From mboxrd@z Thu Jan 1 00:00:00 1970 From: AKASHI Takahiro Date: Wed, 27 Feb 2019 15:47:52 +0900 Subject: [U-Boot] [PATCH v2 4/5] cmd: bootefi: run an EFI application of a specific load option In-Reply-To: <3d7ae831-954a-4fa3-8232-1723059c7f32@gmx.de> References: <20190115025437.11966-1-takahiro.akashi@linaro.org> <20190115025437.11966-5-takahiro.akashi@linaro.org> <97e4c587-2c18-7d5f-757d-1673aa981fd1@gmx.de> <20190227055836.GC20286@linaro.org> <3d7ae831-954a-4fa3-8232-1723059c7f32@gmx.de> Message-ID: <20190227064750.GG20286@linaro.org> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de On Wed, Feb 27, 2019 at 07:31:06AM +0100, Heinrich Schuchardt wrote: > On 2/27/19 6:58 AM, AKASHI Takahiro wrote: > > On Tue, Feb 26, 2019 at 08:30:50PM +0100, Heinrich Schuchardt wrote: > >> On 1/15/19 3:54 AM, AKASHI Takahiro wrote: > >>> With this patch applied, we will be able to selectively execute > >>> an EFI application by specifying a load option, say "1" for Boot0001, > >>> "2" for Boot0002 and so on. > >>> > >>> => bootefi bootmgr 1, or > >>> bootefi bootmgr - 1 > >> > >> You already introduced the support for BootNext. So is there a real benefit? > > > > This is a convenient way of running EFI application directly, > > but I already removed this feature from the next version. > > Please, remove 'run -e' instead because it cannot specify the device > tree needed for booting ARM boards. See my comment for patch#5 first. > > > >>> > >>> Please note that BootXXXX need not be included in "BootOrder". > >>> > >>> Signed-off-by: AKASHI Takahiro > >>> --- > >>> cmd/bootefi.c | 39 ++++++++++++++++++++++++++++----------- > >>> 1 file changed, 28 insertions(+), 11 deletions(-) > >>> > >>> diff --git a/cmd/bootefi.c b/cmd/bootefi.c > >>> index 3be01b49b589..241fd0f987ab 100644 > >>> --- a/cmd/bootefi.c > >>> +++ b/cmd/bootefi.c > >>> @@ -471,16 +471,15 @@ static efi_status_t bootefi_test_prepare > >>> > >>> #endif /* CONFIG_CMD_BOOTEFI_SELFTEST */ > >>> > >>> -static int do_bootefi_bootmgr_exec(void) > >>> +static int do_bootefi_bootmgr_exec(int boot_id) > >>> { > >>> struct efi_device_path *device_path, *file_path; > >>> void *addr; > >>> efi_status_t r; > >>> > >>> - addr = efi_bootmgr_load(EFI_BOOTMGR_DEFAULT_ORDER, > >>> - &device_path, &file_path); > >>> + addr = efi_bootmgr_load(boot_id, &device_path, &file_path); > >>> if (!addr) > >>> - return 1; > >>> + return CMD_RET_FAILURE; > >>> > >>> printf("## Starting EFI application at %p ...\n", addr); > >>> r = do_bootefi_exec(addr, device_path, file_path); > >>> @@ -488,9 +487,9 @@ static int do_bootefi_bootmgr_exec(void) > >>> r & ~EFI_ERROR_MASK); > >>> > >>> if (r != EFI_SUCCESS) > >>> - return 1; > >>> + return CMD_RET_FAILURE; > >>> > >>> - return 0; > >>> + return CMD_RET_SUCCESS; > >>> } > >>> > >>> /* Interpreter command to boot an arbitrary EFI image from memory */ > >>> @@ -546,10 +545,28 @@ static int do_bootefi(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) > >>> } else > >>> #endif > >>> if (!strcmp(argv[1], "bootmgr")) { > >>> - if (efi_handle_fdt(argc > 2 ? argv[2] : NULL)) > >>> - return CMD_RET_FAILURE; > >>> + char *fdtstr, *endp; > >>> + int boot_id = EFI_BOOTMGR_DEFAULT_ORDER; > >>> + > >>> + if (argc > 2) { > >>> + fdtstr = argv[2]; > >>> + /* Special address "-" means no device tree */ > >>> + if (fdtstr[0] == '-') > >>> + fdtstr = NULL; > >>> + > >>> + r = efi_handle_fdt(fdtstr); > >>> + if (r) > >>> + return CMD_RET_FAILURE; > >>> + } > >>> + > >>> + if (argc > 3) { > >>> + boot_id = (int)simple_strtoul(argv[3], &endp, 0); > >>> + if ((argv[3] + strlen(argv[3]) != endp) || > >>> + boot_id > 0xffff) > >>> + return CMD_RET_USAGE; > >>> + } > >>> > >>> - return do_bootefi_bootmgr_exec(); > >>> + return do_bootefi_bootmgr_exec(boot_id); > >> > >> Why not communicate via the BootNext variable? > > > > I don't get your point. > > BootNext and BootOrder are both defined by UEFI specification. > > Instead of changing the interface of do_bootefi_bootmgr_exec() Who care changing an *internal* function? > you could > simply set BootNext. Then the boot manager would pick up the option from > the variable and finally delete the variable. This would result in less > code. No. Even with "run -e," BootNext will disappear after execution. This is a requirement by UEFI spec. Thanks, -Takahiro Akashi > Best regards > > Heinrich > > > > >>> } else { > >>> saddr = argv[1]; > >>> > >>> @@ -590,7 +607,7 @@ static char bootefi_help_text[] = > >>> " Use environment variable efi_selftest to select a single test.\n" > >>> " Use 'setenv efi_selftest list' to enumerate all tests.\n" > >>> #endif > >>> - "bootefi bootmgr [fdt addr]\n" > >>> + "bootefi bootmgr [|'-' []]\n" > >>> " - load and boot EFI payload based on BootOrder/BootXXXX variables.\n" > >>> "\n" > >>> " If specified, the device tree located at gets\n" > >>> @@ -598,7 +615,7 @@ static char bootefi_help_text[] = > >>> #endif > >>> > >>> U_BOOT_CMD( > >>> - bootefi, 3, 0, do_bootefi, > >>> + bootefi, 5, 0, do_bootefi, > >> > >> Why 5? > > > > For additional/optional '-' and . > > But I removed this feature from bootefi. > > > > Thanks, > > -Takahiro Akashi > > > > > >> Best regards > >> > >> Heinrich > >> > >>> "Boots an EFI payload from memory", > >>> bootefi_help_text > >>> ); > >>> > >> > > >