From: Ahmad Fatoum <a.fatoum@pengutronix.de>
To: Fabian Pflug <f.pflug@pengutronix.de>, barebox@lists.infradead.org
Subject: Re: [PATCH 3/4] bootm: use new api to get kernel command line params
Date: Wed, 26 Nov 2025 12:33:20 +0100 [thread overview]
Message-ID: <9c9ad61d-295f-4011-9418-dc2ed8da16a2@pengutronix.de> (raw)
In-Reply-To: <53f649d5cc950aaaadbfae828244e872b3d0688c.camel@pengutronix.de>
Hi,
On 11/26/25 12:29 PM, Fabian Pflug wrote:
> Hey :)
>
> On Wed, 2025-11-26 at 11:25 +0100, Ahmad Fatoum wrote:
>>
>> On 11/26/25 7:42 AM, Fabian Pflug wrote:
>>> Getting root device adn root kernel options seperatily any combining
>>
>> device and root kernel options seperatly and combining
>>
>>> them in this function instead of getting the full string from the helper
>>> functions. This is in preperation for replacing the root= string in the
>>
>> preparation
>>
>>> kernel commandline with something defined during runtime.
>>>
>>> Signed-off-by: Fabian Pflug <f.pflug@pengutronix.de>
>>> ---
>>> common/bootm.c | 34 ++++++++++++++++++++--------------
>>> 1 file changed, 20 insertions(+), 14 deletions(-)
>>>
>>> diff --git a/common/bootm.c b/common/bootm.c
>>> index 17792b2a1d..5d7dc00e3e 100644
>>> --- a/common/bootm.c
>>> +++ b/common/bootm.c
>>> @@ -4,6 +4,7 @@
>>> #include <bootm.h>
>>> #include <bootm-overrides.h>
>>> #include <fs.h>
>>> +#include <fcntl.h>
>>> #include <malloc.h>
>>> #include <memory.h>
>>> #include <block.h>
>>> @@ -841,33 +842,38 @@ int bootm_boot(struct bootm_data *bootm_data)
>>> }
>>>
>>> if (bootm_data->appendroot) {
>>> - char *rootarg;
>>> + char *root;
>>> + char *rootopts;
>>>
>>> if (bootm_data->root_dev) {
>>> const char *root_dev_name = devpath_to_name(bootm_data->root_dev);
>>> struct cdev *root_cdev = cdev_open_by_name(root_dev_name, O_RDONLY);
>>>
>>> - rootarg = cdev_get_linux_rootarg(root_cdev);
>>> - if (!rootarg) {
>>> - rootarg = ERR_PTR(-EINVAL);
>>> -
>>> - if (!root_cdev)
>>> - pr_err("no cdev found for %s, cannot set root= option\n",
>>> - root_dev_name);
>>> - else if (!root_cdev->partuuid[0])
>>> - pr_err("%s doesn't have a PARTUUID, cannot set root= option\n",
>>> - root_dev_name);
>>
>> This warning is useful, because it's very confusing otherwise when a MBR
>> disk image is created without a NT signature.
>>
>>> - }
>>> + cdev_get_linux_root_and_opts(root_cdev, &root, &rootopts);
>>>
>>> if (root_cdev)
>>> cdev_close(root_cdev);
>>> + else
>>> + pr_err("no cdev found for %s, cannot set root= option\n", root_dev_name);
>>> } else {
>>> - rootarg = path_get_linux_rootarg(data->os_file);
>>
>> path_get_linux_rootarg() is stale now right? Can you squash its removal
>> into the patch that gets rid of linux.bootargs?
>
> Then we would have a case, where the function is used, but not available anymore resulting in an unbuildable system,
> which I would like to avoid. I can remove it in an extra commit after this one.
You will have of course to change the users not to call it anymore.
If you add a later commit removing it, you still have an intermittent
breakage that would trip up a bisect or am I missing something?
>
>>
>>> + struct fs_device *fsdev = get_fsdevice_by_path(AT_FDCWD, data->os_file);
>>> + if(fsdev)
>>> + fsdev_get_linux_root_options(fsdev, &root, &rootopts);
>>> + else
>>> + pr_err("no fsdevice under path: %s\n", data->os_file);
>>> }
>>>
>>> - if (IS_ERR(rootarg)) {
>>> + if (!root) {
>>> pr_err("Failed to append kernel cmdline parameter 'root='\n");
>>> } else {
>>> + char *rootarg;
>>> + if(rootopts) {
>>> + rootarg = xasprintf("root=%s %s", root, rootopts);
>>> + free(rootopts);
>>> + } else {
>>> + rootarg = xasprintf("root=%s", root);
>>> + }
>>
>> I wonder if we should add a helper into a header that does this, given
>> how many times this is repeated...
>>
>> char *format_root_bootarg(char *root, char *rootopts);
>>
>> which consumes (and frees) two pointers and returns a new buffer..
>
> Good thinking. But where do we add it?
> * fs.h
> * block.h
> * bootargs.h
bootargs.h sounds suitable.
>
> And also, do we add a third parameter? because I will re replaced in the next commit with "%s=%s %s"
Noticed that too when I read that far. Ye, three arguments sounds good
to me.
> And since you proposed to get rid of path_get_linux_rootarg it will be used with two parameters in one place
> (cdev_get_linux_rootarg) and with three at the other. (bootm_boot)
Can't follow, but I might when I see the code :-)
Cheers,
Ahmad
>
> Kind regard
> Fabian
>
>
>>
>> Cheers,
>> Ahmad
>>
>>> + free(root);
>>> pr_info("Adding \"%s\" to Kernel commandline\n", rootarg);
>>> globalvar_add_simple("linux.bootargs.bootm.appendroot",
>>> rootarg);
>
--
Pengutronix e.K. | |
Steuerwalder Str. 21 | http://www.pengutronix.de/ |
31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
next prev parent reply other threads:[~2025-11-26 11:33 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-24 10:16 [PATCH] common: setting the root= command line parameter Fabian Pflug
2025-11-24 12:00 ` Ahmad Fatoum
2025-11-25 10:22 ` Fabian Pflug
2025-11-25 19:13 ` Ahmad Fatoum
2025-11-26 6:42 ` [PATCH 0/4] make the root= command line parameter variable Fabian Pflug
2025-11-26 6:42 ` [PATCH 1/4] block.h: renamed get_rootargs to get_root Fabian Pflug
2025-11-26 9:50 ` Ahmad Fatoum
2025-11-26 6:42 ` [PATCH 2/4] fs: split rootargs into root and options Fabian Pflug
2025-11-26 10:13 ` Ahmad Fatoum
2025-11-26 10:31 ` Ahmad Fatoum
2025-11-26 6:42 ` [PATCH 3/4] bootm: use new api to get kernel command line params Fabian Pflug
2025-11-26 10:25 ` Ahmad Fatoum
2025-11-26 11:29 ` Fabian Pflug
2025-11-26 11:33 ` Ahmad Fatoum [this message]
2025-11-26 6:42 ` [PATCH 4/4] bootm: introduce bootm.root_arg variable Fabian Pflug
2025-11-26 10:28 ` Ahmad Fatoum
2025-11-27 10:57 ` [PATCH v2 0/4] make the root= command line parameter variable Fabian Pflug
2025-11-27 10:57 ` [PATCH v2 1/4] block.h: renamed get_rootargs to get_root Fabian Pflug
2025-11-27 11:08 ` Ahmad Fatoum
2025-11-27 10:57 ` [PATCH v2 2/4] fs: split rootargs into root and options Fabian Pflug
2025-11-27 10:57 ` [PATCH v2 3/4] bootm: use new api to get kernel command line params Fabian Pflug
2025-11-27 10:57 ` [PATCH v2 4/4] bootm: introduce bootm.root_arg variable Fabian Pflug
2025-11-27 14:25 ` [PATCH v3 0/4] make the root= command line parameter variable Fabian Pflug
2025-11-27 14:25 ` [PATCH v3 1/4] block.h: renamed get_rootargs to get_root Fabian Pflug
2025-11-28 9:45 ` Ahmad Fatoum
2025-11-27 14:25 ` [PATCH v3 2/4] fs: split rootargs into root and options Fabian Pflug
2025-11-28 10:01 ` Ahmad Fatoum
2025-11-27 14:25 ` [PATCH v3 3/4] bootm: use new api to get kernel command line params Fabian Pflug
2025-11-28 10:04 ` Ahmad Fatoum
2025-11-27 14:25 ` [PATCH v3 4/4] bootm: introduce bootm.root_arg variable Fabian Pflug
2025-11-28 10:07 ` Ahmad Fatoum
2025-11-28 14:58 ` [PATCH v4 0/4] make the root= command line parameter variable Fabian Pflug
2025-11-28 14:59 ` [PATCH v4 1/4] block.h: renamed get_rootargs to get_root Fabian Pflug
2025-11-28 14:59 ` [PATCH 2/4] fs: split rootargs into root and options Fabian Pflug
2025-11-28 14:59 ` [PATCH v4 3/4] bootm: use new api to get kernel command line params Fabian Pflug
2025-11-28 14:59 ` [PATCH v4 4/4] bootm: introduce bootm.root_arg variable Fabian Pflug
2025-12-01 7:20 ` [PATCH v5 0/4] make the root= command line parameter variable Fabian Pflug
2025-12-01 7:20 ` [PATCH v5 1/4] block.h: renamed get_rootargs to get_root Fabian Pflug
2025-12-01 7:20 ` [PATCH v5 2/4] fs: split rootargs into root and options Fabian Pflug
2025-12-01 7:20 ` [PATCH v5 3/4] bootm: use new api to get kernel command line params Fabian Pflug
2025-12-01 7:21 ` [PATCH v5 4/4] bootm: introduce bootm.root_param variable Fabian Pflug
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=9c9ad61d-295f-4011-9418-dc2ed8da16a2@pengutronix.de \
--to=a.fatoum@pengutronix.de \
--cc=barebox@lists.infradead.org \
--cc=f.pflug@pengutronix.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox