serial: pmac_zilog: two receive fixes
Subsystems (3)
cover serial: pmac_zilog: two receive fixes
These two patches fix receive problems in pmac_zilog that I ran into while bringing up the serial ports on a Performa 6200CD, a NuBus Power Macintosh. The first stops data received on one port from being left in its flip buffer while the other port is open. The second makes received breaks visible; until now they arrived as plain NUL characters.
I tested both on the 6200, running 7.2.8 with an out-of-tree platform port and a loopback plug on the printer port. Without the series, only 3 of 8 bytes written to ttyPZ1 came back while ttyPZ0 was open, and a break sent with tcsendbreak() arrived as a NUL with no break counted. With the series, all 8 bytes came back, the break was counted, and with PARMRK set it read as FF 00 00.
The series also builds on v7.3-rc7 with pmac32_defconfig and W=1 without warnings. I have not tested 68k Macs or macio Power Macs.
I found these bugs and wrote the patches with the help of Claude Opus 5.5. I have reviewed both patches and run the tests on the 6200 myself.
Tim Douglas (2):
serial: pmac_zilog: push received data to the port that received it
serial: pmac_zilog: report received breaks
drivers/tty/serial/pmac_zilog.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
base-commit: 24a19b953e2a3d0ea5a8c4951a0ab8f32720c907
Signature · 1 line
2.54.0 (Apple Git-157)
PATCH 1/2 serial: pmac_zilog: push received data to the port that received it
pmz_interrupt() drains both channels but, since commit 2e124b4a390c ("TTY: switch tty_flip_buffer_push"), pushes the flip buffer of the port whose handler is running rather than the port that was drained. With both ports open, data received on the other port can be left in its flip buffer until something else pushes it, so its reader may not see it.
--- drivers/tty/serial/pmac_zilog.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/tty/serial/pmac_zilog.c +2 -2
@@ -444,7 +444,7 @@ static irqreturn_t pmz_interrupt(int irq, void *dev_id) skip_a:uart_port_unlock (&uap_a->port);if (push) -tty_flip_buffer_push (&uap->port.state->port); +tty_flip_buffer_push (&uap_a->port.state->port);if (!uap_b) goto out;
@@ -469,7 +469,7 @@ static irqreturn_t pmz_interrupt(int irq, void *dev_id) skip_b:uart_port_unlock (&uap_b->port);if (push) -tty_flip_buffer_push (&uap->port.state->port); +tty_flip_buffer_push (&uap_b->port.state->port); out:return rc;
Signature · 1 line
2.54.0 (Apple Git-157)
[email protected]
AI reviewRe: [PATCH 1/2] serial: pmac_zilog: push received data to the port that received it
Sashiko has reviewed this patch and found no issues. It looks great!
Quoted 11 lines
pmz_interrupt()drains both channels but, since commit2e124b4a390c("TTY: switchtty_flip_buffer_push"), pushes the flip buffer of the port whose handler is running rather than the port that was drained. With both ports open, data received on the other port can be left in its flip buffer until something else pushes it, so its reader may not see it.Fixes 2e124b4a390c ("TTY: switch tty_flip_buffer_push")Assisted-by LLMSigned-off-by Tim Douglas <[email protected]>
Sashiko has reviewed this patch and found no issues. It looks great!
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
PATCH 2/2 serial: pmac_zilog: report received breaks
In reply to Tim Douglaspmz_status_handle() sets PMACZILOG_FLAG_BREAK from RR0, and pmz_receive_chars() clears it on the null character that ends the break without marking that character, then tests BRK_ABRT in RR1, where the SCC does not set it in asynchronous mode. Breaks therefore arrive as plain NULs: brk is never counted and BRKINT, PARMRK and SysRq never see them.
Set BRK_ABRT in r1 for that character, as sunzilog does.
--- drivers/tty/serial/pmac_zilog.c | 1 + 1 file changed, 1 insertion(+)
drivers/tty/serial/pmac_zilog.c +1 -0
@@ -233,6 +233,7 @@ static bool pmz_receive_chars(struct uart_pmac_port *uap) ch &= uap->parity_mask;if (ch ==0 && uap->flags &PMACZILOG_FLAG_BREAK ) { uap->flags &= ~PMACZILOG_FLAG_BREAK ; + r1 |=BRK_ABRT ; }#if defined(CONFIG_MAGIC_SYSRQ ) && defined(CONFIG_SERIAL_CORE_CONSOLE )
Signature · 1 line
2.54.0 (Apple Git-157)
[email protected]
AI reviewRe: [PATCH 2/2] serial: pmac_zilog: report received breaks
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Medium] tty: serial: false break injection on NUL after Rx overrun --
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] tty: serial: false break injection on NUL after Rx overrun
--
commit 009d6cf5b2db60250bcc44251e6de09814cc2746
Author: Tim Douglas <[email protected]>
serial: pmac_zilog: report received breaks
This commit modifies pmz_receive_chars() to set the BRK_ABRT flag in r1 when processing the null character that ends a break. This allows breaks to be properly counted and passed to BRKINT, PARMRK, and SysRq handlers, matching the behavior of sunzilog.
drivers/tty/serial/pmac_zilog.c static bool pmz_receive_chars(struct uart_pmac_port *uap) · line 237
Quoted 10 lines
drivers/tty/serial/pmac_zilog.c +1
-0@@ -233,6 +233,7 @@ static bool pmz_receive_chars(struct uart_pmac_port *uap) ch &= uap->parity_mask;if (ch ==0 && uap->flags &PMACZILOG_FLAG_BREAK ) { uap->flags &= ~PMACZILOG_FLAG_BREAK ; + r1 |=BRK_ABRT ; }
[Severity: Medium]
Does this code risk injecting a false Break event if an Rx FIFO overrun occurs?
pmz_status_handle() sets PMACZILOG_FLAG_BREAK when it receives an EXT status interrupt for a Break. Then pmz_receive_chars() waits for the associated NUL character to clear the flag.
If an Rx FIFO overrun occurs and the hardware drops the expected NUL character, the PMACZILOG_FLAG_BREAK flag will remain set indefinitely.
Since BRK_ABRT is now artificially injected into r1 when this flag is finally cleared, could a subsequent, perfectly valid NUL character (such as Ctrl-@) arriving much later falsely trigger a Break?
This might cause dropped data or trigger an unintended SysRq. Before this patch, the stuck flag was harmless because it was cleared without altering r1.
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=2