Re: [PATCH v2 0/6] Chainsaw efivars.c
On 04/17/2013 01:30 AM, Tom Gundersen wrote: > Hi Matt, > > On Tue, Apr 16, 2013 at 11:41 PM, Matt Fleming wrote: >> This patch series introduces the new efivar_entry API, and splits out the >> major >> parts of efivars.c into new files. In particular, having the efivarfs code >> under fs/ allows building an efivarfs.ko module, which means mount(8) can >> automatically load it. >> >> The remaining EFI code is repositioned under drivers/firmware/efi/. > > These all look good to me, so feel free to add my Reviewed-by. > > Tested on top of 3.9-rc7, with everything built as modules. Everything > works as intended: systemd correctly detects it is on an EFI system, > so mounts efivarfs, which auto-loads the required modules. > > Thanks a lot! Thanks for reviewing and testing! I've added your Reviewed-by and Tested-by tags to all 5 patches (but not the one you authored). I found a bug in PATCH 3, so I've sent out a v3 of the series. Peter, I've pulled these 6 patches into the 'chainsaw' branch in the EFI repo. I think the prudent thing to do (as you previously suggested) is to merge that branch into tip as a separate topic branch, just in case we need to drop it before the merge window opens. -- Matt Fleming, Intel Open Source Technology Center -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2 0/6] Chainsaw efivars.c
On 04/17/2013 01:30 AM, Tom Gundersen wrote: Hi Matt, On Tue, Apr 16, 2013 at 11:41 PM, Matt Fleming m...@console-pimps.org wrote: This patch series introduces the new efivar_entry API, and splits out the major parts of efivars.c into new files. In particular, having the efivarfs code under fs/ allows building an efivarfs.ko module, which means mount(8) can automatically load it. The remaining EFI code is repositioned under drivers/firmware/efi/. These all look good to me, so feel free to add my Reviewed-by. Tested on top of 3.9-rc7, with everything built as modules. Everything works as intended: systemd correctly detects it is on an EFI system, so mounts efivarfs, which auto-loads the required modules. Thanks a lot! Thanks for reviewing and testing! I've added your Reviewed-by and Tested-by tags to all 5 patches (but not the one you authored). I found a bug in PATCH 3, so I've sent out a v3 of the series. Peter, I've pulled these 6 patches into the 'chainsaw' branch in the EFI repo. I think the prudent thing to do (as you previously suggested) is to merge that branch into tip as a separate topic branch, just in case we need to drop it before the merge window opens. -- Matt Fleming, Intel Open Source Technology Center -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2 0/6] Chainsaw efivars.c
Hi Matt, On Tue, Apr 16, 2013 at 11:41 PM, Matt Fleming wrote: > This patch series introduces the new efivar_entry API, and splits out the > major > parts of efivars.c into new files. In particular, having the efivarfs code > under fs/ allows building an efivarfs.ko module, which means mount(8) can > automatically load it. > > The remaining EFI code is repositioned under drivers/firmware/efi/. These all look good to me, so feel free to add my Reviewed-by. Tested on top of 3.9-rc7, with everything built as modules. Everything works as intended: systemd correctly detects it is on an EFI system, so mounts efivarfs, which auto-loads the required modules. Thanks a lot! Tom -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[PATCH v2 0/6] Chainsaw efivars.c
From: Matt Fleming drivers/firmware/efivars.c has grown pretty large and is ~2K lines. Inside efivars.c there's currently, o code for handling EFI variables at the firmware-level o sysfs code for exposing EFI variables o a new EFI variable filesystem o a persistent storage backend all intertwined and smushed together. This situation is only going to get worse as new EFI support is added. We need an interface that hides the EFI variable operations in use so code isn't tempted to access them directly, e.g. efivarfs currently uses '__efivars' which means it doesn't work for CONFIG_GOOGLE_SMI as that uses different variable ops. With this interface in place, we can start moving independent code out into separate files, allowing users to only turn on the functionality that they want. This patch series introduces the new efivar_entry API, and splits out the major parts of efivars.c into new files. In particular, having the efivarfs code under fs/ allows building an efivarfs.ko module, which means mount(8) can automatically load it. The remaining EFI code is repositioned under drivers/firmware/efi/. The series is also available on the 'chainsaw' branch at, git://git.kernel.org/pub/scm/linux/kernel/git/mfleming/linux.git Changes in v2: - Add Reviewed-by/Acked-by tags from Mike - Allow both EFI variable ops to be built as requested by Mike (PATCH 2) - Fix deadlock and remove unnecessary logic changes in the pstore patch reported by Seiji (PATCH 3) Matt Fleming (5): efi: move utf16 string functions to efi.h efivars: Keep a private global pointer to efivars efivars: efivar_entry API efivars: Move pstore code into the new EFI directory efivarfs: Move to fs/efivarfs Tom Gundersen (1): efi: split efisubsystem from efivars MAINTAINERS | 13 +- drivers/firmware/Kconfig | 36 +- drivers/firmware/Makefile |2 +- drivers/firmware/efi/Kconfig | 39 + drivers/firmware/efi/Makefile |6 + drivers/firmware/efi/efi-pstore.c | 244 + drivers/firmware/efi/efi.c| 134 +++ drivers/firmware/efi/efivars.c| 615 +++ drivers/firmware/efi/vars.c | 1046 ++ drivers/firmware/efivars.c| 2171 - drivers/firmware/google/gsmi.c| 30 +- fs/Kconfig|1 + fs/Makefile |1 + fs/efivarfs/Kconfig | 12 + fs/efivarfs/Makefile |7 + fs/efivarfs/file.c| 111 ++ fs/efivarfs/inode.c | 173 +++ fs/efivarfs/internal.h| 22 + fs/efivarfs/super.c | 266 + include/linux/efi.h | 133 ++- 20 files changed, 2828 insertions(+), 2234 deletions(-) create mode 100644 drivers/firmware/efi/Kconfig create mode 100644 drivers/firmware/efi/Makefile create mode 100644 drivers/firmware/efi/efi-pstore.c create mode 100644 drivers/firmware/efi/efi.c create mode 100644 drivers/firmware/efi/efivars.c create mode 100644 drivers/firmware/efi/vars.c delete mode 100644 drivers/firmware/efivars.c create mode 100644 fs/efivarfs/Kconfig create mode 100644 fs/efivarfs/Makefile create mode 100644 fs/efivarfs/file.c create mode 100644 fs/efivarfs/inode.c create mode 100644 fs/efivarfs/internal.h create mode 100644 fs/efivarfs/super.c -- 1.7.10.4 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[PATCH v2 0/6] Chainsaw efivars.c
From: Matt Fleming matt.flem...@intel.com drivers/firmware/efivars.c has grown pretty large and is ~2K lines. Inside efivars.c there's currently, o code for handling EFI variables at the firmware-level o sysfs code for exposing EFI variables o a new EFI variable filesystem o a persistent storage backend all intertwined and smushed together. This situation is only going to get worse as new EFI support is added. We need an interface that hides the EFI variable operations in use so code isn't tempted to access them directly, e.g. efivarfs currently uses '__efivars' which means it doesn't work for CONFIG_GOOGLE_SMI as that uses different variable ops. With this interface in place, we can start moving independent code out into separate files, allowing users to only turn on the functionality that they want. This patch series introduces the new efivar_entry API, and splits out the major parts of efivars.c into new files. In particular, having the efivarfs code under fs/ allows building an efivarfs.ko module, which means mount(8) can automatically load it. The remaining EFI code is repositioned under drivers/firmware/efi/. The series is also available on the 'chainsaw' branch at, git://git.kernel.org/pub/scm/linux/kernel/git/mfleming/linux.git Changes in v2: - Add Reviewed-by/Acked-by tags from Mike - Allow both EFI variable ops to be built as requested by Mike (PATCH 2) - Fix deadlock and remove unnecessary logic changes in the pstore patch reported by Seiji (PATCH 3) Matt Fleming (5): efi: move utf16 string functions to efi.h efivars: Keep a private global pointer to efivars efivars: efivar_entry API efivars: Move pstore code into the new EFI directory efivarfs: Move to fs/efivarfs Tom Gundersen (1): efi: split efisubsystem from efivars MAINTAINERS | 13 +- drivers/firmware/Kconfig | 36 +- drivers/firmware/Makefile |2 +- drivers/firmware/efi/Kconfig | 39 + drivers/firmware/efi/Makefile |6 + drivers/firmware/efi/efi-pstore.c | 244 + drivers/firmware/efi/efi.c| 134 +++ drivers/firmware/efi/efivars.c| 615 +++ drivers/firmware/efi/vars.c | 1046 ++ drivers/firmware/efivars.c| 2171 - drivers/firmware/google/gsmi.c| 30 +- fs/Kconfig|1 + fs/Makefile |1 + fs/efivarfs/Kconfig | 12 + fs/efivarfs/Makefile |7 + fs/efivarfs/file.c| 111 ++ fs/efivarfs/inode.c | 173 +++ fs/efivarfs/internal.h| 22 + fs/efivarfs/super.c | 266 + include/linux/efi.h | 133 ++- 20 files changed, 2828 insertions(+), 2234 deletions(-) create mode 100644 drivers/firmware/efi/Kconfig create mode 100644 drivers/firmware/efi/Makefile create mode 100644 drivers/firmware/efi/efi-pstore.c create mode 100644 drivers/firmware/efi/efi.c create mode 100644 drivers/firmware/efi/efivars.c create mode 100644 drivers/firmware/efi/vars.c delete mode 100644 drivers/firmware/efivars.c create mode 100644 fs/efivarfs/Kconfig create mode 100644 fs/efivarfs/Makefile create mode 100644 fs/efivarfs/file.c create mode 100644 fs/efivarfs/inode.c create mode 100644 fs/efivarfs/internal.h create mode 100644 fs/efivarfs/super.c -- 1.7.10.4 -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2 0/6] Chainsaw efivars.c
Hi Matt, On Tue, Apr 16, 2013 at 11:41 PM, Matt Fleming m...@console-pimps.org wrote: This patch series introduces the new efivar_entry API, and splits out the major parts of efivars.c into new files. In particular, having the efivarfs code under fs/ allows building an efivarfs.ko module, which means mount(8) can automatically load it. The remaining EFI code is repositioned under drivers/firmware/efi/. These all look good to me, so feel free to add my Reviewed-by. Tested on top of 3.9-rc7, with everything built as modules. Everything works as intended: systemd correctly detects it is on an EFI system, so mounts efivarfs, which auto-loads the required modules. Thanks a lot! Tom -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
