| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <ZoK6kW1FJiQ0ip0e@google.com>
Date: Mon, 1 Jul 2024 14:17:53 +0000
From: Sebastian Ene <sebastianene@...gle.com>
To: Oliver Upton <oliver.upton@...ux.dev>
Cc: akpm@...ux-foundation.org, alexghiti@...osinc.com, ankita@...dia.com,
ardb@...nel.org, catalin.marinas@....com,
christophe.leroy@...roup.eu, james.morse@....com,
vdonnefort@...gle.com, mark.rutland@....com, maz@...nel.org,
rananta@...gle.com, ryan.roberts@....com, shahuang@...hat.com,
suzuki.poulose@....com, will@...nel.org, yuzenghui@...wei.com,
kvmarm@...ts.linux.dev, linux-arm-kernel@...ts.infradead.org,
linux-kernel@...r.kernel.org, kernel-team@...roid.com
Subject: Re: [PATCH v7 5/6] KVM: arm64: Initialize the ptdump parser with
stage-2 attributes
On Fri, Jun 28, 2024 at 09:18:16PM +0000, Oliver Upton wrote:
> Hi Seb,
>
> On Fri, Jun 21, 2024 at 12:32:29PM +0000, Sebastian Ene wrote:
> > Define a set of attributes used by the ptdump parser to display the
> > properties of a guest memory region covered by a pagetable descriptor.
> > Build a description of the pagetable levels and initialize the parser
> > with this configuration.
> >
> > Signed-off-by: Sebastian Ene <sebastianene@...gle.com>
>
> This patch should come *before* patch 4, no point in exposing the
> debugfs file if we aren't ready to handle it yet.
>
Gotcha, let me try to reorder them.
> > ---
> > arch/arm64/kvm/ptdump.c | 143 ++++++++++++++++++++++++++++++++++++++--
> > 1 file changed, 137 insertions(+), 6 deletions(-)
> >
> > diff --git a/arch/arm64/kvm/ptdump.c b/arch/arm64/kvm/ptdump.c
> > index 36dc7662729f..cc1d4fdddc6e 100644
> > --- a/arch/arm64/kvm/ptdump.c
> > +++ b/arch/arm64/kvm/ptdump.c
> > @@ -14,6 +14,61 @@
> > #include <kvm_ptdump.h>
> >
> >
> > +#define MARKERS_LEN (2)
> > +#define KVM_PGTABLE_MAX_LEVELS (KVM_PGTABLE_LAST_LEVEL + 1)
> > +
> > +struct kvm_ptdump_guest_state {
> > + struct kvm *kvm;
> > + struct pg_state parser_state;
> > + struct addr_marker ipa_marker[MARKERS_LEN];
> > + struct pg_level level[KVM_PGTABLE_MAX_LEVELS];
> > + struct ptdump_range range[MARKERS_LEN];
> > +};
> > +
> > +static const struct prot_bits stage2_pte_bits[] = {
> > + {
> > + .mask = PTE_VALID,
> > + .val = PTE_VALID,
> > + .set = " ",
> > + .clear = "F",
> > + }, {
> > + .mask = KVM_PTE_LEAF_ATTR_HI_S2_XN | PTE_VALID,
> > + .val = KVM_PTE_LEAF_ATTR_HI_S2_XN | PTE_VALID,
> > + .set = "XN",
> > + .clear = " ",
> > + }, {
> > + .mask = KVM_PTE_LEAF_ATTR_LO_S2_S2AP_R | PTE_VALID,
> > + .val = KVM_PTE_LEAF_ATTR_LO_S2_S2AP_R | PTE_VALID,
> > + .set = "R",
> > + .clear = " ",
> > + }, {
> > + .mask = KVM_PTE_LEAF_ATTR_LO_S2_S2AP_W | PTE_VALID,
> > + .val = KVM_PTE_LEAF_ATTR_LO_S2_S2AP_W | PTE_VALID,
> > + .set = "W",
> > + .clear = " ",
> > + }, {
> > + .mask = KVM_PTE_LEAF_ATTR_LO_S2_AF | PTE_VALID,
> > + .val = KVM_PTE_LEAF_ATTR_LO_S2_AF | PTE_VALID,
> > + .set = "AF",
> > + .clear = " ",
>
> <snip>
>
> > + }, {
> > + .mask = PTE_NG,
> > + .val = PTE_NG,
> > + .set = "FnXS",
> > + .clear = " ",
> > + }, {
> > + .mask = PTE_CONT | PTE_VALID,
> > + .val = PTE_CONT | PTE_VALID,
> > + .set = "CON",
> > + .clear = " ",
> > + }, {
>
> </snip>
>
> Neither of these bits are used at stage-2, why have descriptors for
> them?
>
Atm, we don't make use of the contiguous bit in stage-2 in upstream (but
we have it in some experimental patches). I can remove this, no hard
feelings for them.
> > +static int kvm_ptdump_build_levels(struct pg_level *level, u32 start_lvl)
> > +{
> > + static const char * const level_names[] = {"PGD", "PUD", "PMD", "PTE"};
> > + u32 i = 0;
> > + u64 mask = 0;
> > +
> > + if (start_lvl > 2) {
> > + pr_err("invalid start_lvl %u\n", start_lvl);
> > + return -EINVAL;
> > + }
>
> if (WARN_ON_ONCE(start_lvl >= KVM_PGTABLE_LAST_LEVEL))
> return -EINVAL;
>
I will include this validation, thanks !
> > + for (i = 0; i < ARRAY_SIZE(stage2_pte_bits); i++)
> > + mask |= stage2_pte_bits[i].mask;
> > +
> > + for (i = start_lvl; i < KVM_PGTABLE_MAX_LEVELS; i++) {
> > + strscpy(level[i].name, level_names[i], sizeof(level[i].name));
> > +
> > + level[i].num = ARRAY_SIZE(stage2_pte_bits);
> > + level[i].bits = stage2_pte_bits;
> > + level[i].mask = mask;
> > + }
> > +
> > + if (start_lvl > 0)
> > + strscpy(level[start_lvl].name, level_names[0], sizeof(level_names[0]));
>
> This should pass the size of @dst, not the source. This becomes slightly
> more self-documenting if you use a literal for "PGD" here too.
>
> strscpy(level[start_lvl].name, "PGD", sizeof(level[start_lvl].name));
>
Will use this, thanks for the suggestion !
> > + return 0;
> > +}
> > +
> > +static struct kvm_ptdump_guest_state
> > +*kvm_ptdump_parser_init(struct kvm *kvm)
> > +{
> > + struct kvm_ptdump_guest_state *st;
> > + struct kvm_s2_mmu *mmu = &kvm->arch.mmu;
> > + struct kvm_pgtable *pgtable = mmu->pgt;
> > + int ret;
> > +
> > + st = kzalloc(sizeof(struct kvm_ptdump_guest_state), GFP_KERNEL_ACCOUNT);
> > + if (!st)
> > + return NULL;
> > +
> > + ret = kvm_ptdump_build_levels(&st->level[0], pgtable->start_level);
> > + if (ret)
> > + goto free_with_state;
>
> I don't see any value in the use of goto here, as there isn't any sort
> of cascading initialization / cleanup. This also presents an opportunity
> to get an error back out to the caller.
>
> if (ret) {
> kfree(st);
> return ERR_PTR(ret);
> }
>
Let me remove that goto; statement.
> > @@ -57,22 +176,34 @@ static int kvm_ptdump_guest_show(struct seq_file *m, void *unused)
> > static int kvm_ptdump_guest_open(struct inode *m, struct file *file)
> > {
> > struct kvm *kvm = m->i_private;
> > + struct kvm_ptdump_guest_state *st;
> > int ret;
> >
> > if (!kvm_get_kvm_safe(kvm))
> > return -ENOENT;
> >
> > - ret = single_open(file, kvm_ptdump_guest_show, m->i_private);
> > - if (ret < 0)
> > - kvm_put_kvm(kvm);
> > + st = kvm_ptdump_parser_init(kvm);
> > + if (!st) {
> > + ret = -ENOMEM;
> > + goto free_with_kvm_ref;
> > + }
>
> (with the earlier suggestion)
>
> st = kvm_ptdump_parser_init(kvm);
> if (IS_ERR(st)) {
> ret = PTR_ERR(st);
> goto free_with_kvm_ref;
> }
>
> Otherwise genuine KVM bugs (-EINVAL) are getting lumped into ENOMEM.
>
> --
> Thanks,
> Oliver
Thanks,
Sebastian
Powered by blists - more mailing lists