[SLOF] [PATCH slof 05/13] veth: Compile with -Wextra

Thomas Huth thuth at redhat.com
Fri Jan 29 01:31:54 AEDT 2021


On 27/01/2021 09.57, Alexey Kardashevskiy wrote:
> -Wextra enables a bunch of rather useful checks which this fixes.
> 
> Signed-off-by: Alexey Kardashevskiy <aik at ozlabs.ru>
> ---
>   lib/libveth/veth.h | 2 +-
>   lib/libveth/veth.c | 8 ++++++--
>   2 files changed, 7 insertions(+), 3 deletions(-)
> 
> diff --git a/lib/libveth/veth.h b/lib/libveth/veth.h
> index 23af0eab6211..6a1cb4cb5790 100644
> --- a/lib/libveth/veth.h
> +++ b/lib/libveth/veth.h
> @@ -16,7 +16,7 @@
>   #include <stdint.h>
>   #include <netdriver.h>
>   
> -extern net_driver_t *libveth_open(char *mac_addr, int mac_len, char *reg, int reg_len);
> +extern net_driver_t *libveth_open(char *mac_addr, unsigned mac_len, char *reg, unsigned reg_len);
>   extern void libveth_close(net_driver_t *driver);
>   extern int libveth_read(char *buf, int len, net_driver_t *driver);
>   extern int libveth_write(char *buf, int len, net_driver_t *driver);
> diff --git a/lib/libveth/veth.c b/lib/libveth/veth.c
> index 748730854035..a8e19ba41764 100644
> --- a/lib/libveth/veth.c
> +++ b/lib/libveth/veth.c
> @@ -164,7 +164,7 @@ static int veth_term(net_driver_t *driver)
>   	return 0;
>   }
>   
> -static int veth_receive(char *f_buffer_pc, int f_len_i, net_driver_t *driver)
> +static int veth_receive(char *f_buffer_pc, unsigned f_len_i, net_driver_t *driver)
>   {
>   	int packet = 0;
>   
> @@ -223,10 +223,14 @@ static int veth_xmit(char *f_buffer_pc, int f_len_i, net_driver_t *driver)
>   	return f_len_i;
>   }
>   
> -net_driver_t *libveth_open(char *mac_addr, int mac_len, char *reg, int reg_len)
> +net_driver_t *libveth_open(char *mac_addr, unsigned mac_len, char *reg, unsigned reg_len)
>   {
>   	net_driver_t *driver;
>   
> +	if (reg_len != sizeof(uint32_t)) {
> +		printf("vio reg must 1 cell long\n");
> +		return NULL;
> +	}
>   	driver = SLOF_alloc_mem(sizeof(*driver));
>   	if (!driver) {
>   		printf("Unable to allocate veth driver\n");
> 

Is this patch necessary at all? mac_len is only used to compare with 8, and 
reg_len should not matter since you've finally added -Wno-unused-param ... 
so I think you could simply drop this again?

  Thomas



More information about the SLOF mailing list