[v3] target/s390x: Prohibit target_ulong and TARGET_PAGE_SIZE uses

[PATCH v3 01/10] target/s390x: Use address_space_ldl_be() in read_table_entry()

Posted by Philippe Mathieu-Daudé 2 days, 22 hours ago

address_space_read/write() is meant for accessing random
amount of memory blobs.  When the access size is known,
use the address_space_ld/st() API which can directly swap
endianness.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Eric Farman <farman@linux.ibm.com>
---
 target/s390x/mmu_helper.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

diff --git a/target/s390x/mmu_helper.c b/target/s390x/mmu_helper.c
index 026502a3e40..30f09ec3de4 100644
--- a/target/s390x/mmu_helper.c
+++ b/target/s390x/mmu_helper.c
@@ -108,6 +108,7 @@ static inline bool read_table_entry(CPUS390XState *env, hwaddr gaddr,
                                     uint64_t *entry)
 {
     CPUState *cs = env_cpu(env);
+    MemTxResult ret;
 
     /*
      * According to the PoP, these table addresses are "unpredictably real
@@ -116,13 +117,9 @@ static inline bool read_table_entry(CPUS390XState *env, hwaddr gaddr,
      *
      * We treat them as absolute addresses and don't wrap them.
      */
-    if (unlikely(address_space_read(cs->as, gaddr, MEMTXATTRS_UNSPECIFIED,
-                                    entry, sizeof(*entry)) !=
-                 MEMTX_OK)) {
-        return false;
-    }
-    *entry = be64_to_cpu(*entry);
-    return true;
+    *entry = address_space_ldq_be(cs->as, gaddr, MEMTXATTRS_UNSPECIFIED, &ret);
+
+    return ret == MEMTX_OK;
 }
 
 static int mmu_translate_asce(CPUS390XState *env, target_ulong vaddr,
-- 
2.52.0

Re: [PATCH v3 01/10] target/s390x: Use address_space_ldl_be() in read_table_entry()

Posted by Thomas Huth 9 hours ago

Still:
s/address_space_ldl_be/address_space_ldq_be/
in the subject.

  Thomas


On 06/02/2026 19.19, Philippe Mathieu-Daudé wrote:
> address_space_read/write() is meant for accessing random
> amount of memory blobs.  When the access size is known,
> use the address_space_ld/st() API which can directly swap
> endianness.
> 
> Reviewed-by: Thomas Huth <thuth@redhat.com>
> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> Reviewed-by: Eric Farman <farman@linux.ibm.com>
> ---
>   target/s390x/mmu_helper.c | 11 ++++-------
>   1 file changed, 4 insertions(+), 7 deletions(-)
> 
> diff --git a/target/s390x/mmu_helper.c b/target/s390x/mmu_helper.c
> index 026502a3e40..30f09ec3de4 100644
> --- a/target/s390x/mmu_helper.c
> +++ b/target/s390x/mmu_helper.c
> @@ -108,6 +108,7 @@ static inline bool read_table_entry(CPUS390XState *env, hwaddr gaddr,
>                                       uint64_t *entry)
>   {
>       CPUState *cs = env_cpu(env);
> +    MemTxResult ret;
>   
>       /*
>        * According to the PoP, these table addresses are "unpredictably real
> @@ -116,13 +117,9 @@ static inline bool read_table_entry(CPUS390XState *env, hwaddr gaddr,
>        *
>        * We treat them as absolute addresses and don't wrap them.
>        */
> -    if (unlikely(address_space_read(cs->as, gaddr, MEMTXATTRS_UNSPECIFIED,
> -                                    entry, sizeof(*entry)) !=
> -                 MEMTX_OK)) {
> -        return false;
> -    }
> -    *entry = be64_to_cpu(*entry);
> -    return true;
> +    *entry = address_space_ldq_be(cs->as, gaddr, MEMTXATTRS_UNSPECIFIED, &ret);
> +
> +    return ret == MEMTX_OK;
>   }
>   
>   static int mmu_translate_asce(CPUS390XState *env, target_ulong vaddr,