[v6,08/12] net/dpaa2: fix possible use of uninitialized vars
Checks
Commit Message
This patch fixes 'maybe-uninitialized' warnings reported by compiler
when using LTO.
Compiler warning pointing to this error (with LTO enabled):
error: ‘kg_cfg.extracts[0].masks[0].mask’ may be used uninitialized in
this function [-Werror=maybe-uninitialized]
extr->masks[j].mask = cfg->extracts[i].masks[j].mask;
Fixes: 16bbc98a3e63 ("bus/fslmc: update MC to 10.3.x")
Signed-off-by: Andrzej Ostruszka <aostruszka@marvell.com>
---
drivers/net/dpaa2/base/dpaa2_hw_dpni.c | 1 +
drivers/net/dpaa2/mc/dpkg.c | 2 +-
drivers/net/dpaa2/mc/dpni.c | 9 ++++++---
3 files changed, 8 insertions(+), 4 deletions(-)
Comments
Hi Andrzej
> -----Original Message-----
> From: Andrzej Ostruszka <aostruszka@marvell.com>
> Sent: Tuesday, October 29, 2019 7:42 PM
> To: dev@dpdk.org; Hemant Agrawal <hemant.agrawal@nxp.com>; Sachin
> Saxena <sachin.saxena@nxp.com>
> Cc: mattias.ronnblom@ericsson.com; stephen@networkplumber.org
> Subject: [PATCH v6 08/12] net/dpaa2: fix possible use of uninitialized vars
> Importance: High
>
> This patch fixes 'maybe-uninitialized' warnings reported by compiler when
> using LTO.
>
> Compiler warning pointing to this error (with LTO enabled):
> error: ‘kg_cfg.extracts[0].masks[0].mask’ may be used uninitialized in this
> function [-Werror=maybe-uninitialized]
> extr->masks[j].mask = cfg->extracts[i].masks[j].mask;
>
> Fixes: 16bbc98a3e63 ("bus/fslmc: update MC to 10.3.x")
> Signed-off-by: Andrzej Ostruszka <aostruszka@marvell.com>
> ---
> drivers/net/dpaa2/base/dpaa2_hw_dpni.c | 1 +
> drivers/net/dpaa2/mc/dpkg.c | 2 +-
> drivers/net/dpaa2/mc/dpni.c | 9 ++++++---
> 3 files changed, 8 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
> b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
> index 16555d7b0..47a8bda6a 100644
> --- a/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
> +++ b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
> @@ -51,6 +51,7 @@ rte_pmd_dpaa2_set_custom_hash(uint16_t port_id,
> kg_cfg.extracts[0].type = DPKG_EXTRACT_FROM_DATA;
> kg_cfg.extracts[0].extract.from_data.offset = offset;
> kg_cfg.extracts[0].extract.from_data.size = size;
> + kg_cfg.extracts[0].num_of_byte_masks = 0;
> kg_cfg.num_extracts = 1;
>
> ret = dpkg_prepare_key_cfg(&kg_cfg, p_params); diff --git
> a/drivers/net/dpaa2/mc/dpkg.c b/drivers/net/dpaa2/mc/dpkg.c index
> 80f94f40e..7aa63ea12 100644
> --- a/drivers/net/dpaa2/mc/dpkg.c
> +++ b/drivers/net/dpaa2/mc/dpkg.c
> @@ -63,7 +63,7 @@ dpkg_prepare_key_cfg(const struct dpkg_profile_cfg
> *cfg, uint8_t *key_cfg_buf)
> dpkg_set_field(extr->extract_type, EXTRACT_TYPE,
> cfg->extracts[i].type);
>
[Hemant] Please add a check here to avoid array overrun
+ if (extr->num_of_byte_masks > DPKG_NUM_OF_MASKS)
+ return -EINVAL;
+
> - for (j = 0; j < DPKG_NUM_OF_MASKS; j++) {
> + for (j = 0; j < extr->num_of_byte_masks; j++) {
> extr->masks[j].mask = cfg-
> >extracts[i].masks[j].mask;
> extr->masks[j].offset =
> cfg->extracts[i].masks[j].offset;
> diff --git a/drivers/net/dpaa2/mc/dpni.c b/drivers/net/dpaa2/mc/dpni.c
> index 0950ee007..89a64b0ab 100644
> --- a/drivers/net/dpaa2/mc/dpni.c
> +++ b/drivers/net/dpaa2/mc/dpni.c
> @@ -1839,10 +1839,13 @@ int dpni_set_congestion_notification(struct
> fsl_mc_io *mc_io,
> cmd_params->qtype = qtype;
> cmd_params->tc = tc_id;
> cmd_params->congestion_point = cfg->cg_point;
> - cmd_params->cgid = (uint8_t)cfg->cgid;
> - cmd_params->dest_id = cpu_to_le32(cfg->dest_cfg.dest_id);
> + if (cfg->cg_point == DPNI_CP_CONGESTION_GROUP)
> + cmd_params->cgid = (uint8_t)cfg->cgid;
> + if (cfg->dest_cfg.dest_type != DPNI_DEST_NONE) {
> + cmd_params->dest_id = cpu_to_le32(cfg-
> >dest_cfg.dest_id);
> + cmd_params->dest_priority = cfg->dest_cfg.priority;
> + }
[Hemant] What is the explicit error you got here?
In the calling function, we are resetting the structure to 0. Which should have avoided
Any un-initialized var errors?
struct dpni_congestion_notification_cfg cong_notif_cfg = {0};
> cmd_params->notification_mode = cpu_to_le16(cfg-
> >notification_mode);
> - cmd_params->dest_priority = cfg->dest_cfg.priority;
> cmd_params->message_iova = cpu_to_le64(cfg->message_iova);
> cmd_params->message_ctx = cpu_to_le64(cfg->message_ctx);
> cmd_params->threshold_entry = cpu_to_le32(cfg-
> >threshold_entry);
> --
> 2.17.1
On 11/4/19 12:46 PM, Hemant Agrawal wrote:
> Hi Andrzej
Hello Hemant,
Thank you for your comments.
>> -----Original Message-----
>> From: Andrzej Ostruszka <aostruszka@marvell.com>
[...]
>> a/drivers/net/dpaa2/mc/dpkg.c b/drivers/net/dpaa2/mc/dpkg.c index
>> 80f94f40e..7aa63ea12 100644
>> --- a/drivers/net/dpaa2/mc/dpkg.c
>> +++ b/drivers/net/dpaa2/mc/dpkg.c
>> @@ -63,7 +63,7 @@ dpkg_prepare_key_cfg(const struct dpkg_profile_cfg
>> *cfg, uint8_t *key_cfg_buf)
>> dpkg_set_field(extr->extract_type, EXTRACT_TYPE,
>> cfg->extracts[i].type);
>>
> [Hemant] Please add a check here to avoid array overrun
> + if (extr->num_of_byte_masks > DPKG_NUM_OF_MASKS)
> + return -EINVAL;
> +
Will do, thank you.
>> - for (j = 0; j < DPKG_NUM_OF_MASKS; j++) {
>> + for (j = 0; j < extr->num_of_byte_masks; j++) {
>> extr->masks[j].mask = cfg-
>>> extracts[i].masks[j].mask;
>> extr->masks[j].offset =
>> cfg->extracts[i].masks[j].offset;
>> diff --git a/drivers/net/dpaa2/mc/dpni.c b/drivers/net/dpaa2/mc/dpni.c
>> index 0950ee007..89a64b0ab 100644
>> --- a/drivers/net/dpaa2/mc/dpni.c
>> +++ b/drivers/net/dpaa2/mc/dpni.c
>> @@ -1839,10 +1839,13 @@ int dpni_set_congestion_notification(struct
>> fsl_mc_io *mc_io,
>> cmd_params->qtype = qtype;
>> cmd_params->tc = tc_id;
>> cmd_params->congestion_point = cfg->cg_point;
>> - cmd_params->cgid = (uint8_t)cfg->cgid;
>> - cmd_params->dest_id = cpu_to_le32(cfg->dest_cfg.dest_id);
>> + if (cfg->cg_point == DPNI_CP_CONGESTION_GROUP)
>> + cmd_params->cgid = (uint8_t)cfg->cgid;
>> + if (cfg->dest_cfg.dest_type != DPNI_DEST_NONE) {
>> + cmd_params->dest_id = cpu_to_le32(cfg-
>>> dest_cfg.dest_id);
>> + cmd_params->dest_priority = cfg->dest_cfg.priority;
>> + }
> [Hemant] What is the explicit error you got here?
> In the calling function, we are resetting the structure to 0. Which should have avoided
> Any un-initialized var errors?
> struct dpni_congestion_notification_cfg cong_notif_cfg = {0};
Indeed there is no warning from compiler here at the current version.
The original version was developed against 18.05 and this is what I used
to make compiler happy. I've kept this change (even though compiler
does not report anything here) since to me it looked like a good change
(use the cgid only in case of "congestion group" and use dest_* only
when not DEST_NONE).
You have obviously more authority here so I will remove these changes
(unless you prefer them to stay - in which case please indicate so).
Again thank you for the comments.
Regards
Andrzej
@@ -51,6 +51,7 @@ rte_pmd_dpaa2_set_custom_hash(uint16_t port_id,
kg_cfg.extracts[0].type = DPKG_EXTRACT_FROM_DATA;
kg_cfg.extracts[0].extract.from_data.offset = offset;
kg_cfg.extracts[0].extract.from_data.size = size;
+ kg_cfg.extracts[0].num_of_byte_masks = 0;
kg_cfg.num_extracts = 1;
ret = dpkg_prepare_key_cfg(&kg_cfg, p_params);
@@ -63,7 +63,7 @@ dpkg_prepare_key_cfg(const struct dpkg_profile_cfg *cfg, uint8_t *key_cfg_buf)
dpkg_set_field(extr->extract_type, EXTRACT_TYPE,
cfg->extracts[i].type);
- for (j = 0; j < DPKG_NUM_OF_MASKS; j++) {
+ for (j = 0; j < extr->num_of_byte_masks; j++) {
extr->masks[j].mask = cfg->extracts[i].masks[j].mask;
extr->masks[j].offset =
cfg->extracts[i].masks[j].offset;
@@ -1839,10 +1839,13 @@ int dpni_set_congestion_notification(struct fsl_mc_io *mc_io,
cmd_params->qtype = qtype;
cmd_params->tc = tc_id;
cmd_params->congestion_point = cfg->cg_point;
- cmd_params->cgid = (uint8_t)cfg->cgid;
- cmd_params->dest_id = cpu_to_le32(cfg->dest_cfg.dest_id);
+ if (cfg->cg_point == DPNI_CP_CONGESTION_GROUP)
+ cmd_params->cgid = (uint8_t)cfg->cgid;
+ if (cfg->dest_cfg.dest_type != DPNI_DEST_NONE) {
+ cmd_params->dest_id = cpu_to_le32(cfg->dest_cfg.dest_id);
+ cmd_params->dest_priority = cfg->dest_cfg.priority;
+ }
cmd_params->notification_mode = cpu_to_le16(cfg->notification_mode);
- cmd_params->dest_priority = cfg->dest_cfg.priority;
cmd_params->message_iova = cpu_to_le64(cfg->message_iova);
cmd_params->message_ctx = cpu_to_le64(cfg->message_ctx);
cmd_params->threshold_entry = cpu_to_le32(cfg->threshold_entry);