From: Luis Machado <luis.machado@arm.com>
To: Lancelot SIX <lsix@lancelotsix.com>
Cc: gdb-patches@sourceware.org, jhb@FreeBSD.org, thiago.bauermann@linaro.org
Subject: Re: [PATCH,v3] [aarch64] Fix removal of non-address bits for PAuth
Date: Thu, 22 Sep 2022 17:39:23 +0100 [thread overview]
Message-ID: <b79d6348-af57-9734-ce40-c4eca6cc0e06@arm.com> (raw)
In-Reply-To: <20220922125805.hvekyxcf3nc2i764@ubuntu.lan>
Hi,
On 9/22/22 13:59, Lancelot SIX wrote:
> Hi Luis,
>
> I went through the patch and have a couple of questions above.
>
>> diff --git a/gdb/aarch64-linux-tdep.c b/gdb/aarch64-linux-tdep.c
>> index 15773c75da8..279c8d98f5d 100644
>> --- a/gdb/aarch64-linux-tdep.c
>> +++ b/gdb/aarch64-linux-tdep.c
>> @@ -1787,7 +1787,8 @@ aarch64_linux_report_signal_info (struct gdbarch *gdbarch,
>> uiout->text ("\n");
>>
>> gdb::optional<CORE_ADDR> atag
>> - = aarch64_mte_get_atag (address_significant (gdbarch, fault_addr));
>> + = aarch64_mte_get_atag (gdbarch_remove_non_address_bits (gdbarch,
>> + fault_addr));
>> gdb_byte ltag = aarch64_mte_get_ltag (fault_addr);
>>
>> if (!atag.has_value ())
>> @@ -1961,6 +1962,47 @@ aarch64_linux_decode_memtag_section (struct gdbarch *gdbarch,
>> return tags;
>> }
>>
>> +/* AArch64 implementation of the remove_non_address_bits gdbarch hook. Remove
>> + non address bits from a pointer value. */
>> +
>> +static CORE_ADDR
>> +aarch64_remove_non_address_bits (struct gdbarch *gdbarch, CORE_ADDR pointer)
>> +{
>> + aarch64_gdbarch_tdep *tdep = gdbarch_tdep<aarch64_gdbarch_tdep> (gdbarch);
>> +
>> + /* By default, we assume TBI and discard the top 8 bits plus the VA range
>> + select bit (55). */
>> + CORE_ADDR mask = AARCH64_TOP_BITS_MASK;
>> +
>> + if (tdep->has_pauth ())
>> + {
>> + /* Fetch the PAC masks. These masks are per-process, so we can just
>> + fetch data from whatever thread we have at the moment.
>> +
>> + Also, we have both a code mask and a data mask. For now they are the
>> + same, but this may change in the future. */
>> + struct regcache *regs = get_current_regcache ();
>> + CORE_ADDR cmask, dmask;
>> +
>> + if (regs->cooked_read (tdep->pauth_reg_base, &dmask) != REG_VALID)
>> + dmask = mask;
>> +
>> + if (regs->cooked_read (tdep->pauth_reg_base + 1, &cmask) != REG_VALID)
>> + cmask = mask;
>> +
>> + if (dmask != cmask)
>> + {
>> + /* Warn if the masks are different. */
>> + aarch64_pauth_mask_warning ();
>> + mask |= dmask > cmask? dmask : cmask;
>> + }
>> + else
>> + mask |= cmask;
>
> Here, I am wondering what happens if either cooked_read does not return
> ROG_VALID. Wouldn't cmask/dmask have un-initialized values, making the
> end of the method hazardous?
>
> I guess initializing both to 0 would solve this.
>
If either one of the register reads fail, then we assign the default MASK. Otherwise we
continue with using the mask that's been read from the register.
Does that make sense?
>> + }
>> +
>> + return aarch64_remove_top_bits (pointer, mask);
>> +}
>> +
>> static void
>> aarch64_linux_init_abi (struct gdbarch_info info, struct gdbarch *gdbarch)
>> {
>> index 0f73286f145..d9c4b994850 100644
>> --- a/gdb/arch/aarch64.c
>> +++ b/gdb/arch/aarch64.c
>> @@ -58,3 +58,30 @@ aarch64_create_target_description (const aarch64_features &features)
>>
>> return tdesc.release ();
>> }
>> +
>> +/* See arch/aarch64.h. */
>> +
>> +CORE_ADDR
>> +aarch64_remove_top_bits (CORE_ADDR pointer, CORE_ADDR mask)
>> +{
>> + /* The VA range select bit is 55. This bit tells us if we have a
>> + kernel-space address or a user-space address. */
>> + bool kernel_address = (pointer & VA_RANGE_SELECT_BIT_MASK) != 0;
>> +
>
> I am wondering: is this Linux specific or is this valid accross all
> configurations? If this is linux specific, is aarch64.c the right place
> to implement this?
Although this is only used on Linux systems at the moment, it is not technically Linux-specific.
When we say kernel-space, it means the other half of the VA space (non-user).
I have an upcoming patch (relying on this one) to use this function for bare metal pointer authentication
support (with user QEMU). So I think it makes sense to have this function in arch-specific code
and not in Linux-specific code.
>
> Best,
> Lancelot.
>
>> + /* Remove the top non-address bits. */
>> + pointer &= ~mask;
>> +
>> + /* Sign-extend if we have a kernel-space address. */
>> + if (kernel_address)
>> + pointer |= mask;
>> +
>> + return pointer;
>> +}
>> +
>> +/* See arch/aarch64.h. */
>> +
>> +void
>> +aarch64_pauth_mask_warning ()
>> +{
>> + warning (_("Pointer authentication masks for code (C) and data (D) differ"));
>> +}
next prev parent reply other threads:[~2022-09-22 16:39 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-07-05 14:00 [PATCH] [AArch64] " Luis Machado
2022-07-05 18:12 ` John Baldwin
2022-07-06 11:38 ` Lancelot SIX
2022-07-08 11:36 ` Luis Machado
2022-07-11 11:55 ` [PATCH,v2] [aarch64] " Luis Machado
2022-07-18 8:16 ` [Ping v1][PATCH,v2] " Luis Machado
2022-08-01 11:09 ` [Ping v2][PATCH,v2] " Luis Machado
2022-08-08 11:34 ` [Ping v3][PATCH,v2] " Luis Machado
2022-08-18 15:49 ` [Ping v4][PATCH,v2] " Luis Machado
2022-08-18 23:47 ` [PATCH,v2] " Thiago Jung Bauermann
2022-08-19 9:52 ` Luis Machado
2022-08-19 14:06 ` Thiago Jung Bauermann
2022-08-23 20:29 ` [PATCH,v3] " Luis Machado
2022-08-24 18:44 ` Thiago Jung Bauermann
2022-09-01 9:29 ` [PING][PATCH,v3] " Luis Machado
2022-09-07 8:21 ` Luis Machado
2022-09-12 12:47 ` Luis Machado
2022-09-20 12:26 ` Luis Machado
2022-09-22 12:59 ` [PATCH,v3] " Lancelot SIX
2022-09-22 16:39 ` Luis Machado [this message]
2022-09-23 7:58 ` Lancelot SIX
2022-10-03 11:37 ` [PING][PATCH,v3] " Luis Machado
2022-10-10 12:18 ` Luis Machado
2022-10-17 10:04 ` Luis Machado
2022-10-25 13:52 ` Luis Machado
2022-11-10 1:00 ` Luis Machado
2022-11-29 22:19 ` Luis Machado
2022-12-09 16:42 ` Luis Machado
2022-12-09 19:14 ` [PATCH,v3] " Simon Marchi
2022-12-12 14:21 ` Luis Machado
2022-12-12 15:07 ` Simon Marchi
2022-12-12 17:13 ` [PATCH v4] " Luis Machado
2022-12-12 18:54 ` Simon Marchi
2022-12-13 9:18 ` Luis Machado
2022-12-13 10:27 ` [PATCH v5] " Luis Machado
2022-12-16 10:57 ` [PATCH v6] " Luis Machado
2022-12-16 11:20 ` Luis Machado
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=b79d6348-af57-9734-ce40-c4eca6cc0e06@arm.com \
--to=luis.machado@arm.com \
--cc=gdb-patches@sourceware.org \
--cc=jhb@FreeBSD.org \
--cc=lsix@lancelotsix.com \
--cc=thiago.bauermann@linaro.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 read-only IMAP folder(s) and NNTP newsgroup(s).