mirror of
https://mirrors.bfsu.edu.cn/git/linux.git
synced 2024-12-13 05:54:23 +08:00
efacae6d4c
Prepare to mark sensitive kernel structures for randomization by making sure they're using designated initializers. These were identified during allyesconfig builds of x86, arm, and arm64, with most initializer fixes extracted from grsecurity. For these cases, terminate the list with { }, which will be zero-filled, instead of undesignated NULLs. Signed-off-by: Kees Cook <keescook@chromium.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
245 lines
5.9 KiB
C
245 lines
5.9 KiB
C
/*
|
|
* QLogic iSCSI Offload Driver
|
|
* Copyright (c) 2016 Cavium Inc.
|
|
*
|
|
* This software is available under the terms of the GNU General Public License
|
|
* (GPL) Version 2, available from the file COPYING in the main directory of
|
|
* this source tree.
|
|
*/
|
|
|
|
#include "qedi.h"
|
|
#include "qedi_dbg.h"
|
|
|
|
#include <linux/uaccess.h>
|
|
#include <linux/debugfs.h>
|
|
#include <linux/module.h>
|
|
|
|
int do_not_recover;
|
|
static struct dentry *qedi_dbg_root;
|
|
|
|
void
|
|
qedi_dbg_host_init(struct qedi_dbg_ctx *qedi,
|
|
struct qedi_debugfs_ops *dops,
|
|
const struct file_operations *fops)
|
|
{
|
|
char host_dirname[32];
|
|
struct dentry *file_dentry = NULL;
|
|
|
|
sprintf(host_dirname, "host%u", qedi->host_no);
|
|
qedi->bdf_dentry = debugfs_create_dir(host_dirname, qedi_dbg_root);
|
|
if (!qedi->bdf_dentry)
|
|
return;
|
|
|
|
while (dops) {
|
|
if (!(dops->name))
|
|
break;
|
|
|
|
file_dentry = debugfs_create_file(dops->name, 0600,
|
|
qedi->bdf_dentry, qedi,
|
|
fops);
|
|
if (!file_dentry) {
|
|
QEDI_INFO(qedi, QEDI_LOG_DEBUGFS,
|
|
"Debugfs entry %s creation failed\n",
|
|
dops->name);
|
|
debugfs_remove_recursive(qedi->bdf_dentry);
|
|
return;
|
|
}
|
|
dops++;
|
|
fops++;
|
|
}
|
|
}
|
|
|
|
void
|
|
qedi_dbg_host_exit(struct qedi_dbg_ctx *qedi)
|
|
{
|
|
debugfs_remove_recursive(qedi->bdf_dentry);
|
|
qedi->bdf_dentry = NULL;
|
|
}
|
|
|
|
void
|
|
qedi_dbg_init(char *drv_name)
|
|
{
|
|
qedi_dbg_root = debugfs_create_dir(drv_name, NULL);
|
|
if (!qedi_dbg_root)
|
|
QEDI_INFO(NULL, QEDI_LOG_DEBUGFS, "Init of debugfs failed\n");
|
|
}
|
|
|
|
void
|
|
qedi_dbg_exit(void)
|
|
{
|
|
debugfs_remove_recursive(qedi_dbg_root);
|
|
qedi_dbg_root = NULL;
|
|
}
|
|
|
|
static ssize_t
|
|
qedi_dbg_do_not_recover_enable(struct qedi_dbg_ctx *qedi_dbg)
|
|
{
|
|
if (!do_not_recover)
|
|
do_not_recover = 1;
|
|
|
|
QEDI_INFO(qedi_dbg, QEDI_LOG_DEBUGFS, "do_not_recover=%d\n",
|
|
do_not_recover);
|
|
return 0;
|
|
}
|
|
|
|
static ssize_t
|
|
qedi_dbg_do_not_recover_disable(struct qedi_dbg_ctx *qedi_dbg)
|
|
{
|
|
if (do_not_recover)
|
|
do_not_recover = 0;
|
|
|
|
QEDI_INFO(qedi_dbg, QEDI_LOG_DEBUGFS, "do_not_recover=%d\n",
|
|
do_not_recover);
|
|
return 0;
|
|
}
|
|
|
|
static struct qedi_list_of_funcs qedi_dbg_do_not_recover_ops[] = {
|
|
{ "enable", qedi_dbg_do_not_recover_enable },
|
|
{ "disable", qedi_dbg_do_not_recover_disable },
|
|
{ NULL, NULL }
|
|
};
|
|
|
|
struct qedi_debugfs_ops qedi_debugfs_ops[] = {
|
|
{ "gbl_ctx", NULL },
|
|
{ "do_not_recover", qedi_dbg_do_not_recover_ops},
|
|
{ "io_trace", NULL },
|
|
{ NULL, NULL }
|
|
};
|
|
|
|
static ssize_t
|
|
qedi_dbg_do_not_recover_cmd_write(struct file *filp, const char __user *buffer,
|
|
size_t count, loff_t *ppos)
|
|
{
|
|
size_t cnt = 0;
|
|
struct qedi_dbg_ctx *qedi_dbg =
|
|
(struct qedi_dbg_ctx *)filp->private_data;
|
|
struct qedi_list_of_funcs *lof = qedi_dbg_do_not_recover_ops;
|
|
|
|
if (*ppos)
|
|
return 0;
|
|
|
|
while (lof) {
|
|
if (!(lof->oper_str))
|
|
break;
|
|
|
|
if (!strncmp(lof->oper_str, buffer, strlen(lof->oper_str))) {
|
|
cnt = lof->oper_func(qedi_dbg);
|
|
break;
|
|
}
|
|
|
|
lof++;
|
|
}
|
|
return (count - cnt);
|
|
}
|
|
|
|
static ssize_t
|
|
qedi_dbg_do_not_recover_cmd_read(struct file *filp, char __user *buffer,
|
|
size_t count, loff_t *ppos)
|
|
{
|
|
size_t cnt = 0;
|
|
|
|
if (*ppos)
|
|
return 0;
|
|
|
|
cnt = sprintf(buffer, "do_not_recover=%d\n", do_not_recover);
|
|
cnt = min_t(int, count, cnt - *ppos);
|
|
*ppos += cnt;
|
|
return cnt;
|
|
}
|
|
|
|
static int
|
|
qedi_gbl_ctx_show(struct seq_file *s, void *unused)
|
|
{
|
|
struct qedi_fastpath *fp = NULL;
|
|
struct qed_sb_info *sb_info = NULL;
|
|
struct status_block *sb = NULL;
|
|
struct global_queue *que = NULL;
|
|
int id;
|
|
u16 prod_idx;
|
|
struct qedi_ctx *qedi = s->private;
|
|
unsigned long flags;
|
|
|
|
seq_puts(s, " DUMP CQ CONTEXT:\n");
|
|
|
|
for (id = 0; id < MIN_NUM_CPUS_MSIX(qedi); id++) {
|
|
spin_lock_irqsave(&qedi->hba_lock, flags);
|
|
seq_printf(s, "=========FAST CQ PATH [%d] ==========\n", id);
|
|
fp = &qedi->fp_array[id];
|
|
sb_info = fp->sb_info;
|
|
sb = sb_info->sb_virt;
|
|
prod_idx = (sb->pi_array[QEDI_PROTO_CQ_PROD_IDX] &
|
|
STATUS_BLOCK_PROD_INDEX_MASK);
|
|
seq_printf(s, "SB PROD IDX: %d\n", prod_idx);
|
|
que = qedi->global_queues[fp->sb_id];
|
|
seq_printf(s, "DRV CONS IDX: %d\n", que->cq_cons_idx);
|
|
seq_printf(s, "CQ complete host memory: %d\n", fp->sb_id);
|
|
seq_puts(s, "=========== END ==================\n\n\n");
|
|
spin_unlock_irqrestore(&qedi->hba_lock, flags);
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
static int
|
|
qedi_dbg_gbl_ctx_open(struct inode *inode, struct file *file)
|
|
{
|
|
struct qedi_dbg_ctx *qedi_dbg = inode->i_private;
|
|
struct qedi_ctx *qedi = container_of(qedi_dbg, struct qedi_ctx,
|
|
dbg_ctx);
|
|
|
|
return single_open(file, qedi_gbl_ctx_show, qedi);
|
|
}
|
|
|
|
static int
|
|
qedi_io_trace_show(struct seq_file *s, void *unused)
|
|
{
|
|
int id, idx = 0;
|
|
struct qedi_ctx *qedi = s->private;
|
|
struct qedi_io_log *io_log;
|
|
unsigned long flags;
|
|
|
|
seq_puts(s, " DUMP IO LOGS:\n");
|
|
spin_lock_irqsave(&qedi->io_trace_lock, flags);
|
|
idx = qedi->io_trace_idx;
|
|
for (id = 0; id < QEDI_IO_TRACE_SIZE; id++) {
|
|
io_log = &qedi->io_trace_buf[idx];
|
|
seq_printf(s, "iodir-%d:", io_log->direction);
|
|
seq_printf(s, "tid-0x%x:", io_log->task_id);
|
|
seq_printf(s, "cid-0x%x:", io_log->cid);
|
|
seq_printf(s, "lun-%d:", io_log->lun);
|
|
seq_printf(s, "op-0x%02x:", io_log->op);
|
|
seq_printf(s, "0x%02x%02x%02x%02x:", io_log->lba[0],
|
|
io_log->lba[1], io_log->lba[2], io_log->lba[3]);
|
|
seq_printf(s, "buflen-%d:", io_log->bufflen);
|
|
seq_printf(s, "sgcnt-%d:", io_log->sg_count);
|
|
seq_printf(s, "res-0x%08x:", io_log->result);
|
|
seq_printf(s, "jif-%lu:", io_log->jiffies);
|
|
seq_printf(s, "blk_req_cpu-%d:", io_log->blk_req_cpu);
|
|
seq_printf(s, "req_cpu-%d:", io_log->req_cpu);
|
|
seq_printf(s, "intr_cpu-%d:", io_log->intr_cpu);
|
|
seq_printf(s, "blk_rsp_cpu-%d\n", io_log->blk_rsp_cpu);
|
|
|
|
idx++;
|
|
if (idx == QEDI_IO_TRACE_SIZE)
|
|
idx = 0;
|
|
}
|
|
spin_unlock_irqrestore(&qedi->io_trace_lock, flags);
|
|
return 0;
|
|
}
|
|
|
|
static int
|
|
qedi_dbg_io_trace_open(struct inode *inode, struct file *file)
|
|
{
|
|
struct qedi_dbg_ctx *qedi_dbg = inode->i_private;
|
|
struct qedi_ctx *qedi = container_of(qedi_dbg, struct qedi_ctx,
|
|
dbg_ctx);
|
|
|
|
return single_open(file, qedi_io_trace_show, qedi);
|
|
}
|
|
|
|
const struct file_operations qedi_dbg_fops[] = {
|
|
qedi_dbg_fileops_seq(qedi, gbl_ctx),
|
|
qedi_dbg_fileops(qedi, do_not_recover),
|
|
qedi_dbg_fileops_seq(qedi, io_trace),
|
|
{ },
|
|
};
|