
> --- a/drivers/nvmem/core.c
> +++ b/drivers/nvmem/core.c
> @@ -383,6 +383,7 @@ static int nvmem_cell_info_to_nvmem_cell(struct
> nvmem_device *nvmem,
> dev_err(&nvmem->dev,
> "cell %s unaligned to nvmem stride %d\n",
> cell->name, nvmem->stride);
> + kfree_const(cell->name);
> return -EINVAL;
> }
>
> @@ -1465,8 +1466,10 @@ ssize_t nvmem_device_cell_read(struct nvmem_device
> *nvmem,
> return rc;
>
> rc = __nvmem_cell_read(nvmem, &cell, buf, &len);
> - if (rc)
> + if (rc) {
> + kfree_const(cell->name);
> return rc;
> + }
>
> return len;
> }
> @@ -1494,7 +1497,11 @@ int nvmem_device_cell_write(struct nvmem_device
> *nvmem,
> if (rc)
> return rc;
>
> - return nvmem_cell_write(&cell, buf, cell.bytes);
> + rc = nvmem_cell_write(&cell, buf, cell.bytes);
> + if (rc)
> + kfree_const(cell->name);
> +
> + return rc;
> }
> EXPORT_SYMBOL_GPL(nvmem_device_cell_write);
> ------------------------>cut<---------------------------
>
> --srini
>

But is it really needed to kstrdup(cell->name) for nvmem_device_cell_{read,write} ?
It is used only for log error in case the unaligned access did not
pass the check.

> > not freed in the end, or I missed something ?
> >
> > >
> > > Below change should just fix this the reported issue!
> > > ------------------------>cut<---------------------------
> > >
> > > diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
> > > index 6cd3edb2eaf6..9fb9112fe75d 100644
> > > --- a/drivers/nvmem/core.c
> > > +++ b/drivers/nvmem/core.c
> > > @@ -383,6 +383,7 @@ static int nvmem_cell_info_to_nvmem_cell(struct
> > > nvmem_device *nvmem,
> > > dev_err(&nvmem->dev,
> > > "cell %s unaligned to nvmem stride %d\n",
> > > cell->name, nvmem->stride);
> > > + kfree_const(cell->name);
> > > return -EINVAL;
> > > }
> > >
> > > ------------------------>cut<---------------------------
> > >
> > > I don't see a point in the way your patch try to fix this!!
> > >
> > >
> > > --srini
> > >
> > > >
> > > > diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c
> > > > index 6cd3edb2eaf6..e6d1bc414faf 100644
> > > > --- a/drivers/nvmem/core.c
> > > > +++ b/drivers/nvmem/core.c
> > > > @@ -361,16 +361,15 @@ static void nvmem_cell_add(struct nvmem_cell *cell)
> > > > blocking_notifier_call_chain(&nvmem_notifier, NVMEM_CELL_ADD, cell);
> > > > }
> > > > -static int nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> > > > - const struct nvmem_cell_info *info,
> > > > - struct nvmem_cell *cell)
> > > > +static int
> > > > +__nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> > > > + const struct nvmem_cell_info *info,
> > > > + struct nvmem_cell *cell)
> > > > {
> > > > cell->nvmem = nvmem;
> > > > cell->offset = info->offset;
> > > > cell->bytes = info->bytes;
> > > > - cell->name = kstrdup_const(info->name, GFP_KERNEL);
> > > > - if (!cell->name)
> > > > - return -ENOMEM;
> > > > + cell->name = info->name;
> > > > cell->bit_offset = info->bit_offset;
> > > > cell->nbits = info->nbits;
> > > > @@ -382,13 +381,31 @@ static int nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> > > > if (!IS_ALIGNED(cell->offset, nvmem->stride)) {
> > > > dev_err(&nvmem->dev,
> > > > "cell %s unaligned to nvmem stride %d\n",
> > > > - cell->name, nvmem->stride);
> > > > + cell->name ?: "<unknown>", nvmem->stride);
> > > > return -EINVAL;
> > > > }
> > > > return 0;
> > > > }
> > > > +static int
> > > > +nvmem_cell_info_to_nvmem_cell(struct nvmem_device *nvmem,
> > > > + const struct nvmem_cell_info *info,
> > > > + struct nvmem_cell *cell)
> > > > +{
> > > > + int err;
> > > > +
> > > > + err = __nvmem_cell_info_to_nvmem_cell(nvmem, info, cell);
> > > > + if (err)
> > > > + return err;
> > > > +
> > > > + cell->name = kstrdup_const(info->name, GFP_KERNEL);
> > > > + if (!cell->name)
> > > > + return -ENOMEM;
> > > > +
> > > > + return 0;
> > > > +}
> > > > +
> > > > /**
> > > > * nvmem_add_cells() - Add cell information to an nvmem device
> > > > *
> > > > @@ -1460,7 +1477,7 @@ ssize_t nvmem_device_cell_read(struct nvmem_device *nvmem,
> > > > if (!nvmem)
> > > > return -EINVAL;
> > > > - rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> > > > + rc = __nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> > > > if (rc)
> > > > return rc;
> > > > @@ -1490,7 +1507,7 @@ int nvmem_device_cell_write(struct nvmem_device *nvmem,
> > > > if (!nvmem)
> > > > return -EINVAL;
> > > > - rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> > > > + rc = __nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell);
> > > > if (rc)
> > > > return rc;
> > > >

2020-09-23 15:52:40

by Srinivas Kandagatla

[permalink] [raw]
Subject: Re: [PATCH v2] nvmem: core: fix possibly memleak when use nvmem_cell_info_to_nvmem_cell()



On 23/09/2020 15:51, Vadym Kochan wrote:
>> - return nvmem_cell_write(&cell, buf, cell.bytes);
>> + rc = nvmem_cell_write(&cell, buf, cell.bytes);
>> + if (rc)
>> + kfree_const(cell->name);
>> +
>> + return rc;
>> }
>> EXPORT_SYMBOL_GPL(nvmem_device_cell_write);
>> ------------------------>cut<---------------------------
>>
>> --srini
>>
> But is it really needed to kstrdup(cell->name) for nvmem_device_cell_{read,write} ?
This boils down to if we want to use same api to parse nvmem_cell_info
or not!

If we want to keep this simple, we can either explicitly add free for
successful caller to nvmem_cell_info_to_nvmem_cell()!

Or

use something like what you did, but new api needs more clarity!
May be renaming __nvmem_cell_info_to_nvmem_cell to
nvmem_cell_info_to_nvmem_cell_no_alloc would clarify that a bit!

Also can you make sure that linewrapping on function names be inline
with existing code.

Please send v3 with that changes!


--srini
> It is used only for log error in case the unaligned access did not
> pass the check

2020-09-23 16:04:46

by Vadym Kochan

[permalink] [raw]
Subject: Re: [PATCH v2] nvmem: core: fix possibly memleak when use nvmem_cell_info_to_nvmem_cell()

On Wed, Sep 23, 2020 at 04:51:06PM +0100, Srinivas Kandagatla wrote:
>
>
> On 23/09/2020 15:51, Vadym Kochan wrote:
> > > - return nvmem_cell_write(&cell, buf, cell.bytes);
> > > + rc = nvmem_cell_write(&cell, buf, cell.bytes);
> > > + if (rc)
> > > + kfree_const(cell->name);
> > > +
> > > + return rc;
> > > }
> > > EXPORT_SYMBOL_GPL(nvmem_device_cell_write);
> > > ------------------------>cut<---------------------------
> > >
> > > --srini
> > >
> > But is it really needed to kstrdup(cell->name) for nvmem_device_cell_{read,write} ?
> This boils down to if we want to use same api to parse nvmem_cell_info or
> not!
>
> If we want to keep this simple, we can either explicitly add free for
> successful caller to nvmem_cell_info_to_nvmem_cell()!
>

I think that such additional kfree_const(cell->name) handling adds more
complexity for error handling, also to my understanding usually
resource allocation should be done in the called func in case of error
was returned.

> Or
>
> use something like what you did, but new api needs more clarity!
> May be renaming __nvmem_cell_info_to_nvmem_cell to
> nvmem_cell_info_to_nvmem_cell_no_alloc would clarify that a bit!
>

Yes, I agree that naming should be better, actually "__" already points
to it's unsafety (no kstrdup() is used), but of course additional suffix
would be better.

> Also can you make sure that linewrapping on function names be inline with
> existing code.

You mean do not do such func attributes breaking as I did (moved them
line upper) ?

>
> Please send v3 with that changes!
>
>
> --srini
> > It is used only for log error in case the unaligned access did not
> > pass the check

2020-09-23 16:07:41

by Srinivas Kandagatla

[permalink] [raw]
Subject: Re: [PATCH v2] nvmem: core: fix possibly memleak when use nvmem_cell_info_to_nvmem_cell()



On 23/09/2020 17:02, Vadym Kochan wrote:
> You mean do not do such func attributes breaking as I did (moved them
> line upper) ?
yes, for consistency reasons!

--srini
>

2020-09-23 16:25:06

by Vadym Kochan

[permalink] [raw]
Subject: Re: [PATCH v2] nvmem: core: fix possibly memleak when use nvmem_cell_info_to_nvmem_cell()

On Wed, Sep 23, 2020 at 04:51:06PM +0100, Srinivas Kandagatla wrote:
>
>
> On 23/09/2020 15:51, Vadym Kochan wrote:
> > > - return nvmem_cell_write(&cell, buf, cell.bytes);
> > > + rc = nvmem_cell_write(&cell, buf, cell.bytes);
> > > + if (rc)
> > > + kfree_const(cell->name);
> > > +
> > > + return rc;
> > > }
> > > EXPORT_SYMBOL_GPL(nvmem_device_cell_write);
> > > ------------------------>cut<---------------------------
> > >
> > > --srini
> > >
> > But is it really needed to kstrdup(cell->name) for nvmem_device_cell_{read,write} ?
> This boils down to if we want to use same api to parse nvmem_cell_info or
> not!
>
> If we want to keep this simple, we can either explicitly add free for
> successful caller to nvmem_cell_info_to_nvmem_cell()!
>
> Or
>
> use something like what you did, but new api needs more clarity!
> May be renaming __nvmem_cell_info_to_nvmem_cell to
> nvmem_cell_info_to_nvmem_cell_no_alloc would clarify that a bit!
>

Naming is most difficult thing, what about __nvmem_cell_info_to_nvmem_cell_{unsafe,nodup}() ?
At least this is an indication to be carefully here.

> Also can you make sure that linewrapping on function names be inline with
> existing code.
>
> Please send v3 with that changes!
>
>
> --srini
> > It is used only for log error in case the unaligned access did not
> > pass the check

2020-09-23 16:26:53

by Srinivas Kandagatla

[permalink] [raw]
Subject: Re: [PATCH v2] nvmem: core: fix possibly memleak when use nvmem_cell_info_to_nvmem_cell()



On 23/09/2020 17:23, Vadym Kochan wrote:
> On Wed, Sep 23, 2020 at 04:51:06PM +0100, Srinivas Kandagatla wrote:
>>
>>
>> On 23/09/2020 15:51, Vadym Kochan wrote:
>>>> - return nvmem_cell_write(&cell, buf, cell.bytes);
>>>> + rc = nvmem_cell_write(&cell, buf, cell.bytes);
>>>> + if (rc)
>>>> + kfree_const(cell->name);
>>>> +
>>>> + return rc;
>>>> }
>>>> EXPORT_SYMBOL_GPL(nvmem_device_cell_write);
>>>> ------------------------>cut<---------------------------
>>>>
>>>> --srini
>>>>
>>> But is it really needed to kstrdup(cell->name) for nvmem_device_cell_{read,write} ?
>> This boils down to if we want to use same api to parse nvmem_cell_info or
>> not!
>>
>> If we want to keep this simple, we can either explicitly add free for
>> successful caller to nvmem_cell_info_to_nvmem_cell()!
>>
>> Or
>>
>> use something like what you did, but new api needs more clarity!
>> May be renaming __nvmem_cell_info_to_nvmem_cell to
>> nvmem_cell_info_to_nvmem_cell_no_alloc would clarify that a bit!
>>
>
> Naming is most difficult thing, what about __nvmem_cell_info_to_nvmem_cell_{unsafe,nodup}() ?
> At least this is an indication to be carefully here.

nvmem_cell_info_to_nvmem_cell_nodup() should be good!

--srini
>
>> Also can you make sure that linewrapping on function names be inline with
>> existing code.
>>
>> Please send v3 with that changes!
>>
>>
>> --srini
>>> It is used only for log error in case the unaligned access did not
>>> pass the check