Message ID | 20200828165513.25649-1-mjw@redhat.com |
---|---|
State | New |
Headers | show |
Series | elf.h: Add aarch64 bti/pac dynamic tag constants | expand |
* Mark Wielaard via Libc-alpha: > From: Mark Wielaard <mark@klomp.org> > > Constants double checked against binutils and the ELF for the Arm 64-bit > Architecture (AArch64) Release 2020Q2 document. > --- > elf/elf.h | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/elf/elf.h b/elf/elf.h > index 197b557d15..ff9f1dadd5 100644 > --- a/elf/elf.h > +++ b/elf/elf.h > @@ -2875,6 +2875,8 @@ enum > #define R_AARCH64_IRELATIVE 1032 /* STT_GNU_IFUNC relocation. */ > > /* AArch64 specific values for the Dyn d_tag field. */ > +#define DT_AARCH64_BTI_PLT (DT_LOPROC + 1) > +#define DT_AARCH64_PAC_PLT (DT_LOPROC + 3) > #define DT_AARCH64_VARIANT_PCS (DT_LOPROC + 5) > #define DT_AARCH64_NUM 6 I think DT_AARCH64_PAC_PLT isn't really used, so maybe mentioned that in the commit message? There's no PAC PLT with glibc, and people are expected to use BIND_NOW. (I hope I remember this correctly.) Thanks, Florian
Hi Florian, On Fri, 2020-08-28 at 20:03 +0200, Florian Weimer wrote: > I think DT_AARCH64_PAC_PLT isn't really used, so maybe mentioned that > in the commit message? There's no PAC PLT with glibc, and people are > expected to use BIND_NOW. OK, how about with the attached commit message? Thanks, Mark
* Mark Wielaard: > Hi Florian, > > On Fri, 2020-08-28 at 20:03 +0200, Florian Weimer wrote: >> I think DT_AARCH64_PAC_PLT isn't really used, so maybe mentioned that >> in the commit message? There's no PAC PLT with glibc, and people are >> expected to use BIND_NOW. > > OK, how about with the attached commit message? Yes, this is what I had in mind, but please wait for an ack from an AArch64 maintainer. Thanks, Florian
On Fri, 2020-08-28 at 21:04 +0200, Florian Weimer wrote: > * Mark Wielaard: > > On Fri, 2020-08-28 at 20:03 +0200, Florian Weimer wrote: > > > I think DT_AARCH64_PAC_PLT isn't really used, so maybe mentioned > > > that > > > in the commit message? There's no PAC PLT with glibc, and people > > > are > > > expected to use BIND_NOW. > > > > OK, how about with the attached commit message? > > Yes, this is what I had in mind, but please wait for an ack from an > AArch64 maintainer. OK. Ping. Any AArch64 maintainer want to ack or nack this commit? Thanks, Mark
* Mark Wielaard <mark@klomp.org> [2020-09-02 13:28:02 +0200]: > On Fri, 2020-08-28 at 21:04 +0200, Florian Weimer wrote: > > * Mark Wielaard: > > > On Fri, 2020-08-28 at 20:03 +0200, Florian Weimer wrote: > > > > I think DT_AARCH64_PAC_PLT isn't really used, so maybe mentioned > > > > that > > > > in the commit message? There's no PAC PLT with glibc, and people > > > > are > > > > expected to use BIND_NOW. > > > > > > OK, how about with the attached commit message? > > > > Yes, this is what I had in mind, but please wait for an ack from an > > AArch64 maintainer. > > OK. Ping. Any AArch64 maintainer want to ack or nack this commit? i'm on a long holiday, i will be back on sept 21. the patch looks good, but i wont use my @arm.com email until i'm back. > > Thanks, > > Mark > From 3b925f004607cebc167af907c053bd7ef7146a3b Mon Sep 17 00:00:00 2001 > From: Mark Wielaard <mark@klomp.org> > Date: Fri, 28 Aug 2020 18:44:16 +0200 > Subject: [PATCH] elf.h: Add aarch64 bti/pac dynamic tag constants > > Constants double checked against binutils and the ELF for the Arm 64-bit > Architecture (AArch64) Release 2020Q2 document. > > Only BTI PLT is used in glibc, there's no PAC PLT with glibc, and people > are expected to use BIND_NOW. > --- > elf/elf.h | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/elf/elf.h b/elf/elf.h > index 197b557d15..ff9f1dadd5 100644 > --- a/elf/elf.h > +++ b/elf/elf.h > @@ -2875,6 +2875,8 @@ enum > #define R_AARCH64_IRELATIVE 1032 /* STT_GNU_IFUNC relocation. */ > > /* AArch64 specific values for the Dyn d_tag field. */ > +#define DT_AARCH64_BTI_PLT (DT_LOPROC + 1) > +#define DT_AARCH64_PAC_PLT (DT_LOPROC + 3) > #define DT_AARCH64_VARIANT_PCS (DT_LOPROC + 5) > #define DT_AARCH64_NUM 6 > > -- > 2.18.4 >
On Sun, 2020-09-06 at 22:21 +0200, Szabolcs Nagy wrote: > * Mark Wielaard <mark@klomp.org> [2020-09-02 13:28:02 +0200]: > > On Fri, 2020-08-28 at 21:04 +0200, Florian Weimer wrote: > > > * Mark Wielaard: > > > > On Fri, 2020-08-28 at 20:03 +0200, Florian Weimer wrote: > > > > > I think DT_AARCH64_PAC_PLT isn't really used, so maybe mentioned > > > > > that in the commit message? There's no PAC PLT with glibc, and > > > > > people are expected to use BIND_NOW. > > > > > > > > OK, how about with the attached commit message? > > > > > > Yes, this is what I had in mind, but please wait for an ack from an > > > AArch64 maintainer. > > > > OK. Ping. Any AArch64 maintainer want to ack or nack this commit? > > i'm on a long holiday, i will be back on sept 21. > > the patch looks good Thanks pushed. Have a nice holiday. Cheers, Mark
diff --git a/elf/elf.h b/elf/elf.h index 197b557d15..ff9f1dadd5 100644 --- a/elf/elf.h +++ b/elf/elf.h @@ -2875,6 +2875,8 @@ enum #define R_AARCH64_IRELATIVE 1032 /* STT_GNU_IFUNC relocation. */ /* AArch64 specific values for the Dyn d_tag field. */ +#define DT_AARCH64_BTI_PLT (DT_LOPROC + 1) +#define DT_AARCH64_PAC_PLT (DT_LOPROC + 3) #define DT_AARCH64_VARIANT_PCS (DT_LOPROC + 5) #define DT_AARCH64_NUM 6
From: Mark Wielaard <mark@klomp.org> Constants double checked against binutils and the ELF for the Arm 64-bit Architecture (AArch64) Release 2020Q2 document. --- elf/elf.h | 2 ++ 1 file changed, 2 insertions(+)