[PATCH v2 1/3] grub-editenv: Verify the variable size

2019-01-16 Thread Prarit Bhargava
It is possible to execute 'grub-editenv - set ="some data"', which results in an unremoveable entry ="some data" Verify the variable has a size before setting a value. Signed-off-by: Prarit Bhargava Cc: mleit...@redhat.com Cc: pjo...@redhat.com Cc: javi...@redhat.com Cc: ar...@redhat.com Cc: Da

[PATCH v2 3/3] grub: add extend variable functionality

2019-01-16 Thread Prarit Bhargava
Customers and users of the kernel are commenting that there is no way to update a grub variable without copy and pasting the existing data. For example, [10:57 AM root@intel-wildcatpass-07 grub-2.02]# ./grub-editenv list saved_entry=0 next_entry= kernelopts=root=/dev/mapper/rhel_intel--wildcatpas

[PATCH v2 0/3] grub: add grub variable extend functionality

2019-01-16 Thread Prarit Bhargava
This patchset allows a user to update an existing grub variable on the commandline. v2: - Split into 3 patches. - fix "=value" bug. An additional consequence of this is that the value pointer p in set_variable() cannot underflow. - make grub_envblk_iterate return an in

[PATCH v2 2/3] grub: Make grub_envblk_iterate() return an int

2019-01-16 Thread Prarit Bhargava
grub_envblk_iterate() is a void. Future functions will require the ability to interpret return codes from the iteration, so grub_envblk_iterate() should be an int. The value of 0 returned from the hook functions is overloaded and cannot be parsed by callers to grub_envblk_iterate() for error hand

Re: [PATCH 1/1] Upgrade Gnulib; switch to bootstrap tool

2019-01-16 Thread Daniel Kiper
On Wed, Jan 09, 2019 at 12:29:30PM +, Colin Watson wrote: > Upgrade Gnulib files to 20190105. > > It's much easier to maintain GRUB's use of portability support files > from Gnulib when the process is automatic and driven by a single > configuration file, rather than by maintainers occasionally

Re: GRUB Xen PVH chainloading

2019-01-16 Thread Daniel Kiper
On Mon, Jan 07, 2019 at 01:43:11PM +0100, Juergen Gross wrote: > On 07/01/2019 12:41, Colin Watson wrote: > > Hi, > > > > I'm working on integrating the recently-merged PVH support for GRUB into > > the Debian packages. As a result I find myself thinking about how to > > handle the two-stage boot

Re: MKPasswd idempotency patch

2019-01-16 Thread Daniel Kiper
On Tue, Dec 25, 2018 at 12:01:12PM +0100, grub-de...@agowa338.de wrote: > Hello, > > I want to use grub-mkpasswd inside of a ansible playbook, therefore a > idempotency or check feature is required. Without it, there is no > posibility to verify, that the configuration is already inplace. > Therefo

Re: [PATCH v2 1/3] grub-editenv: Verify the variable size

2019-01-16 Thread Daniel Kiper
On Wed, Jan 16, 2019 at 01:34:41PM -0500, Prarit Bhargava wrote: > It is possible to execute 'grub-editenv - set ="some data"', > which results in an unremoveable entry > > ="some data" Please explain why it happens. > Verify the variable has a size before setting a value. s/variable/argument/ s

Re: [PATCH v2 3/3] grub: add extend variable functionality

2019-01-16 Thread Daniel Kiper
On Wed, Jan 16, 2019 at 01:34:43PM -0500, Prarit Bhargava wrote: > Customers and users of the kernel are commenting that there is no way to > update > a grub variable without copy and pasting the existing data. > > For example, > > [10:57 AM root@intel-wildcatpass-07 grub-2.02]# ./grub-editenv lis

Re: [PATCH v2 2/3] grub: Make grub_envblk_iterate() return an int

2019-01-16 Thread Daniel Kiper
On Wed, Jan 16, 2019 at 01:34:42PM -0500, Prarit Bhargava wrote: > grub_envblk_iterate() is a void. Future functions will require the > ability to interpret return codes from the iteration, so > grub_envblk_iterate() should be an int. > > The value of 0 returned from the hook functions is overload

Re: [PATCH] uboot: Add the missing disk write operation support

2019-01-16 Thread Daniel Kiper
On Tue, Jan 15, 2019 at 04:59:56PM +0200, Cristian Ciocaltea wrote: > uboot_disk_write() is currently lacking the write support > to storage devices because, historically, those devices did not > implement block_write() in U-Boot. > > The solution has been tested using a patched U-Boot loading > an