hw/char/escc: Lower irq when transmit buffer is filled
PATCH v3 1/1 hw/char/escc: Lower irq when transmit buffer is filled
The SCC/ESCC will briefly stop asserting an interrupt when the transmit FIFO is filled.
This code doesn't model the transmit FIFO/shift register so the pending transmit interrupt is never deasserted which means that an edge-triggered interrupt controller will never see the low-to-high transition it needs to raise another interrupt. The practical consequence of this is that guest firmware with an interrupt service routine for the ESCC that does not send all of the data it has immediately will stop sending data if the following sequence of events occurs:
1. Disable processor interrupts
2. Write a character to the ESCC
3. Add additional characters to a buffer which is drained by the ISR
4. Enable processor interrupts
In this case, the first character will be sent, the interrupt will fire and the ISR will output the second character. Since the pending transmit interrupt remains asserted, no additional interrupts will ever fire.
This behavior was triggered by firmware for an embedded system with a Z85C30 which necessitated this patch.
This patch fixes that situation by explicitly lowering the IRQ when a character is written to the buffer and no other interrupts are currently pending.
---
I added a sentence about the Z85C30 necessitating this to the commit message.
hw/char/escc.c | 7 +++++++ 1 file changed, 7 insertions(+)
hw/char/escc.c +7 -0
@@ -509,6 +509,13 @@ static void escc_mem_write(void *opaque, hwaddr addr,break ;case SERIAL_DATA :trace_escc_mem_writeb_data (CHN_C (s), val); +/* +* Lower the irq when data is written to the Tx buffer and no other +* interrupts are currently pending. The irq will be raised again once +* the Tx buffer becomes empty below. +*/ + s->txint =0 ; +escc_update_irq (s); s->tx = val;if (s->wregs[W_TXCTRL2 ] &TXCTRL2_TXEN ) {// tx enabled if (qemu_chr_fe_backend_connected (&s->chr)) {
Signature · 1 line
2.20.1 (Apple Git-117)
Laurent Vivier
AppliedQuoted 46 lines · Stephen Checkoway
The SCC/ESCC will briefly stop asserting an interrupt when the transmit FIFO is filled.
This code doesn't model the transmit FIFO/shift register so the pending transmit interrupt is never deasserted which means that an edge-triggered interrupt controller will never see the low-to-high transition it needs to raise another interrupt. The practical consequence of this is that guest firmware with an interrupt service routine for the ESCC that does not send all of the data it has immediately will stop sending data if the following sequence of events occurs:
1. Disable processor interrupts
2. Write a character to the ESCC
3. Add additional characters to a buffer which is drained by the ISR
4. Enable processor interruptsIn this case, the first character will be sent, the interrupt will fire and the ISR will output the second character. Since the pending transmit interrupt remains asserted, no additional interrupts will ever fire.
This behavior was triggered by firmware for an embedded system with a Z85C30 which necessitated this patch.
This patch fixes that situation by explicitly lowering the IRQ when a character is written to the buffer and no other interrupts are currently pending.
Signed-off-by Stephen Checkoway <[email protected]>Reviewed-by Philippe Mathieu-Daudé <[email protected]>---
I added a sentence about the Z85C30 necessitating this to the commit message.
hw/char/escc.c | 7 +++++++ 1 file changed, 7 insertions(+)hw/char/escc.c +0
-0@@ -509,6 +509,13 @@ static void escc_mem_write(void *opaque, hwaddr addr,break ;case SERIAL_DATA :trace_escc_mem_writeb_data (CHN_C (s), val);
hw/char/escc.c static void escc_mem_write(void *opaque, hwaddr addr, · line 521
Quoted 10 lines
hw/char/escc.c +7
-0+/* +* Lower the irq when data is written to the Tx buffer and no other +* interrupts are currently pending. The irq will be raised again once +* the Tx buffer becomes empty below. +*/ + s->txint =0 ; +escc_update_irq (s); s->tx = val;if (s->wregs[W_TXCTRL2 ] &TXCTRL2_TXEN ) {// tx enabled if (qemu_chr_fe_backend_connected (&s->chr)) {
Applied to my trivial-patches branch.
Paolo, Marc-André, if you disagree with this change or prefer to take it through one of your queues, I can remove it from mine. Let me know.
Thanks,
Laurent
Quoted 52 lines · Laurent Vivier
Quoted 51 lines · Stephen Checkoway
The SCC/ESCC will briefly stop asserting an interrupt when the transmit FIFO is filled.
This code doesn't model the transmit FIFO/shift register so the pending transmit interrupt is never deasserted which means that an edge-triggered interrupt controller will never see the low-to-high transition it needs to raise another interrupt. The practical consequence of this is that guest firmware with an interrupt service routine for the ESCC that does not send all of the data it has immediately will stop sending data if the following sequence of events occurs:
1. Disable processor interrupts
2. Write a character to the ESCC
3. Add additional characters to a buffer which is drained by the ISR
4. Enable processor interruptsIn this case, the first character will be sent, the interrupt will fire and the ISR will output the second character. Since the pending transmit interrupt remains asserted, no additional interrupts will ever fire.
This behavior was triggered by firmware for an embedded system with a Z85C30 which necessitated this patch.
This patch fixes that situation by explicitly lowering the IRQ when a character is written to the buffer and no other interrupts are currently pending.
Signed-off-by Stephen Checkoway <[email protected]>Reviewed-by Philippe Mathieu-Daudé <[email protected]>---
I added a sentence about the Z85C30 necessitating this to the commit message.
hw/char/escc.c | 7 +++++++ 1 file changed, 7 insertions(+)hw/char/escc.c +6
-0@@ -509,6 +509,13 @@ static void escc_mem_write(void *opaque, hwaddr addr,break ;case SERIAL_DATA :trace_escc_mem_writeb_data (CHN_C (s), val); +/* +* Lower the irq when data is written to the Tx buffer and no other +* interrupts are currently pending. The irq will be raised again once +* the Tx buffer becomes empty below. +*/ + s->txint =0 ;
Quoted 8 lines
hw/char/escc.c static void escc_mem_write(void *opaque, hwaddr addr, · line 521
Quoted 5 lines
hw/char/escc.c +1
-0+escc_update_irq (s); s->tx = val;if (s->wregs[W_TXCTRL2 ] &TXCTRL2_TXEN ) {// tx enabled if (qemu_chr_fe_backend_connected (&s->chr)) {Applied to my trivial-patches branch.
Mark, Artyom, are you OK with this patch?
Quoted 8 lines
Paolo, Marc-André, if you disagree with this change or prefer to take it through one of your queues, I can remove it from mine. Let me know.
Thanks,
Laurent
Quoted 62 lines · Philippe Mathieu-Daudé
Quoted 52 lines · Laurent Vivier
Quoted 51 lines · Stephen Checkoway
The SCC/ESCC will briefly stop asserting an interrupt when the transmit FIFO is filled.
This code doesn't model the transmit FIFO/shift register so the pending transmit interrupt is never deasserted which means that an edge-triggered interrupt controller will never see the low-to-high transition it needs to raise another interrupt. The practical consequence of this is that guest firmware with an interrupt service routine for the ESCC that does not send all of the data it has immediately will stop sending data if the following sequence of events occurs:
1. Disable processor interrupts
2. Write a character to the ESCC
3. Add additional characters to a buffer which is drained by the ISR
4. Enable processor interruptsIn this case, the first character will be sent, the interrupt will fire and the ISR will output the second character. Since the pending transmit interrupt remains asserted, no additional interrupts will ever fire.
This behavior was triggered by firmware for an embedded system with a Z85C30 which necessitated this patch.
This patch fixes that situation by explicitly lowering the IRQ when a character is written to the buffer and no other interrupts are currently pending.
Signed-off-by Stephen Checkoway <[email protected]>Reviewed-by Philippe Mathieu-Daudé <[email protected]>---
I added a sentence about the Z85C30 necessitating this to the commit message.
hw/char/escc.c | 7 +++++++ 1 file changed, 7 insertions(+)hw/char/escc.c +6
-0@@ -509,6 +509,13 @@ static void escc_mem_write(void *opaque, hwaddr addr,break ;case SERIAL_DATA :trace_escc_mem_writeb_data (CHN_C (s), val); +/* +* Lower the irq when data is written to the Tx buffer and no other +* interrupts are currently pending. The irq will be raised again once +* the Tx buffer becomes empty below. +*/ + s->txint =0 ;Quoted 8 lines
hw/char/escc.c static void escc_mem_write(void *opaque, hwaddr addr, · line 521
Quoted 5 lines
hw/char/escc.c +1
-0+escc_update_irq (s); s->tx = val;if (s->wregs[W_TXCTRL2 ] &TXCTRL2_TXEN ) {// tx enabled if (qemu_chr_fe_backend_connected (&s->chr)) {Applied to my trivial-patches branch.
Quoted 1 line
Mark, Artyom, are you OK with this patch?
I started testing this with my OpenBIOS test images at the start of the week, but unfortunately got distracted by real life :)
I've now finished and confirmed there are no regressions in my local tests, so I'll include this in the PR I am planning to send shortly containing the leon3 updates.
ATB,
Mark.
Quoted 64 lines · Mark Cave-Ayland
Quoted 62 lines · Philippe Mathieu-Daudé
Quoted 60 lines · Laurent Vivier
Quoted 46 lines · Stephen Checkoway
The SCC/ESCC will briefly stop asserting an interrupt when the transmit FIFO is filled.
This code doesn't model the transmit FIFO/shift register so the pending transmit interrupt is never deasserted which means that an edge-triggered interrupt controller will never see the low-to-high transition it needs to raise another interrupt. The practical consequence of this is that guest firmware with an interrupt service routine for the ESCC that does not send all of the data it has immediately will stop sending data if the following sequence of events occurs:
1. Disable processor interrupts
2. Write a character to the ESCC
3. Add additional characters to a buffer which is drained by the ISR
4. Enable processor interruptsIn this case, the first character will be sent, the interrupt will fire and the ISR will output the second character. Since the pending transmit interrupt remains asserted, no additional interrupts will ever fire.
This behavior was triggered by firmware for an embedded system with a Z85C30 which necessitated this patch.
This patch fixes that situation by explicitly lowering the IRQ when a character is written to the buffer and no other interrupts are currently pending.
Signed-off-by Stephen Checkoway <[email protected]>Reviewed-by Philippe Mathieu-Daudé <[email protected]>---
I added a sentence about the Z85C30 necessitating this to the commit message.
hw/char/escc.c | 7 +++++++ 1 file changed, 7 insertions(+)hw/char/escc.c +0
-0@@ -509,6 +509,13 @@ static void escc_mem_write(void *opaque, hwaddr addr,break ;case SERIAL_DATA :trace_escc_mem_writeb_data (CHN_C (s), val);hw/char/escc.c static void escc_mem_write(void *opaque, hwaddr addr, · line 521
Quoted 10 lines
hw/char/escc.c +7
-0+/* +* Lower the irq when data is written to the Tx buffer and no other +* interrupts are currently pending. The irq will be raised again once +* the Tx buffer becomes empty below. +*/ + s->txint =0 ; +escc_update_irq (s); s->tx = val;if (s->wregs[W_TXCTRL2 ] &TXCTRL2_TXEN ) {// tx enabled if (qemu_chr_fe_backend_connected (&s->chr)) {Applied to my trivial-patches branch.
Quoted 7 lines
Quoted 1 line
Mark, Artyom, are you OK with this patch?
I started testing this with my OpenBIOS test images at the start of the week, but unfortunately got distracted by real life :)
I've now finished and confirmed there are no regressions in my local tests, so I'll include this in the PR I am planning to send shortly containing the leon3 updates.
Hi Mark,
I've already included a leon3 patch ("hw/sparc/leon3: Allow load of uImage firmwares") in the PR I sent yesterday for the trivial branch. I've removed from my PR this patch about the ESCC, so you can take it.
Thanks,
Laurent
Quoted 5 lines · Laurent Vivier
Quoted 5 lines
Quoted 1 line
Mark, Artyom, are you OK with this patch?
I started testing this with my OpenBIOS test images at the start of the week, but unfortunately got distracted by real life :)
Quoted 8 lines
Quoted 2 lines
I've now finished and confirmed there are no regressions in my local tests, so I'll include this in the PR I am planning to send shortly containing the leon3 updates.
Hi Mark,
I've already included a leon3 patch ("hw/sparc/leon3: Allow load of uImage firmwares") in the PR I sent yesterday for the trivial branch. I've removed from my PR this patch about the ESCC, so you can take it.
Hi Laurent,
Understood, and thanks for looking out for the patch. My first week back has been quite busy so I'm still playing catch-up here...
ATB,
Mark.