Serial.c Bug
Hi,
I have found a bug. It is in support of serial cards which uses
memory for I/O
insted of ports. I made a patch for serial.c and fix one place, but probably the problem like this one could be somewhere else.
If you try to use setserial with such cards you will get "Address in
use" (-EADDRINUSE)
Best regards,
Kurakin Roman
serial.c renamed from serial.c.orig +4 -0
@@ -2077,6 +2077,7 @@unsigned int i,change_irq,change_port;int retval =0 ;unsigned long new_port; +unsigned long new_mem;if (copy_from_user (&new_serial,new_info,sizeof (new_serial)))return -EFAULT ;
@@ -2087,6 +2088,8 @@if (HIGH_BITS_OFFSET ) new_port += (unsigned long ) new_serial.port_high <<HIGH_BITS_OFFSET ; + new_mem = new_serial.iomem_base; + change_irq = new_serial.irq != state->irq; change_port = (new_port != ((int ) state->port)) || (new_serial.hub6 != state->hub6);
@@ -2127,6 +2130,7 @@for (i =0 ; i <NR_PORTS ; i++)if ((state != &rs_table[i]) && (rs_table[i].port == new_port) && + (rs_table[i].iomem_base == new_mem) && rs_table[i].type)return -EADDRINUSE ; }
Quoted 4 lines · Roman Kurakin
I have found a bug. It is in support of serial cards which uses
memory for I/O insted of ports. I made a patch forserial.cand fix one place, but probably the problem like this one could be somewhere else.
I've got this fish caught in the my serial driver rewrite - the driver always handles the requesting and freeing of the resources. If it is unable to request the resources, then you will receive a suitable error when trying to configure two ports.
Please note that I'm not about to take on maintainence of the current serial.c driver, except where I spot obvious bugs.
I'd recommend that you pass this one to Marcelo to incorporate (only after he's got his feet on the ground again. 8)) It looks sensible.
--
Russell King ([email protected]) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html
Thanks
Quoted 30 lines · Russell King
Quoted 5 lines · Roman Kurakin
I have found a bug. It is in support of serial cards which uses
memory for I/O insted of ports. I made a patch forserial.cand fix one place, but probably the problem like this one could be somewhere else.I've got this fish caught in the my serial driver rewrite - the driver always handles the requesting and freeing of the resources. If it is unable to request the resources, then you will receive a suitable error when trying to configure two ports.
Please note that I'm not about to take on maintainence of the current
serial.cdriver, except where I spot obvious bugs.I'd recommend that you pass this one to Marcelo to incorporate (only after he's got his feet on the ground again. 8)) It looks sensible.
--
Russell King ([email protected]) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/