[PATCH] efidebug: avoid 'dfu_alt_info not defined' message

[AKASHI Takahiro]

On Mon, Jan 17, 2022 at 01:33:51PM +0100, Heinrich Schuchardt wrote:
> On 1/17/22 01:45, AKASHI Takahiro wrote:
> > On Sat, Jan 15, 2022 at 02:18:21AM +0100, Heinrich Schuchardt wrote:
> > > If variable dfu_alt_info is not defined duplicate messages are displayed.
> > > 
> > >      => efidebug boot dump
> > >      Scanning disk mmc2.blk...
> > >      Scanning disk mmc1.blk...
> > >      Scanning disk mmc0.blk...
> > >      Found 3 disks
> > >      No EFI system partition
> > >      "dfu_alt_info" env variable not defined!
> > >      Probably dfu_alt_info not defined
> > >      "dfu_alt_info" env variable not defined!
> > >      Probably dfu_alt_info not defined
> > > 
> > > Remove the 'Probably dfu_alt_info not defined' message.
> > > Instead write a warning if the variable contains no entities.
> > > 
> > > Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt at canonical.com>
> > > ---
> > >   lib/efi_loader/efi_firmware.c | 7 +++++--
> > >   1 file changed, 5 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/lib/efi_loader/efi_firmware.c b/lib/efi_loader/efi_firmware.c
> > > index a1b88dbfc2..519a47267c 100644
> > > --- a/lib/efi_loader/efi_firmware.c
> > > +++ b/lib/efi_loader/efi_firmware.c
> > > @@ -128,8 +128,11 @@ static efi_status_t efi_get_dfu_info(
> > >   	size_t names_len, total_size;
> > >   	int dfu_num, i;
> > >   	u16 *name, *next;
> > > +	int ret;
> > > -	dfu_init_env_entities(NULL, NULL);
> > > +	ret = dfu_init_env_entities(NULL, NULL);
> > > +	if (ret)
> > > +		return EFI_SUCCESS;
> > 
> > Do you want to return EFI_SUCCESS here?
> > It is mandatory that "dfu_alt_info" be defined so that the current
> > FMP drivers should work.
> > So if the variable is not defined, I think,
> > efi_firmware_[fit|raw]_get_image_info() should fail.
> > Even if it returns EFI_SUCCESS, descriptor_count should be set to zero.
> 
> This patch did not change the return value. It only removed superfluous
> messages that I saw everytime I used the sandbox.

The return value is the same, but

> For dfu_num == 0 we are returning EFI_SUCCESS. If the number is 0 due to the
> content of dfu_alt_info or due to its absence, should not make a difference
> in the return code. Writing a message if dfu_alt_info exists but contains no
> entity seems reasonable.

With your patch, the value of "descriptor_count", as well as the other
arguments, will remain uncertain even we see EFI_SUCCESS at return.

Since dfu_init_env_entities() returns -EINVAL in case "dfu_alt_info" is
not defined, efi_get_dfu_info(), hence efi_firmware_[fit|raw]_get_image_info(),
returns immediately without setting any value to "descriptor_count".

Have you ever tested your code?

> If you think EFI_SUCCESS is not correct in both cases, please, send a
> followup patch for review.

NAK.
We are discussing the issue against your patch.
Why not do so in this thread?

-Takahiro Akashi


> Best regards
> 
> Heinrich
> 
> > 
> > -Takahiro Akashi
> > 
> > >   	names_len = 0;
> > >   	dfu_num = 0;
> > > @@ -138,7 +141,7 @@ static efi_status_t efi_get_dfu_info(
> > >   		dfu_num++;
> > >   	}
> > >   	if (!dfu_num) {
> > > -		log_warning("Probably dfu_alt_info not defined\n");
> > > +		log_warning("No entities in dfu_alt_info\n");
> > >   		*image_info_size = 0;
> > >   		dfu_free_entities();
> > > -- 
> > > 2.33.1
> > > 
>