There are some conflicts between staging and media trees, as reported by Stephen Rothwell <sfr@canb.auug.org.au>. So, merge from staging. * tag 'staging-4.15-rc1': (775 commits) staging: lustre: add SPDX identifiers to all lustre files staging: greybus: Remove redundant license text staging: greybus: add SPDX identifiers to all greybus driver files staging: ccree: simplify ioread/iowrite staging: ccree: simplify registers access staging: ccree: simplify error handling logic staging: ccree: remove dead code staging: ccree: handle limiting of DMA masks staging: ccree: copy IV to DMAable memory staging: fbtft: remove redundant initialization of buf staging: sm750fb: Fix parameter mistake in poke32 staging: wilc1000: Fix bssid buffer offset in Txq staging: fbtft: fb_ssd1331: fix mirrored display staging: android: Fix checkpatch.pl error staging: greybus: loopback: convert loopback to use generic async operations staging: greybus: operation: add private data with get/set accessors staging: greybus: loopback: Fix iteration count on async path staging: greybus: loopback: Hold per-connection mutex across operations staging: greybus/loopback: use ktime_get() for time intervals staging: fsl-dpaa2/eth: Extra headroom in RX buffers ...
Showing the changes to include/uapi/linux (1 of 1302 files). Show all
Unified · Split
| @@ -52,21 +52,30 @@ · 51 unmodified lines | |||
| 52 | * (non-running threads are de facto in such a | 52 | * (non-running threads are de facto in such a |
| 53 | * state). This only covers threads from the | 53 | * state). This only covers threads from the |
| 54 | * same processes as the caller thread. This | 54 | * same processes as the caller thread. This |
| 55 | * command returns 0. The "expedited" commands | 55 | * command returns 0 on success. The |
| 56 | * complete faster than the non-expedited ones, | 56 | * "expedited" commands complete faster than |
| 57 | * they never block, but have the downside of | 57 | * the non-expedited ones, they never block, |
| 58 | * causing extra overhead. | 58 | * but have the downside of causing extra |
| 59 | * overhead. A process needs to register its | ||
| 60 | * intent to use the private expedited command | ||
| 61 | * prior to using it, otherwise this command | ||
| 62 | * returns -EPERM. | ||
| 63 | * @MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED: | ||
| 64 | * Register the process intent to use | ||
| 65 | * MEMBARRIER_CMD_PRIVATE_EXPEDITED. Always | ||
| 66 | * returns 0. | ||
| 59 | * | 67 | * |
| 60 | * Command to be passed to the membarrier system call. The commands need to | 68 | * Command to be passed to the membarrier system call. The commands need to |
| 61 | * be a single bit each, except for MEMBARRIER_CMD_QUERY which is assigned to | 69 | * be a single bit each, except for MEMBARRIER_CMD_QUERY which is assigned to |
| 62 | * the value 0. | 70 | * the value 0. |
| 63 | */ | 71 | */ |
| 64 | enum membarrier_cmd { | 72 | enum membarrier_cmd { |
| 65 | MEMBARRIER_CMD_QUERY = 0, | 73 | MEMBARRIER_CMD_QUERY = 0, |
| 66 | MEMBARRIER_CMD_SHARED = (1 << 0), | 74 | MEMBARRIER_CMD_SHARED = (1 << 0), |
| 67 | /* reserved for MEMBARRIER_CMD_SHARED_EXPEDITED (1 << 1) */ | 75 | /* reserved for MEMBARRIER_CMD_SHARED_EXPEDITED (1 << 1) */ |
| 68 | /* reserved for MEMBARRIER_CMD_PRIVATE (1 << 2) */ | 76 | /* reserved for MEMBARRIER_CMD_PRIVATE (1 << 2) */ |
| 69 | MEMBARRIER_CMD_PRIVATE_EXPEDITED = (1 << 3), | 77 | MEMBARRIER_CMD_PRIVATE_EXPEDITED = (1 << 3), |
| 78 | MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED = (1 << 4), | ||
| 70 | }; | 79 | }; |
| 71 | 80 | ||
| 72 | #endif /* _UAPI_LINUX_MEMBARRIER_H */ | 81 | #endif /* _UAPI_LINUX_MEMBARRIER_H */ |