#include "../../wusbcore/wusbhc.h"
#include "whcd.h"
static int whc_update_di(struct whc *whc, int idx)
{
int offset = idx / 32;
u32 bit = 1 << (idx % 32);
le_writel(bit, whc->base + WUSBDIBUPDATED + offset);
return whci_wait_for(&whc->umc->dev,
whc->base + WUSBDIBUPDATED + offset, bit, 0,
100, "DI update");
}
/*
* WHCI starts MMCs based on there being a valid GTK so these need
* only start/stop the asynchronous and periodic schedules and send a
* channel stop command.
*/
int whc_wusbhc_start(struct wusbhc *wusbhc)
{
struct whc *whc = wusbhc_to_whc(wusbhc);
asl_start(whc);
pzl_start(whc);
return 0;
}
void whc_wusbhc_stop(struct wusbhc *wusbhc, int delay)
{
struct whc *whc = wusbhc_to_whc(wusbhc);
u32 stop_time, now_time;
int ret;
pzl_stop(whc);
asl_stop(whc);
now_time = le_readl(whc->base + WUSBTIME) & WUSBTIME_CHANNEL_TIME_MASK;
stop_time = (now_time + ((delay * 8) << 7)) & 0x00ffffff;
ret = whc_do_gencmd(whc, WUSBGENCMDSTS_CHAN_STOP, stop_time, NULL, 0);
if (ret == 0)
msleep(delay);
}
int whc_mmcie_add(struct wusbhc *wusbhc, u8 interval, u8 repeat_cnt,
u8 handle, struct wuie_hdr *wuie)
{
struct whc *whc = wusbhc_to_whc(wusbhc);
u32 params;
params = (interval << 24)
| (repeat_cnt << 16)
| (wuie->bLength << 8)
| handle;
return whc_do_gencmd(whc, WUSBGENCMDSTS_MMCIE_ADD, params, wuie, wuie->bLength);
}
int whc_mmcie_rm(struct wusbhc *wusbhc, u8 handle)
{
struct whc *whc = wusbhc_to_whc(wusbhc);
u32 params;
params = handle;
return whc_do_gencmd(whc, WUSBGENCMDSTS_MMCIE_RM, params, NULL, 0);
}
int whc_bwa_set(struct wusbhc *wusbhc, s8 stream_index, const struct uwb_mas_bm *mas_bm)
{
struct whc *whc = wusbhc_to_whc(wusbhc);
if (stream_index >= 0)
whc_write_wusbcmd(whc, WUSBCMD_WUSBSI_MASK, WUSBCMD_WUSBSI(stream_index));
return whc_do_gencmd(whc, WUSBGENCMDSTS_SET_MAS, 0, (void *)mas_bm, sizeof(*mas_bm));
}
int whc_dev_info_set(struct wusbhc *wusbhc, struct wusb_dev *wusb_dev)
{
struct whc *whc = wusbhc_to_whc(wusbhc);
int idx = wusb_dev->port_idx;
struct di_buf_entry *di = &whc->di_buf[idx];
int ret;
mutex_lock(&whc->mutex);
uwb_mas_bm_copy_le(di->availability_info, &wusb_dev->availability);
di->addr_sec_info &= ~(WHC_DI_DISABLE | WHC_DI_DEV_ADDR_MASK);
di->addr_sec_info |= WHC_DI_DEV_ADDR(wusb_dev->addr);
ret = whc_update_di(whc, idx);
mutex_unlock(&whc->mutex);
return ret;
}
/*
* Set the number of Device Notification Time Slots (DNTS) and enable
* device notifications.
*/
int whc_set_num_dnts(struct wusbhc *wusbhc, u8 interval, u8 slots)
{
struct whc *whc = wusbhc_to_whc(wusbhc);
u32 dntsctrl;
dntsctrl = WUSBDNTSCTRL_ACTIVE
| WUSBDNTSCTRL_INTERVAL(interval)
| WUSBDNTSCTRL_SLOTS(slots);
le_writel(dntsctrl, whc->base + WUSBDNTSCTRL);
return 0;
}
static int whc_set_key(struct whc *whc, u8 key_index, uint32_t tkid,
const void *key, size_t key_size, bool is_gtk)
{
uint32_t setkeycmd;
uint32_t seckey[4];
int i;
int ret;
memcpy(seckey, key, key_size);
setkeycmd = WUSBSETSECKEYCMD_SET | WUSBSETSECKEYCMD_IDX(key_index);
if (is_gtk)
setkeycmd |= WUSBSETSECKEYCMD_GTK;
le_writel(tkid, whc->base + WUSBTKID);
for (i = 0; i < 4; i++)
le_writel(seckey[i], whc->base + WUSBSECKEY + 4*i);
le_writel(setkeycmd, whc->base + WUSBSETSECKEYCMD);
ret = whci_wait_for(&whc->umc->dev, whc->base + WUSBSETSECKEYCMD,
WUSBSETSECKEYCMD_SET, 0, 100, "set key");
return ret;
}
/**
* whc_set_ptk - set the PTK to use for a device.
*
* The index into the key table for this PTK is the same as the
* device's port index.
*/
int whc_set_ptk(struct wusbhc *wusbhc, u8 port_idx, u32 tkid,
const void *ptk, size_t key_size)
{
struct whc *whc = wusbhc_to_whc(wusbhc);
struct di_buf_entry *di = &whc->di_buf[port_idx];
int ret;
mutex_lock(&whc->mutex);
if (ptk) {
ret = whc_set_key(whc, port_idx, tkid, ptk, key_size, false);
if (ret)
goto out;
di->addr_sec_info &= ~WHC_DI_KEY_IDX_MASK;
di->addr_sec_info |= WHC_DI_SECURE | WHC_DI_KEY_IDX(port_idx);
} else
di->addr_sec_info &= ~WHC_DI_SECURE;
ret = whc_update_di(whc, port_idx);
out:
mutex_unlock(&whc->mutex);
return ret;
}
/**
* whc_set_gtk - set the GTK for subsequent broadcast packets
*
* The GTK is stored in the last entry in the key table (the previous
* N_DEVICES entries are for the per-device PTKs).
*/
int whc_set_gtk(struct wusbhc *wusbhc, u32 tkid,
const void *gtk, size_t key_size)
{
struct whc *whc = wusbhc_to_whc(wusbhc);
int ret;
mutex_lock(&whc->mutex);
ret = whc_set_key(whc, whc->n_devices, tkid, gtk, key_size, true);
mutex_unlock(&whc->mutex);
return ret;
}
int whc_set_cluster_id(struct whc *whc, u8 bcid)
{
whc_write_wusbcmd(whc, WUSBCMD_BCID_MASK, WUSBCMD_BCID(bcid));
return 0;
}
c9a316e77f6dfba96c543e55b6672d5a37e'>patch)
tree | 98fe974ee4e20121253de7f61fc8d01bdb3821c1 /kernel/power/suspend.c |
parent | 2c5d9555d6d937966d79d4c6529a5f7b9206e405 (diff) |
drm/i915: Check for NULL i915_vma in intel_unpin_fb_obj()
I've seen this trigger twice now, where the i915_gem_object_to_ggtt()
call in intel_unpin_fb_obj() returns NULL, resulting in an oops
immediately afterwards as the (inlined) call to i915_vma_unpin_fence()
tries to dereference it.
It seems to be some race condition where the object is going away at
shutdown time, since both times happened when shutting down the X
server. The call chains were different:
- VT ioctl(KDSETMODE, KD_TEXT):
intel_cleanup_plane_fb+0x5b/0xa0 [i915]
drm_atomic_helper_cleanup_planes+0x6f/0x90 [drm_kms_helper]
intel_atomic_commit_tail+0x749/0xfe0 [i915]
intel_atomic_commit+0x3cb/0x4f0 [i915]
drm_atomic_commit+0x4b/0x50 [drm]
restore_fbdev_mode+0x14c/0x2a0 [drm_kms_helper]
drm_fb_helper_restore_fbdev_mode_unlocked+0x34/0x80 [drm_kms_helper]
drm_fb_helper_set_par+0x2d/0x60 [drm_kms_helper]
intel_fbdev_set_par+0x18/0x70 [i915]
fb_set_var+0x236/0x460
fbcon_blank+0x30f/0x350
do_unblank_screen+0xd2/0x1a0
vt_ioctl+0x507/0x12a0
tty_ioctl+0x355/0xc30
do_vfs_ioctl+0xa3/0x5e0
SyS_ioctl+0x79/0x90
entry_SYSCALL_64_fastpath+0x13/0x94
- i915 unpin_work workqueue:
intel_unpin_work_fn+0x58/0x140 [i915]
process_one_work+0x1f1/0x480
worker_thread+0x48/0x4d0
kthread+0x101/0x140
and this patch purely papers over the issue by adding a NULL pointer
check and a WARN_ON_ONCE() to avoid the oops that would then generally
make the machine unresponsive.
Other callers of i915_gem_object_to_ggtt() seem to also check for the
returned pointer being NULL and warn about it, so this clearly has
happened before in other places.
[ Reported it originally to the i915 developers on Jan 8, applying the
ugly workaround on my own now after triggering the problem for the
second time with no feedback.
This is likely to be the same bug reported as
https://bugs.freedesktop.org/show_bug.cgi?id=98829
https://bugs.freedesktop.org/show_bug.cgi?id=99134
which has a patch for the underlying problem, but it hasn't gotten to
me, so I'm applying the workaround. ]
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>