From: Peter Robinson <pbrobinson@gmail.com>
To: Josh Boyer <jwboyer@kernel.org>
Cc: Linux Firmware <linux-firmware@kernel.org>
Subject: Re: [PATCH v2] Add support for compressing firmware in copy-firmware.sh
Date: Mon, 8 Feb 2021 13:50:57 +0000 [thread overview]
Message-ID: <CALeDE9O+uDmJVGvRLXc4sE00X-8TGkDE-8Hcxs_wDfjo7vgMuQ@mail.gmail.com> (raw)
In-Reply-To: <CA+5PVA7c2_mrndhqhrX5DzGcpaW4Y_-Q3HedYXtjGq5S+wO8nA@mail.gmail.com>
On Mon, Feb 8, 2021 at 1:47 PM Josh Boyer <jwboyer@kernel.org> wrote:
>
> On Sat, Jan 23, 2021 at 5:26 AM Peter Robinson <pbrobinson@gmail.com> wrote:
> >
> > As of kernel 5.3 there's initial support for loading compressed firmware.
> > At this stage the only supported compression methis is "xz -C crc32" but
> > this option brings significant benefits.
> >
> > Signed-off-by: Peter Robinson <pbrobinson@gmail.com>
> > ---
> >
> > v2: quote filename for xz command
>
> That seems fixed, I ran into another issue.
>
> > Makefile | 4 ++++
> > copy-firmware.sh | 47 +++++++++++++++++++++++++++++++----------------
> > 2 files changed, 35 insertions(+), 16 deletions(-)
> >
> > diff --git a/Makefile b/Makefile
> > index e1c362f..9a48471 100644
> > --- a/Makefile
> > +++ b/Makefile
> > @@ -11,3 +11,7 @@ check:
> > install:
> > mkdir -p $(DESTDIR)$(FIRMWAREDIR)
> > ./copy-firmware.sh $(DESTDIR)$(FIRMWAREDIR)
> > +
> > +installcompress:
> > + mkdir -p $(DESTDIR)$(FIRMWAREDIR)
> > + ./copy-firmware.sh -C $(DESTDIR)$(FIRMWAREDIR)
> > diff --git a/copy-firmware.sh b/copy-firmware.sh
> > index 9b46b63..0dd2e5c 100755
> > --- a/copy-firmware.sh
> > +++ b/copy-firmware.sh
> > @@ -6,6 +6,7 @@
> >
> > verbose=:
> > prune=no
> > +compress=no
> >
> > while test $# -gt 0; do
> > case $1 in
> > @@ -19,6 +20,11 @@ while test $# -gt 0; do
> > shift
> > ;;
> >
> > + -C | --compress)
> > + compress=yes
> > + shift
> > + ;;
> > +
> > *)
> > if test "x$destdir" != "x"; then
> > echo "ERROR: unknown command-line options: $@"
> > @@ -31,40 +37,49 @@ while test $# -gt 0; do
> > esac
> > done
> >
> > +if test "x$compress" = "xyes"; then
> > + cmpxtn=.xz
> > + grep '^File:' WHENCE | sed -e's/^File: *//g' -e's/"//g' | while read f; do
> > + test -f "$f" || continue
> > + $verbose "compressing $f"
> > + xz -C crc32 "$f"
> > + done
> > +fi
>
> This compresses the firmware in the local git directory. If you ever
> run this once, it'll never compress any updated files if you run it
> again and instead you'll get a bunch of "xz: <file>: File exists"
> errors:
>
> https://paste.centos.org/view/75c479da
>
> That seems counter-intuitive and leaves everyone's git tree with a ton
> of untracked files. Options I see:
>
> 1) Compress in place and copy, as-is in the patch, but then go back
> and remove all the compressed files after the copy
> 2) Compress to a temp location, move to the destination from there
> 3) Use xz to directly output to the destination file instead of compress/copy
>
> I kind of like 3. Thoughts?
Yes, I think that makes sense, I'd been looking at this from a distro
packaging perspective where it's generally run from the tarfile output
and then that is disgarded so less of an issue. Will take another pass
at it.
> josh
>
> > +
> > grep '^File:' WHENCE | sed -e's/^File: *//g' -e's/"//g' | while read f; do
> > - test -f "$f" || continue
> > - $verbose "copying file $f"
> > - mkdir -p $destdir/$(dirname "$f")
> > - cp -d "$f" $destdir/"$f"
> > + test -f "$f$cmpxtn" || continue
> > + $verbose "copying file $f$cmpxtn"
> > + mkdir -p $destdir/$(dirname "$f$cmpxtn")
> > + cp -d "$f$cmpxtn" $destdir/"$f$cmpxtn"
> > done
> >
> > grep -E '^Link:' WHENCE | sed -e's/^Link: *//g' -e's/-> //g' | while read f d; do
> > - if test -L "$f"; then
> > - test -f "$destdir/$f" && continue
> > - $verbose "copying link $f"
> > - mkdir -p $destdir/$(dirname "$f")
> > + if test -L "$f$cmpxtn"; then
> > + test -f "$destdir/$f$cmpxtn" && continue
> > + $verbose "copying link $f$cmpxtn"
> > + mkdir -p $destdir/$(dirname "$f$cmpxtn")
> > cp -d "$f" $destdir/"$f"
> >
> > if test "x$d" != "x"; then
> > - target=`readlink "$f"`
> > + target=`readlink "$f$cmpxtn"`
> >
> > if test "x$target" != "x$d"; then
> > $verbose "WARNING: inconsistent symlink target: $target != $d"
> > else
> > if test "x$prune" != "xyes"; then
> > - $verbose "WARNING: unneeded symlink detected: $f"
> > + $verbose "WARNING: unneeded symlink detected: $f$cmpxtn"
> > else
> > - $verbose "WARNING: pruning unneeded symlink $f"
> > - rm -f "$f"
> > + $verbose "WARNING: pruning unneeded symlink $f$cmpxtn"
> > + rm -f "$f$cmpxtn"
> > fi
> > fi
> > else
> > - $verbose "WARNING: missing target for symlink $f"
> > + $verbose "WARNING: missing target for symlink $f$cmpxtn"
> > fi
> > else
> > - $verbose "creating link $f -> $d"
> > - mkdir -p $destdir/$(dirname "$f")
> > - ln -sf "$d" "$destdir/$f"
> > + $verbose "creating link $f$cmpxtn -> $d$cmpxtn"
> > + mkdir -p $destdir/$(dirname "$f$cmpxtn")
> > + ln -sf "$d$cmpxtn" "$destdir/$f$cmpxtn"
> > fi
> > done
> >
> > --
> > 2.29.2
> >
prev parent reply other threads:[~2021-02-08 13:51 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-01-23 10:26 [PATCH v2] Add support for compressing firmware in copy-firmware.sh Peter Robinson
2021-02-08 13:47 ` Josh Boyer
2021-02-08 13:50 ` Peter Robinson [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=CALeDE9O+uDmJVGvRLXc4sE00X-8TGkDE-8Hcxs_wDfjo7vgMuQ@mail.gmail.com \
--to=pbrobinson@gmail.com \
--cc=jwboyer@kernel.org \
--cc=linux-firmware@kernel.org \
/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;
as well as URLs for NNTP newsgroup(s).