From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ilias Apalodimas Date: Fri, 19 Feb 2021 14:02:52 +0200 Subject: [PATCH 1/2] efi_loader: multi part device paths to text In-Reply-To: <21e27436-649b-8ba8-1755-4050c19719ae@gmx.de> References: <20210218173044.127456-1-xypron.glpk@gmx.de> <20210218173044.127456-2-xypron.glpk@gmx.de> <21e27436-649b-8ba8-1755-4050c19719ae@gmx.de> Message-ID: List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de On Fri, Feb 19, 2021 at 12:45:38PM +0100, Heinrich Schuchardt wrote: > On 19.02.21 11:13, Ilias Apalodimas wrote: > > Hi Heinrich > > > > On Thu, Feb 18, 2021 at 06:30:43PM +0100, Heinrich Schuchardt wrote: > >> Our current implementation of > >> EFI_DEVICE_PATH_TO_TEXT_PROTOCOL.ConvertDevicePathToText() truncates multi > >> part device paths after the first part. We should convert all parts. > >> > >> Render device path instance ends as commas. This is not explicitly > >> described in the UEFI spec but mimics what EDK II does. > >> > >> Signed-off-by: Heinrich Schuchardt > >> --- > >> lib/efi_loader/efi_device_path_to_text.c | 17 ++++++++++++----- > >> 1 file changed, 12 insertions(+), 5 deletions(-) > >> > >> diff --git a/lib/efi_loader/efi_device_path_to_text.c b/lib/efi_loader/efi_device_path_to_text.c > >> index 1aaa9f94fa..81b8ac23ba 100644 > >> --- a/lib/efi_loader/efi_device_path_to_text.c > >> +++ b/lib/efi_loader/efi_device_path_to_text.c > >> @@ -369,11 +369,18 @@ static uint16_t EFIAPI *efi_convert_device_path_to_text( > >> > >> if (!device_path) > >> goto out; > >> - while (device_path && > >> - str + MAX_NODE_LEN < buffer + MAX_PATH_LEN) { > >> - *str++ = '/'; > >> - str = efi_convert_single_device_node_to_text(str, device_path); > >> - device_path = efi_dp_next(device_path); > >> + while (device_path && str + MAX_NODE_LEN < buffer + MAX_PATH_LEN) { > >> + if (device_path->type == DEVICE_PATH_TYPE_END) { > >> + if (device_path->sub_type != > >> + DEVICE_PATH_SUB_TYPE_INSTANCE_END) > > > > maybe it's better to check explicitly for '== DEVICE_PATH_SUB_TYPE_END'? > > Thank you for reviewing. > > I want to leave the loop when an unexpected value occurs. Fair enough Reviewed-by: Ilias Apalodimas > > Best regards > > Heinrich > > > > >> + break; > >> + *str++ = ','; > >> + } else { > >> + *str++ = '/'; > >> + str = efi_convert_single_device_node_to_text( > >> + str, device_path); > >> + } > >> + *(u8 **)&device_path += device_path->length; > >> } > >> > >> text = efi_str_to_u16(buffer); > >> -- > >> 2.30.0 > >> > > > > Cheers > > /Ilias > > >