lpfc.rst (3221B)
1.. SPDX-License-Identifier: GPL-2.0 2 3========================= 4LPFC Driver Release Notes 5========================= 6 7 8.. important:: 9 10 Starting in the 8.0.17 release, the driver began to be targeted strictly 11 toward the upstream kernel. As such, we removed #ifdefs for older kernels 12 (pre 2.6.10). The 8.0.16 release should be used if the driver is to be 13 run on one of the older kernels. 14 15 The proposed modifications to the transport layer for FC remote ports 16 and extended attribute support is now part of the upstream kernel 17 as of 2.6.12. We no longer need to provide patches for this support, 18 nor a *full* version which has old an new kernel support. 19 20 The driver now requires a 2.6.12 (if pre-release, 2.6.12-rc1) or later 21 kernel. 22 23 Please heed these dependencies.... 24 25 26The following information is provided for additional background on the 27history of the driver as we push for upstream acceptance. 28 29Cable pull and temporary device Loss: 30 31 In older revisions of the lpfc driver, the driver internally queued i/o 32 received from the midlayer. In the cases where a cable was pulled, link 33 jitter, or a device temporarily loses connectivity (due to its cable 34 being removed, a switch rebooting, or a device reboot), the driver could 35 hide the disappearance of the device from the midlayer. I/O's issued to 36 the LLDD would simply be queued for a short duration, allowing the device 37 to reappear or link come back alive, with no inadvertent side effects 38 to the system. If the driver did not hide these conditions, i/o would be 39 errored by the driver, the mid-layer would exhaust its retries, and the 40 device would be taken offline. Manual intervention would be required to 41 re-enable the device. 42 43 The community supporting kernel.org has driven an effort to remove 44 internal queuing from all LLDDs. The philosophy is that internal 45 queuing is unnecessary as the block layer already performs the 46 queuing. Removing the queues from the LLDD makes a more predictable 47 and more simple LLDD. 48 49 As a potential new addition to kernel.org, the 8.x driver was asked to 50 have all internal queuing removed. Emulex complied with this request. 51 In explaining the impacts of this change, Emulex has worked with the 52 community in modifying the behavior of the SCSI midlayer so that SCSI 53 devices can be temporarily suspended while transport events (such as 54 those described) can occur. 55 56 The proposed patch was posted to the linux-scsi mailing list. The patch 57 is contained in the 2.6.10-rc2 (and later) patch kits. As such, this 58 patch is part of the standard 2.6.10 kernel. 59 60 By default, the driver expects the patches for block/unblock interfaces 61 to be present in the kernel. No #define needs to be set to enable support. 62 63 64Kernel Support 65============== 66 67 This source package is targeted for the upstream kernel only. (See notes 68 at the top of this file). It relies on interfaces that are slowing 69 migrating into the kernel.org kernel. 70 71 At this time, the driver requires the 2.6.12 (if pre-release, 2.6.12-rc1) 72 kernel. 73 74 If a driver is needed for older kernels please utilize the 8.0.16 75 driver sources. 76 77 78Patches 79======= 80 81 Thankfully, at this time, patches are not needed.