Skip to contentSkip to conversation

hw/net/can: Add missing fallthrough statements

PATCH Applied Thomas Huth 3 participants 4 messages Jun 30, 2020 – Sep 1, 2020

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 hw/net/can: Add missing fallthrough statements

Add fallthrough annotations to be able to compile the code without warnings when using -Wimplicit-fallthrough in our CFLAGS. Looking at the code, it seems like the fallthrough is indeed intended here, so the comments should be appropriate.

---
 hw/net/can/can_sja1000.c | 2 ++
 1 file changed, 2 insertions(+)
hw/net/can/can_sja1000.c +2 -0
index ea915a023a..299932998a 100644
@@ -523,6 +523,7 @@ void can_sja_mem_write(CanSJA1000State *s, hwaddr addr, uint64_t val,
             break;
         case 16: /* RX frame information addr16-28. */
             s->status_pel |= (1 << 5); /* Set transmit status. */
+            /* fallthrough */
         case 17 ... 28:
             if (s->mode & 0x01) { /* Reset mode */
                 if (addr < 24) {
@@ -620,6 +621,7 @@ void can_sja_mem_write(CanSJA1000State *s, hwaddr addr, uint64_t val,
             break;
         case 10:
             s->status_bas |= (1 << 5); /* Set transmit status. */
+            /* fallthrough */
         case 11 ... 19:
             if ((s->control & 0x01) == 0) { /* Operation mode */
                 s->tx_buff[addr - 10] = val; /* Store to TX buffer directly. */
Signature · 1 line
2.18.1
Reviewed-by on v1 1/1
#1 lore raw

Le 30/06/2020 à 09:55, Thomas Huth a écrit :

Quoted 24 lines

Add fallthrough annotations to be able to compile the code without warnings when using -Wimplicit-fallthrough in our CFLAGS. Looking at the code, it seems like the fallthrough is indeed intended here, so the comments should be appropriate.

---
 hw/net/can/can_sja1000.c | 2 ++
 1 file changed, 2 insertions(+)
hw/net/can/can_sja1000.c +1 -0
index ea915a023a..299932998a 100644
@@ -523,6 +523,7 @@ void can_sja_mem_write(CanSJA1000State *s, hwaddr addr, uint64_t val,
             break;
         case 16: /* RX frame information addr16-28. */
             s->status_pel |= (1 << 5); /* Set transmit status. */
+            /* fallthrough */
         case 17 ... 28:
             if (s->mode & 0x01) { /* Reset mode */
                 if (addr < 24) {

hw/net/can/can_sja1000.c void can_sja_mem_write(CanSJA1000State *s, hwaddr addr, uint64_t val, · line 627

Quoted 7 lines
hw/net/can/can_sja1000.c +1 -0
@@ -620,6 +621,7 @@ void can_sja_mem_write(CanSJA1000State *s, hwaddr addr, uint64_t val,
             break;
         case 10:
             s->status_bas |= (1 << 5); /* Set transmit status. */
+            /* fallthrough */
         case 11 ... 19:
             if ((s->control & 0x01) == 0) { /* Operation mode */
                 s->tx_buff[addr - 10] = val; /* Store to TX buffer directly. */
cc Pavel Pisa <[email protected]>
Reviewed-by
#2 lore raw

Hello Laurent and others,

Quoted 35 lines · On Monday 06 of July 2020 18:35:50 Laurent Vivier wrote:

Le 30/06/2020 à 09:55, Thomas Huth a écrit :

Quoted 30 lines

Add fallthrough annotations to be able to compile the code without warnings when using -Wimplicit-fallthrough in our CFLAGS. Looking at the code, it seems like the fallthrough is indeed intended here, so the comments should be appropriate.

---
 hw/net/can/can_sja1000.c | 2 ++
 1 file changed, 2 insertions(+)
hw/net/can/can_sja1000.c +2 -0
index ea915a023a..299932998a 100644
@@ -523,6 +523,7 @@ void can_sja_mem_write(CanSJA1000State *s, hwaddr
 addr, uint64_t val, break;
         case 16: /* RX frame information addr16-28. */
             s->status_pel |= (1 << 5); /* Set transmit status. */
+            /* fallthrough */
         case 17 ... 28:
             if (s->mode & 0x01) { /* Reset mode */
                 if (addr < 24) {
@@ -620,6 +621,7 @@ void can_sja_mem_write(CanSJA1000State *s, hwaddr
 addr, uint64_t val, break;
         case 10:
             s->status_bas |= (1 << 5); /* Set transmit status. */
+            /* fallthrough */
         case 11 ... 19:
             if ((s->control & 0x01) == 0) { /* Operation mode */
                 s->tx_buff[addr - 10] = val; /* Store to TX buffer
Quoted 1 line

directly. */

cc Pavel Pisa <[email protected]>
Quoted 1 line

The fallthrough is intentional for sure but I have gone through datasheet and checked why the status bit is set there and my conclusion is that to mimic real HW the status bit should not be set there. In the fact, it should be set and immediately (in a future delayed) reset after SJA_CMR transmit request write. This would mimic real hardware more closely. May it be I send patch in future when more of our developed CAN support is added to QEMU. The status bit behavior has no influence on actual Linux SJA1000 driver anyway.

So for now, I confirm that adding /* fallthrough */ is correct step forward.

By the way, we have prepared CAN FD support for QEMU, the CAN core update and device model to emulate our open-source/design/hardware CTU CAN FD IP core

https://gitlab.fel.cvut.cz/canbus/ctucanfd_ip_core

QEMU emulation

https://gitlab.fel.cvut.cz/canbus/qemu-canbus/-/commits/charvj10-canfd

I hope to find time to add patch to document CAN support to CAN FD extension and send whole series this week. Stay tuned, please.

Best wishes,

Pavel

Signature · 9 lines
                Pavel Pisa
    phone:      +420 603531357
    e-mail:     [email protected]
    Department of Control Engineering FEE CVUT
    Karlovo namesti 13, 121 35, Prague 2
    university: http://dce.fel.cvut.cz/
    personal:   http://cmp.felk.cvut.cz/~pisa
    projects:   https://www.openhub.net/accounts/ppisa
    CAN related:http://canbus.pages.fel.cvut.cz/
#3 lore raw

Le 14/07/2020 à 00:09, Pavel Pisa a écrit :

Quoted 51 lines

Hello Laurent and others,

Quoted 35 lines · On Monday 06 of July 2020 18:35:50 Laurent Vivier wrote:

Le 30/06/2020 à 09:55, Thomas Huth a écrit :

Quoted 30 lines

Add fallthrough annotations to be able to compile the code without warnings when using -Wimplicit-fallthrough in our CFLAGS. Looking at the code, it seems like the fallthrough is indeed intended here, so the comments should be appropriate.

---
 hw/net/can/can_sja1000.c | 2 ++
 1 file changed, 2 insertions(+)
hw/net/can/can_sja1000.c +2 -0
index ea915a023a..299932998a 100644
@@ -523,6 +523,7 @@ void can_sja_mem_write(CanSJA1000State *s, hwaddr
 addr, uint64_t val, break;
         case 16: /* RX frame information addr16-28. */
             s->status_pel |= (1 << 5); /* Set transmit status. */
+            /* fallthrough */
         case 17 ... 28:
             if (s->mode & 0x01) { /* Reset mode */
                 if (addr < 24) {
@@ -620,6 +621,7 @@ void can_sja_mem_write(CanSJA1000State *s, hwaddr
 addr, uint64_t val, break;
         case 10:
             s->status_bas |= (1 << 5); /* Set transmit status. */
+            /* fallthrough */
         case 11 ... 19:
             if ((s->control & 0x01) == 0) { /* Operation mode */
                 s->tx_buff[addr - 10] = val; /* Store to TX buffer
Quoted 1 line

directly. */

cc Pavel Pisa <[email protected]>
Quoted 1 line

The fallthrough is intentional for sure but I have gone through datasheet and checked why the status bit is set there and my conclusion is that to mimic real HW the status bit should not be set there. In the fact, it should be set and immediately (in a future delayed) reset after SJA_CMR transmit request write. This would mimic real hardware more closely. May it be I send patch in future when more of our developed CAN support is added to QEMU. The status bit behavior has no influence on actual Linux SJA1000 driver anyway.

Quoted 5 lines

So for now, I confirm that adding /* fallthrough */ is correct step forward.

Applied to my trivial-patches branch.

Thanks,
Laurent