On Mon, Feb 14, 2022 at 11:43:10PM -0800, Lucas De Marchi wrote:
> On Sun, Feb 13, 2022 at 02:13:39PM +0100, Alexey Gladkov wrote:
> > On Sat, Feb 12, 2022 at 09:43:35PM -1000, Lucas De Marchi wrote:
> > > The modules.builtin.alias.bin is way larger than the
> > > modules.builtin.bin. On a normal "distro kernel":
> > >
> > > 21k modules.builtin.alias.bin
> > > 11k modules.builtin.bin
> > >
> > > >From the kernel we get both modules.builtin and modules.builtin.modinfo.
> > > depmod generates modules.builtin.bin and modules.builtin.alias.bin
> > > from them respectively. modules.bultin is not going away: it's not
> > > deprecated by the new index added. So, let's just stop duplicating the
> > > information inside modules.builtin.alias.bin and just use the other
> > > index.
> >
> > The modules.builtin contains only module names. The modules.builtin.modinfo
> > contains full info about builtin modules including names.
> >
> > I thought that if there is complete information about the modules, then
> > reading the index with only the names does not make sense. And only in the
> > absence of modules.builtin.modinfo, you need to fallback to the index
> > with the names.
>
> yeah, but most of the time we really need only the module name, so we
> can optimize for that. And since we are not getting rid of the other
> index, we can simply use it first the same way that for modules we first
> do lookup on lookup modules.dep and only later on modules.aliases.
Yes and no. We can be sure that we don't need aliases. The argument passed
to utilities can be very similar to the name of a module:
$ modinfo fs-ext4
name: ext4
filename: (builtin)
softdep: pre: crc32c
license: GPL
file: fs/ext4/ext4
description: Fourth Extended Filesystem
author: Remy Card, Stephen Tweedie, Andrew Morton, Andreas Dilger, Theodore Ts'o and others
alias: fs-ext4
alias: ext3
alias: fs-ext3
By the way, crc32c is also an alias:
$ modinfo crc32c
filename: /lib/modules/5.14.0.61-centos-alt1.el9/kernel/arch/x86/crypto/crc32c-intel.ko
alias: crypto-crc32c-intel
alias: crc32c-intel
alias: crypto-crc32c
alias: crc32c
license: GPL
description: CRC32c (Castagnoli) optimization using Intel Hardware.
author: Austin Zhang <austin.zhang@xxxxxxxxx>, Kent Liu <kent.liu@xxxxxxxxx>
rhelversion: 9.0
srcversion: 1F2B6A533C8243A4017180A
alias: cpu:type:x86,ven*fam*mod*:feature:*0094*
depends:
retpoline: Y
intree: Y
name: crc32c_intel
vermagic: 5.14.0.61-centos-alt1.el9 SMP preempt mod_unload modversions
This is because there are multiple implementations of crc32c. So, information
about alias of builtin modules is almost always needed.
> >
> > > ---
> > > libkmod/libkmod-module.c | 12 +++++++-----
> > > 1 file changed, 7 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/libkmod/libkmod-module.c b/libkmod/libkmod-module.c
> > > index 6f7747c..6423339 100644
> > > --- a/libkmod/libkmod-module.c
> > > +++ b/libkmod/libkmod-module.c
> > > @@ -576,13 +576,15 @@ KMOD_EXPORT int kmod_module_new_from_lookup(struct kmod_ctx *ctx,
> > > err = kmod_lookup_alias_from_aliases_file(ctx, alias, list);
> > > CHECK_ERR_AND_FINISH(err, fail, list, finish);
> > >
> > > + DBG(ctx, "lookup modules.builtin %s\n", alias);
> > > + err = kmod_lookup_alias_from_builtin_file(ctx, alias, list);
> > > + CHECK_ERR_AND_FINISH(err, fail, list, finish);
> > > +
> > > DBG(ctx, "lookup modules.builtin.modinfo %s\n", alias);
> > > err = kmod_lookup_alias_from_kernel_builtin_file(ctx, alias, list);
> > > - if (err == -ENOSYS) {
> > > - /* Optional index missing, try the old one */
> > > - DBG(ctx, "lookup modules.builtin %s\n", alias);
> > > - err = kmod_lookup_alias_from_builtin_file(ctx, alias, list);
> > > - }
> > > + /* Optional index missing, ignore */
> > > + if (err == -ENOSYS)
> > > + err = 0;
> > > CHECK_ERR_AND_FINISH(err, fail, list, finish);
> >
> > Yep. Looks good for me.
>
> thanks
> Lucas De Marchi
>
> >
> > --
> > Rgrds, legion
> >
>
--
Rgrds, legion
