mailarchive of the ptxdist mailing list
 help / color / mirror / Atom feed
From: Alexander Aring <alex.aring@gmail.com>
To: ptxdist@pengutronix.de
Subject: Re: [ptxdist] [PATCH 2/2] genimage: add workaround to work with glibc 2.22
Date: Tue, 25 Aug 2015 14:41:06 +0200	[thread overview]
Message-ID: <20150825124100.GA20806@omega> (raw)
In-Reply-To: <1440338218-18404-2-git-send-email-alex.aring@gmail.com>

On Sun, Aug 23, 2015 at 03:56:58PM +0200, Alexander Aring wrote:
> This patch adds a workaround for check if value argument of setenv is
> null. In all glibc versions setenv(..., NULL, ...) ends in an undef
> behaviour, since glibc 2.22 it will end in a core dump.
> 
> This is a workaround because it adds to every setenv a null check. I
> think some setenv are optional and some need to be set like name. Maybe
> we should make then a clear upstream patch which checks on this. I
> currently don't know which setenv is optional and which not. In my case
> of creating vfat image it seems some of them are optional like
> IMAGEMOUNTPOINT.
> 
> Signed-off-by: Alexander Aring <alex.aring@gmail.com>
> ---
>  ...ge-add-wrapper-for-handling-setenv-.-NULL.patch | 68 ++++++++++++++++++++++
>  patches/genimage-7/series                          |  4 ++
>  2 files changed, 72 insertions(+)
>  create mode 100644 patches/genimage-7/0001-genimage-add-wrapper-for-handling-setenv-.-NULL.patch
>  create mode 100644 patches/genimage-7/series
> 
> diff --git a/patches/genimage-7/0001-genimage-add-wrapper-for-handling-setenv-.-NULL.patch b/patches/genimage-7/0001-genimage-add-wrapper-for-handling-setenv-.-NULL.patch
> new file mode 100644
> index 0000000..79a2c3e
> --- /dev/null
> +++ b/patches/genimage-7/0001-genimage-add-wrapper-for-handling-setenv-.-NULL.patch
> @@ -0,0 +1,68 @@
> +From: Alexander Aring <alex.aring@gmail.com>
> +Date: Sun, 23 Aug 2015 13:17:38 +0200
> +Subject: [PATCH] genimage: add wrapper for handling setenv(..., NULL, ...)
> +
> +This patch adds a wrapper when calling setenv(..., NULL, ...) to a
> +no-op. Since glibc 2.22 this will end in a core dump.
> +
> +Signed-off-by: Alexander Aring <alex.aring@gmail.com>
> +---
> + genimage.c | 29 ++++++++++++++++++++---------
> + 1 file changed, 20 insertions(+), 9 deletions(-)
> +
> +diff --git a/genimage.c b/genimage.c
> +index e80127a2fbb9..9f60cc092eb4 100644
> +--- a/genimage.c
> ++++ b/genimage.c
> +@@ -239,6 +239,17 @@ static int image_generate(struct image *image)
> + 	return 0;
> + }
> + 
> ++/*
> ++ * Simple wrapper for setenv which doesn't run setenv if value == NULL
> ++ */
> ++static int image_setenv(const char *name, const char *value, int overwrite)
> ++{
> ++	if (!value)
> ++		return 0;
> ++	else
> ++		return setenv(name, value, overwrite);
> ++}

This sould be:

static inline int image_setenv(const char *name, const char *value, int overwrite)
{
	return setenv(name, value ? value : "", overwrite);
}

Current solution is 100% broken if generating multiple images in the
below list_for_each_entry loop it will not overwrite the previous
setting.

> ++
> + static LIST_HEAD(flashlist);
> + 
> + static int parse_flashes(cfg_t *cfg)
> +@@ -598,10 +609,10 @@ int main(int argc, char *argv[])
> + 		if (ret)
> + 			goto err_out;
> + 	}
> +-	setenv("OUTPUTPATH", imagepath(), 1);
> +-	setenv("INPUTPATH", inputpath(), 1);
> +-	setenv("ROOTPATH", rootpath(), 1);
> +-	setenv("TMPPATH", tmppath(), 1);
> ++	image_setenv("OUTPUTPATH", imagepath(), 1);
> ++	image_setenv("INPUTPATH", inputpath(), 1);
> ++	image_setenv("ROOTPATH", rootpath(), 1);
> ++	image_setenv("TMPPATH", tmppath(), 1);
> + 
> + 	ret = systemp(NULL, "mkdir -p %s", imagepath());
> + 	if (ret)
> +@@ -613,13 +624,13 @@ int main(int argc, char *argv[])
> + 
> + 	list_for_each_entry(image, &images, list) {

I mean this loop here.

> + 		char *sizestr;
> +-		setenv("IMAGE", image->file, 1);
> +-		setenv("IMAGEOUTFILE", imageoutfile(image), 1);
> +-		setenv("IMAGENAME", image->name, 1);
> ++		image_setenv("IMAGE", image->file, 1);
> ++		image_setenv("IMAGEOUTFILE", imageoutfile(image), 1);
> ++		image_setenv("IMAGENAME", image->name, 1);
> + 		asprintf(&sizestr, "%lld", image->size);
> +-		setenv("IMAGESIZE", sizestr, 1);
> ++		image_setenv("IMAGESIZE", sizestr, 1);
> + 		free(sizestr);
> +-		setenv("IMAGEMOUNTPOINT", image->mountpoint, 1);
> ++		image_setenv("IMAGEMOUNTPOINT", image->mountpoint, 1);
> + 		ret = image_generate(image);
> + 		if (ret) {
> + 			image_error(image, "failed to generate %s\n", image->file);

I will not send a v2 until mol's opinion according to that solution.

- Alex

-- 
ptxdist mailing list
ptxdist@pengutronix.de

      reply	other threads:[~2015-08-25 12:41 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-08-23 13:56 [ptxdist] [PATCH 1/2] wpan-tools: version bump 0.4 -> 0.5 Alexander Aring
2015-08-23 13:56 ` [ptxdist] [PATCH 2/2] genimage: add workaround to work with glibc 2.22 Alexander Aring
2015-08-25 12:41   ` Alexander Aring [this message]

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=20150825124100.GA20806@omega \
    --to=alex.aring@gmail.com \
    --cc=ptxdist@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