fix CVE-2026-4046, CVE-2026-5450, and CVE-2026-5928 in glibc

[Xi Ruoyao]

I'm really not a fan of "AI assisted" security vulnerability mining but ... it already happens.

I've uploaded a consolidated patch containing the fixes for them, the fixes for two prior vulnerabilities (currently as a sed command in book), and two fixes for Linux 7.0 so we can use it in mid-May update.

There are also CVE-2026-5435 and CVE-2026-6238 but they are not committed upstream yet and they only affect some interfaces deprecated for years, so I didn't include them here. If the fixes get committed before mid-May update we can append them into the patch.

[Xi Ruoyao]

​https://linuxfromscratch.org/patches/downloads/glibc/glibc-2.43-consolidated-1.patch

[Bruce Dubbs]

I need clarification.

In chapter 5 we use glibc-fhs-1.patch. Is that still needed? I think yes, but I'm not sure. Is the glibc-2.43-consolidated-1.patch needed in chapter 5? I think no, but again I'm not sure. It probably wouldn't hurt though.

In Chapter 8, we now have a sed and the fhs patch. Are they still needed in addition to the glibc-2.43-consolidated-1.patch? My understanding is that the sed is incorporated into the consolidated patch and the fhs patch is not. Correct?

[Douglas R. Reno]

The consolidated patch will be needed in Chapter 5, otherwise we can't build glibc against the Linux 7 API headers

[Xi Ruoyao]

Yes, the sed is in the patch but the FHS patch is not.

(I didn't consolidate the FHS patch because it seems we are supposed to keep the FHS patch for a lot of future glibc versions).

[Bruce Dubbs]

Fixed at commit 6d990d4871. Leaving open for security advisory.