2018-06-11 08:54:30

by Zhouyang Jia

Subject: [PATCH] usb: storage: add error handling for kcalloc

When kcalloc fails, the lack of error-handling code may
cause unexpected results.

This patch adds error-handling code after calling kcalloc.

Signed-off-by: Zhouyang Jia <[email protected]>
---
drivers/usb/storage/alauda.c | 5 +++++
1 file changed, 5 insertions(+)

diff --git a/drivers/usb/storage/alauda.c b/drivers/usb/storage/alauda.c
index 900591d..c56355c 100644
--- a/drivers/usb/storage/alauda.c
+++ b/drivers/usb/storage/alauda.c
@@ -437,6 +437,11 @@ static int alauda_init_media(struct us_data *us)
+ MEDIA_INFO(us).blockshift + MEDIA_INFO(us).pageshift);
MEDIA_INFO(us).pba_to_lba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
MEDIA_INFO(us).lba_to_pba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
+ if ((MEDIA_INFO(us).pba_to_lba == NULL)
+ || (MEDIA_INFO(us).lba_to_pba == NULL)) {
+ pr_warn("alauda_init_media: memory allocation failed\n");
+ return USB_STOR_TRANSPORT_ERROR;
+ }

if (alauda_reset_media(us) != USB_STOR_XFER_GOOD)
return USB_STOR_TRANSPORT_ERROR;
--
2.7.4

2018-06-12 14:32:12

by Alan Stern

Subject: Re: [PATCH] usb: storage: add error handling for kcalloc

On Mon, 11 Jun 2018, Zhouyang Jia wrote:

> When kcalloc fails, the lack of error-handling code may
> cause unexpected results.
>
> This patch adds error-handling code after calling kcalloc.
>
> Signed-off-by: Zhouyang Jia <[email protected]>
> ---
> drivers/usb/storage/alauda.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/drivers/usb/storage/alauda.c b/drivers/usb/storage/alauda.c
> index 900591d..c56355c 100644
> --- a/drivers/usb/storage/alauda.c
> +++ b/drivers/usb/storage/alauda.c
> @@ -437,6 +437,11 @@ static int alauda_init_media(struct us_data *us)
> + MEDIA_INFO(us).blockshift + MEDIA_INFO(us).pageshift);
> MEDIA_INFO(us).pba_to_lba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
> MEDIA_INFO(us).lba_to_pba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
> + if ((MEDIA_INFO(us).pba_to_lba == NULL)
> + || (MEDIA_INFO(us).lba_to_pba == NULL)) {
> + pr_warn("alauda_init_media: memory allocation failed\n");
> + return USB_STOR_TRANSPORT_ERROR;
> + }

pr_info() isn't a good routine to use here, because it doesn't print
the device or driver name.

In any case, a log message isn't necessary. kcalloc() will already
put a message in the log if either of the allocations fails.

Alan Stern

>
> if (alauda_reset_media(us) != USB_STOR_XFER_GOOD)
> return USB_STOR_TRANSPORT_ERROR;
>

2018-06-14 13:31:21

by Zhouyang Jia

Subject: [PATCH v2] usb: storage: add error handling for kcalloc

When kcalloc fails, the lack of error-handling code may
cause unexpected results.

This patch adds error-handling code after calling kcalloc.

Signed-off-by: Zhouyang Jia <[email protected]>
---
v1->v2:
- Remove pr_warn statement.
---
drivers/usb/storage/alauda.c | 3 +++
1 file changed, 3 insertions(+)

diff --git a/drivers/usb/storage/alauda.c b/drivers/usb/storage/alauda.c
index 900591d..4e17609 100644
--- a/drivers/usb/storage/alauda.c
+++ b/drivers/usb/storage/alauda.c
@@ -437,6 +437,9 @@ static int alauda_init_media(struct us_data *us)
+ MEDIA_INFO(us).blockshift + MEDIA_INFO(us).pageshift);
MEDIA_INFO(us).pba_to_lba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
MEDIA_INFO(us).lba_to_pba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
+ if ((MEDIA_INFO(us).pba_to_lba == NULL)
+ || (MEDIA_INFO(us).lba_to_pba == NULL))
+ return USB_STOR_TRANSPORT_ERROR;

if (alauda_reset_media(us) != USB_STOR_XFER_GOOD)
return USB_STOR_TRANSPORT_ERROR;
--
2.7.4

2018-06-14 15:00:06

by Alan Stern

Subject: Re: [PATCH v2] usb: storage: add error handling for kcalloc

On Thu, 14 Jun 2018, Zhouyang Jia wrote:

> When kcalloc fails, the lack of error-handling code may
> cause unexpected results.
>
> This patch adds error-handling code after calling kcalloc.
>
> Signed-off-by: Zhouyang Jia <[email protected]>
> ---
> v1->v2:
> - Remove pr_warn statement.
> ---
> drivers/usb/storage/alauda.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/usb/storage/alauda.c b/drivers/usb/storage/alauda.c
> index 900591d..4e17609 100644
> --- a/drivers/usb/storage/alauda.c
> +++ b/drivers/usb/storage/alauda.c
> @@ -437,6 +437,9 @@ static int alauda_init_media(struct us_data *us)
> + MEDIA_INFO(us).blockshift + MEDIA_INFO(us).pageshift);
> MEDIA_INFO(us).pba_to_lba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
> MEDIA_INFO(us).lba_to_pba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
> + if ((MEDIA_INFO(us).pba_to_lba == NULL)
> + || (MEDIA_INFO(us).lba_to_pba == NULL))
> + return USB_STOR_TRANSPORT_ERROR;
>
> if (alauda_reset_media(us) != USB_STOR_XFER_GOOD)
> return USB_STOR_TRANSPORT_ERROR;
>

Acked-by: Alan Stern <[email protected]>

2018-06-25 12:34:35

by Greg Kroah-Hartman

Subject: Re: [PATCH v2] usb: storage: add error handling for kcalloc

On Thu, Jun 14, 2018 at 09:29:11PM +0800, Zhouyang Jia wrote:
> When kcalloc fails, the lack of error-handling code may
> cause unexpected results.
>
> This patch adds error-handling code after calling kcalloc.
>
> Signed-off-by: Zhouyang Jia <[email protected]>
> Acked-by: Alan Stern <[email protected]>
> ---
> v1->v2:
> - Remove pr_warn statement.
> ---
> drivers/usb/storage/alauda.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/usb/storage/alauda.c b/drivers/usb/storage/alauda.c
> index 900591d..4e17609 100644
> --- a/drivers/usb/storage/alauda.c
> +++ b/drivers/usb/storage/alauda.c
> @@ -437,6 +437,9 @@ static int alauda_init_media(struct us_data *us)
> + MEDIA_INFO(us).blockshift + MEDIA_INFO(us).pageshift);
> MEDIA_INFO(us).pba_to_lba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
> MEDIA_INFO(us).lba_to_pba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
> + if ((MEDIA_INFO(us).pba_to_lba == NULL)
> + || (MEDIA_INFO(us).lba_to_pba == NULL))
> + return USB_STOR_TRANSPORT_ERROR;

You just leaked memory if only one of these succeeded :(

2018-06-25 15:24:15

by Alan Stern

Subject: Re: [PATCH v2] usb: storage: add error handling for kcalloc

On Mon, 25 Jun 2018, Greg Kroah-Hartman wrote:

> On Thu, Jun 14, 2018 at 09:29:11PM +0800, Zhouyang Jia wrote:
> > When kcalloc fails, the lack of error-handling code may
> > cause unexpected results.
> >
> > This patch adds error-handling code after calling kcalloc.
> >
> > Signed-off-by: Zhouyang Jia <[email protected]>
> > Acked-by: Alan Stern <[email protected]>
> > ---
> > v1->v2:
> > - Remove pr_warn statement.
> > ---
> > drivers/usb/storage/alauda.c | 3 +++
> > 1 file changed, 3 insertions(+)
> >
> > diff --git a/drivers/usb/storage/alauda.c b/drivers/usb/storage/alauda.c
> > index 900591d..4e17609 100644
> > --- a/drivers/usb/storage/alauda.c
> > +++ b/drivers/usb/storage/alauda.c
> > @@ -437,6 +437,9 @@ static int alauda_init_media(struct us_data *us)
> > + MEDIA_INFO(us).blockshift + MEDIA_INFO(us).pageshift);
> > MEDIA_INFO(us).pba_to_lba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
> > MEDIA_INFO(us).lba_to_pba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
> > + if ((MEDIA_INFO(us).pba_to_lba == NULL)
> > + || (MEDIA_INFO(us).lba_to_pba == NULL))
> > + return USB_STOR_TRANSPORT_ERROR;
>
> You just leaked memory if only one of these succeeded :(

That's not really true. The memory gets deallocated later on in any
case, when alauda_info_destructor() calls alauda_free_maps() if not
before.

More troubling is the fact that this routine (i.e, alauda_init_media)
gets called from only one place, in alauda_check_media, and the caller
completely ignores the return value! Furthermore, the caller always
returns USB_STOR_TRANSPORT_FAILED.

So on the whole, I don't think this patch is going to make any
difference to the driver's operation.

Alan Stern

2018-06-28 16:25:39

by Greg Kroah-Hartman

Subject: Re: [PATCH v2] usb: storage: add error handling for kcalloc

On Mon, Jun 25, 2018 at 11:22:34AM -0400, Alan Stern wrote:
> On Mon, 25 Jun 2018, Greg Kroah-Hartman wrote:
>
> > On Thu, Jun 14, 2018 at 09:29:11PM +0800, Zhouyang Jia wrote:
> > > When kcalloc fails, the lack of error-handling code may
> > > cause unexpected results.
> > >
> > > This patch adds error-handling code after calling kcalloc.
> > >
> > > Signed-off-by: Zhouyang Jia <[email protected]>
> > > Acked-by: Alan Stern <[email protected]>
> > > ---
> > > v1->v2:
> > > - Remove pr_warn statement.
> > > ---
> > > drivers/usb/storage/alauda.c | 3 +++
> > > 1 file changed, 3 insertions(+)
> > >
> > > diff --git a/drivers/usb/storage/alauda.c b/drivers/usb/storage/alauda.c
> > > index 900591d..4e17609 100644
> > > --- a/drivers/usb/storage/alauda.c
> > > +++ b/drivers/usb/storage/alauda.c
> > > @@ -437,6 +437,9 @@ static int alauda_init_media(struct us_data *us)
> > > + MEDIA_INFO(us).blockshift + MEDIA_INFO(us).pageshift);
> > > MEDIA_INFO(us).pba_to_lba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
> > > MEDIA_INFO(us).lba_to_pba = kcalloc(num_zones, sizeof(u16*), GFP_NOIO);
> > > + if ((MEDIA_INFO(us).pba_to_lba == NULL)
> > > + || (MEDIA_INFO(us).lba_to_pba == NULL))
> > > + return USB_STOR_TRANSPORT_ERROR;
> >
> > You just leaked memory if only one of these succeeded :(
>
> That's not really true. The memory gets deallocated later on in any
> case, when alauda_info_destructor() calls alauda_free_maps() if not
> before.
>
> More troubling is the fact that this routine (i.e, alauda_init_media)
> gets called from only one place, in alauda_check_media, and the caller
> completely ignores the return value! Furthermore, the caller always
> returns USB_STOR_TRANSPORT_FAILED.
>
> So on the whole, I don't think this patch is going to make any
> difference to the driver's operation.

Ok, I'm just going to drop it then.

thanks,

greg k-h