kernel/linux.git/drivers/serial/serial_core.c, branch linux-2.6.16.y

[SERIAL] Fix oops when removing suspended serial port

2007-03-28T19:28:32+00:00

A serial card might have been removed when the system is resumed. This results in a suspended port being shut down, which results in the ports shutdown method being called twice in a row. This causes BUGs. Avoid this by tracking the suspended state separately from the initialised state. Signed-off-by: Russell King Signed-off-by: Adrian Bunk

[PATCH] serial core: work around sub-driver bugs

2006-03-08T22:14:00+00:00

We're presently getting oopses because Bluetooth (and possibly other) drivers are calling core functions after things have been shut down. So rather than oopsing, let's drop a warning then take avoiding action, so the machine survives. Once all the sub-drivers are fixed up we can remove the take-avoiding-action part. Signed-off-by: Pavel Machek Cc: Russell King Cc: Marcel Holtmann Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

[SERIAL] initialize spinlock for port failed to setup console

2006-02-02T20:53:39+00:00

It seems serial_core intend to initialize port->lock just once for each ports. This is done in uart_set_options() for console, and in uart_add_one_port() for other ports. But there is a case the port->lock is not initialized by serial_core. If the setup function for the console was failed, it will not call uart_set_options() but the port is marked as console (uart_console(port) returns 1). It can happen if console was PCI port which can not detected at the time of register_console. This patch is to initialize port->lock for such console port. With this change, most of spin_lock_init() (some of them are labeled "Temporary fix.") in low-level serial drivers can be omitted. Signed-off-by: Atsushi Nemoto Signed-off-by: Andrew Morton Signed-off-by: Russell King

[SERIAL] Make uart_port flags a bitwise type

2006-01-21T23:03:28+00:00

Same reasoning as commit 747c8a55946ed037bf7d62454c3c599c02af2262 but this time we're making uart_port flags a bitwise type - not all of these flags correspond with the old ASYNC_ flags, so there is the possibility for bugs if the wrong ASYNC_* constants are used. Always use UPF_* constants for uart_port->flags. Signed-off-by: Russell King

[SERIAL] Make port->ops constant

2006-01-21T22:45:50+00:00

No one should write to the port->ops structure, so make it constant. Signed-off-by: Russell King

[SERIAL] convert uart_state.sem to uart_state.mutex

2006-01-13T21:37:07+00:00

semaphore to mutex conversion. the conversion was generated via scripts, and the result was validated automatically via a script as well. build and boot tested. Signed-off-by: Ingo Molnar Signed-off-by: Russell King

[SERIAL] turn serial semaphores into mutexes

2006-01-12T18:44:32+00:00

Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven Signed-off-by: Russell King

[SERIAL] Fix clocal wakeup problem

2006-01-07T23:11:23+00:00

Jim Alexander reported a problem where "if one calls open() in blocking mode with CLOCAL off, the 8250.c driver under the 2.6 kernel (or at least 2.6.8 and 2.6.10) does not wake up the blocked process when DCD is asserted." Fix this by enabling modem status interrupts immediately before we read the carrier detect status. Thanks to Jim for reporting the problem and testing the fix. Signed-off-by: Russell King

[SERIAL] Fix matching of MMIO ports

2006-01-04T18:09:44+00:00

The function uart_match_port() incorrectly compares the ioremap'd virtual addresses of ports instead of the physical address to find duplicate ports for MMIO based UARTs. This fixes it. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Russell King

[SERIAL] mark several serial tables const

2005-11-28T21:04:11+00:00

This patch marks a few serial data structures const, moving them to .rodata where they won't false-share cachelines with things that get written to. Signed-off-by: Arjan van de Ven Signed-off-by: Russell King