mirror of
https://github.com/AuxXxilium/linux_dsm_epyc7002.git
synced 2024-12-10 09:46:38 +07:00
7d12e780e0
Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
172 lines
4.4 KiB
C
172 lines
4.4 KiB
C
/*
|
|
* QLogic Fibre Channel HBA Driver
|
|
* Copyright (c) 2003-2005 QLogic Corporation
|
|
*
|
|
* See LICENSE.qla2xxx for copyright and licensing details.
|
|
*/
|
|
|
|
static __inline__ uint16_t qla2x00_debounce_register(volatile uint16_t __iomem *);
|
|
/*
|
|
* qla2x00_debounce_register
|
|
* Debounce register.
|
|
*
|
|
* Input:
|
|
* port = register address.
|
|
*
|
|
* Returns:
|
|
* register value.
|
|
*/
|
|
static __inline__ uint16_t
|
|
qla2x00_debounce_register(volatile uint16_t __iomem *addr)
|
|
{
|
|
volatile uint16_t first;
|
|
volatile uint16_t second;
|
|
|
|
do {
|
|
first = RD_REG_WORD(addr);
|
|
barrier();
|
|
cpu_relax();
|
|
second = RD_REG_WORD(addr);
|
|
} while (first != second);
|
|
|
|
return (first);
|
|
}
|
|
|
|
static __inline__ int qla2x00_normalize_dma_addr(
|
|
dma_addr_t *e_addr, uint32_t *e_len,
|
|
dma_addr_t *ne_addr, uint32_t *ne_len);
|
|
|
|
/**
|
|
* qla2x00_normalize_dma_addr() - Normalize an DMA address.
|
|
* @e_addr: Raw DMA address
|
|
* @e_len: Raw DMA length
|
|
* @ne_addr: Normalized second DMA address
|
|
* @ne_len: Normalized second DMA length
|
|
*
|
|
* If the address does not span a 4GB page boundary, the contents of @ne_addr
|
|
* and @ne_len are undefined. @e_len is updated to reflect a normalization.
|
|
*
|
|
* Example:
|
|
*
|
|
* ffffabc0ffffeeee (e_addr) start of DMA address
|
|
* 0000000020000000 (e_len) length of DMA transfer
|
|
* ffffabc11fffeeed end of DMA transfer
|
|
*
|
|
* Is the 4GB boundary crossed?
|
|
*
|
|
* ffffabc0ffffeeee (e_addr)
|
|
* ffffabc11fffeeed (e_addr + e_len - 1)
|
|
* 00000001e0000003 ((e_addr ^ (e_addr + e_len - 1))
|
|
* 0000000100000000 ((e_addr ^ (e_addr + e_len - 1)) & ~(0xffffffff)
|
|
*
|
|
* Compute start of second DMA segment:
|
|
*
|
|
* ffffabc0ffffeeee (e_addr)
|
|
* ffffabc1ffffeeee (0x100000000 + e_addr)
|
|
* ffffabc100000000 (0x100000000 + e_addr) & ~(0xffffffff)
|
|
* ffffabc100000000 (ne_addr)
|
|
*
|
|
* Compute length of second DMA segment:
|
|
*
|
|
* 00000000ffffeeee (e_addr & 0xffffffff)
|
|
* 0000000000001112 (0x100000000 - (e_addr & 0xffffffff))
|
|
* 000000001fffeeee (e_len - (0x100000000 - (e_addr & 0xffffffff))
|
|
* 000000001fffeeee (ne_len)
|
|
*
|
|
* Adjust length of first DMA segment
|
|
*
|
|
* 0000000020000000 (e_len)
|
|
* 0000000000001112 (e_len - ne_len)
|
|
* 0000000000001112 (e_len)
|
|
*
|
|
* Returns non-zero if the specified address was normalized, else zero.
|
|
*/
|
|
static __inline__ int
|
|
qla2x00_normalize_dma_addr(
|
|
dma_addr_t *e_addr, uint32_t *e_len,
|
|
dma_addr_t *ne_addr, uint32_t *ne_len)
|
|
{
|
|
int normalized;
|
|
|
|
normalized = 0;
|
|
if ((*e_addr ^ (*e_addr + *e_len - 1)) & ~(0xFFFFFFFFULL)) {
|
|
/* Compute normalized crossed address and len */
|
|
*ne_addr = (0x100000000ULL + *e_addr) & ~(0xFFFFFFFFULL);
|
|
*ne_len = *e_len - (0x100000000ULL - (*e_addr & 0xFFFFFFFFULL));
|
|
*e_len -= *ne_len;
|
|
|
|
normalized++;
|
|
}
|
|
return (normalized);
|
|
}
|
|
|
|
static __inline__ void qla2x00_poll(scsi_qla_host_t *);
|
|
static inline void
|
|
qla2x00_poll(scsi_qla_host_t *ha)
|
|
{
|
|
ha->isp_ops.intr_handler(0, ha);
|
|
}
|
|
|
|
static __inline__ void qla2x00_check_fabric_devices(scsi_qla_host_t *);
|
|
/*
|
|
* This routine will wait for fabric devices for
|
|
* the reset delay.
|
|
*/
|
|
static __inline__ void qla2x00_check_fabric_devices(scsi_qla_host_t *ha)
|
|
{
|
|
uint16_t fw_state;
|
|
|
|
qla2x00_get_firmware_state(ha, &fw_state);
|
|
}
|
|
|
|
/**
|
|
* qla2x00_issue_marker() - Issue a Marker IOCB if necessary.
|
|
* @ha: HA context
|
|
* @ha_locked: is function called with the hardware lock
|
|
*
|
|
* Returns non-zero if a failure occured, else zero.
|
|
*/
|
|
static inline int
|
|
qla2x00_issue_marker(scsi_qla_host_t *ha, int ha_locked)
|
|
{
|
|
/* Send marker if required */
|
|
if (ha->marker_needed != 0) {
|
|
if (ha_locked) {
|
|
if (__qla2x00_marker(ha, 0, 0, MK_SYNC_ALL) !=
|
|
QLA_SUCCESS)
|
|
return (QLA_FUNCTION_FAILED);
|
|
} else {
|
|
if (qla2x00_marker(ha, 0, 0, MK_SYNC_ALL) !=
|
|
QLA_SUCCESS)
|
|
return (QLA_FUNCTION_FAILED);
|
|
}
|
|
ha->marker_needed = 0;
|
|
}
|
|
return (QLA_SUCCESS);
|
|
}
|
|
|
|
static inline uint8_t *host_to_fcp_swap(uint8_t *, uint32_t);
|
|
static inline uint8_t *
|
|
host_to_fcp_swap(uint8_t *fcp, uint32_t bsize)
|
|
{
|
|
uint32_t *ifcp = (uint32_t *) fcp;
|
|
uint32_t *ofcp = (uint32_t *) fcp;
|
|
uint32_t iter = bsize >> 2;
|
|
|
|
for (; iter ; iter--)
|
|
*ofcp++ = swab32(*ifcp++);
|
|
|
|
return fcp;
|
|
}
|
|
|
|
static inline int qla2x00_is_reserved_id(scsi_qla_host_t *, uint16_t);
|
|
static inline int
|
|
qla2x00_is_reserved_id(scsi_qla_host_t *ha, uint16_t loop_id)
|
|
{
|
|
if (IS_QLA24XX(ha) || IS_QLA54XX(ha))
|
|
return (loop_id > NPH_LAST_HANDLE);
|
|
|
|
return ((loop_id > ha->last_loop_id && loop_id < SNS_FIRST_LOOP_ID) ||
|
|
loop_id == MANAGEMENT_SERVER || loop_id == BROADCAST);
|
|
};
|