RFC: drop ISA support in the synlink tty driver?
Hi Greg and Jiri,
I've been working hard to get rid of the remaining callers the pass a NULL struct device to the DMA mapping functions and am almost done.
The only non-trivial driver is the synclink driver, which has legacy early 90s style ISA support that doesn't use the device model at all.
In theory we could convert it to an isa_driver, but without testing that seems rather dangerous. So for now I would suggest that we remove the ISA support in this driver - if anyone cares enough we can resurrect it from the git history and convert it to use the driver model.
PATCH 1/1 tty/synclink: remove ISA support
The ISA support in this driver is horribly outdated and the last driver that passes a NULL dev argument to the DMA mapping routines.
Drop the support, if someone wants to go through the work of converting it to a proper isa_driver we can resurrect it.
--- drivers/tty/synclink.c | 54 ------------------------------------------ 1 file changed, 54 deletions(-)
drivers/tty/synclink.c +0 -54
@@ -4325,41 +4325,6 @@ static int mgsl_init_tty(void)return 0 ; } -/* enumerate user specified ISA adapters -*/ -static void mgsl_enum_isa_devices (void ) -{ -struct mgsl_struct *info; -int i; - -/* Check for user specified ISA devices */ - -for (i=0 ;(i <MAX_ISA_DEVICES ) && io[i] && irq[i]; i++){ -if ( debug_level >=DEBUG_LEVEL_INFO ) -printk ("ISA device specified io=%04X,irq=%d,dma=%d\n" , - io[i], irq[i], dma[i] ); - - info =mgsl_allocate_device (); -if ( !info ) { -/* error allocating device instance data */ -if ( debug_level >=DEBUG_LEVEL_ERROR ) -printk ("can't allocate device instance data.\n" ); -continue ; - } - -/* Copy user configuration info to device instance data */ - info->io_base = (unsigned int )io[i]; - info->irq_level = (unsigned int )irq[i]; - info->irq_level =irq_canonicalize (info->irq_level); - info->dma_level = (unsigned int )dma[i]; - info->bus_type =MGSL_BUS_TYPE_ISA ; - info->io_addr_size =16 ; - info->irq_flags =0 ; - -mgsl_add_device ( info ); - } -} -static void synclink_cleanup (void ) {int rc;
@@ -4403,7 +4368,6 @@ static int __init synclink_init(void)printk ("%s %s\n" , driver_name, driver_version); -mgsl_enum_isa_devices ();if ((rc =pci_register_driver (&synclink_pci_driver)) <0 )printk ("%s:failed to register PCI driver, error=%d\n" ,__FILE__,rc);else
@@ -5025,12 +4989,6 @@ static void usc_set_sdlc_mode( struct mgsl_struct *info ) info->mbre_bit =BIT8 ;outw (BIT8 , info->io_base );/* set Master Bus Enable (DCAR) */ -if (info->bus_type ==MGSL_BUS_TYPE_ISA ) { -/* Enable DMAEN (Port 7, Bit 14) */ -/* This connects the DMA request signal to the ISA bus */ -usc_OutReg (info,PCR , (u16)((usc_InReg (info,PCR ) |BIT15 ) & ~BIT14 )); - } -/* DMA Control Register (DCR) * * <15..14> 10 Priority mode = Alternating Tx/Rx
@@ -6007,12 +5965,6 @@ static void usc_set_async_mode( struct mgsl_struct *info )usc_EnableMasterIrqBit ( info ); -if (info->bus_type ==MGSL_BUS_TYPE_ISA ) { -/* Enable INTEN (Port 6, Bit12) */ -/* This connects the IRQ request signal to the ISA bus */ -usc_OutReg (info,PCR , (u16)((usc_InReg (info,PCR ) |BIT13 ) & ~BIT12 )); - } -if (info->params.loopback) { info->loopback_bits =0x300 ;outw (0x0300 , info->io_base +CCAR );
@@ -6107,12 +6059,6 @@ static void usc_set_sync_mode( struct mgsl_struct *info )usc_loopback_frame ( info );usc_set_sdlc_mode ( info ); -if (info->bus_type ==MGSL_BUS_TYPE_ISA ) { -/* Enable INTEN (Port 6, Bit12) */ -/* This connects the IRQ request signal to the ISA bus */ -usc_OutReg (info,PCR , (u16)((usc_InReg (info,PCR ) |BIT13 ) & ~BIT12 )); - } -usc_enable_aux_clock (info, info->params.clock_speed);if (info->params.loopback)
Signature · 1 line
2.20.1
Greg Kroah-Hartman
maintainerQuoted 5 lines · Christoph Hellwig
Hi Greg and Jiri,
I've been working hard to get rid of the remaining callers the pass a NULL struct device to the DMA mapping functions and am almost done.
Quoted 8 lines
The only non-trivial driver is the synclink driver, which has legacy early 90s style ISA support that doesn't use the device model at all.
In theory we could convert it to an
isa_driver, but without testing that seems rather dangerous. So for now I would suggest that we remove the ISA support in this driver - if anyone cares enough we can resurrect it from the git history and convert it to use the driver model.
No objection from me at all, I'll go queue this up now, thanks.
greg k-h
Quoted 9 lines · Greg Kroah-Hartman
Quoted 8 lines · Christoph Hellwig
Hi Greg and Jiri,
I've been working hard to get rid of the remaining callers the pass a NULL struct device to the DMA mapping functions and am almost done.
The only non-trivial driver is the synclink driver, which has legacy early 90s style ISA support that doesn't use the device model at all.
Quoted 7 lines
Quoted 5 lines
In theory we could convert it to an
isa_driver, but without testing that seems rather dangerous. So for now I would suggest that we remove the ISA support in this driver - if anyone cares enough we can resurrect it from the git history and convert it to use the driver model.No objection from me at all, I'll go queue this up now, thanks.
Agreed, but I would kill also the MGSL_BUS_TYPE_ISA macro proper.
thanks,
Signature · 2 lines
js suse labs
Greg Kroah-Hartman
maintainerQuoted 16 lines · Jiri Slaby
Quoted 15 lines · Greg Kroah-Hartman
Quoted 13 lines · Christoph Hellwig
Hi Greg and Jiri,
I've been working hard to get rid of the remaining callers the pass a NULL struct device to the DMA mapping functions and am almost done.
The only non-trivial driver is the synclink driver, which has legacy early 90s style ISA support that doesn't use the device model at all.
In theory we could convert it to an
isa_driver, but without testing that seems rather dangerous. So for now I would suggest that we remove the ISA support in this driver - if anyone cares enough we can resurrect it from the git history and convert it to use the driver model.
Quoted 3 lines
Quoted 1 line
No objection from me at all, I'll go queue this up now, thanks.
Agreed, but I would kill also the
MGSL_BUS_TYPE_ISAmacro proper.
It's just a #define, in a uapi file, so we should probably leave it as userspace programs _might_ depend on it. I have no idea why, but oh well...
thanks,
greg k-h
Quoted 3 lines · Greg Kroah-Hartman
It's just a #define, in a uapi file, so we should probably leave it as userspace programs _might_ depend on it. I have no idea why, but oh well...
Uh, I didn't even think this could be in a uapi header. Ok, it makes sense then.
thanks,
Signature · 2 lines
js suse labs